­­­Divergent Neurodegenerative Patterns: Comparison of FDG-PET- and MRI-based Alzheimer’s Disease Subtypes

­­­Divergent Neurodegenerative Patterns: Comparison of FDG-PET- and MRI-based Alzheimer’s Disease Subtypes | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article ­­­Divergent Neurodegenerative Patterns: Comparison of FDG-PET- and MRI-based Alzheimer’s Disease Subtypes Sophia H. Wheatley, Rosaleena Mohanty, Konstantinos Poulakis, and 6 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-4454593/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 22 Nov, 2024 Read the published version in Brain Communications → Version 1 posted You are reading this latest preprint version Abstract [ 18 F] fluorodeoxyglucose (FDG)-PET and MRI are key imaging markers for neurodegeneration in Alzheimer's disease. It is well-established that parieto-temporal hypometabolism on FDG-PET is closely associated with medial temporal atrophy on MRI in Alzheimer's disease. Substantial biological heterogeneity, expressed as distinct subtypes of hypometabolism or atrophy patterns, has been previously described in Alzheimer's disease using data-driven and hypothesis-driven methods. However, the link between these two imaging modalities has not yet been explored in the context of Alzheimer's disease subtypes. To investigate this link, the current study utilised FDG-PET and MRI scans from 180 amyloid-beta positive Alzheimer's disease dementia patients and 176 amyloid-beta negative cognitively normal controls from the Alzheimer’s Disease Neuroimaging Initiative (ADNI). Random forest hierarchical clustering, a data-driven model for identifying subtypes, was implemented in the two modalities: one with standard uptake value ratios and the other with grey matter volumes. Five subtypes hypometabolism- and atrophy-based subtypes were identified, exhibiting both cortical-predominant and limbic-predominant patterns although with differing percentages and clinical presentations. Three cortical-predominant hypometabolism subtypes found were: Cortical Predominant (32%), Cortical Predominant+ (11%), Cortical Predominant posterior (9%); and two limbic-predominant hypometabolism subtypes: Limbic Predominant (36%) and Limbic Predominant (13%). In addition, minimal and diffuse neurodegeneration subtypes were observed from the MRI data. The five atrophy subtypes were found: Cortical Predominant (19%), Limbic Predominant (27%), Diffuse (28%), Diffuse+ (6%) and Minimal (19%). Inter-modality comparisons showed that all FDG-PET subtypes displayed medial temporal atrophy, whereas the distinct MRI subtypes showed topographically similar hypometabolism. Further, allocations of FDG-PET and MRI subtypes were not consistent when compared at an individual-level. Additional analysis comparing the data-driven clustering model with prior hypothesis-driven methods showed only partial agreement between these subtyping methods. FDG-PET subtypes had greater differences between limbic-predominant and cortical-predominant patterns and MRI subtypes had greater differences in severity of atrophy. In conclusion, this study highlighted that Alzheimer's disease subtypes identified using both FDG-PET and MRI capture distinct pathways showing cortical versus limbic predominance of neurodegeneration. However, the subtypes do not share a bidirectional relationship between modalities and are thus not interchangeable. Nuclear Medicine & Medical Imaging Neurobiology of Disease Alzheimer's disease data-driven models subtypes hypometabolism atrophy PET MRI Figures Figure 1 Figure 2 Figure 3 Figure 4 Introduction Alzheimer’s disease (AD) is the most common neurodegenerative disease 1 and is heterogeneous in nature 2 . Over the last years, distinct AD subtypes displaying different clinical and biological features have been identified 2,3 . Despite this, there are discrepancies in findings across the studies, especially at an individual-level 4 . The common consensus from AD subtyping research is the presence of distinct cortical-predominant versus limbic-predominant profiles of neurodegeneration, which primarily reflect differential distributions of tau neurofibrillary tangle (NFTs) pathology in neuropathological findings. Corticolimbic subtypes were first identified from neuropathological data by distributions of NFTs in cortical and limbic regions. Such AD subtypes were coined as ‘Hippocampal Sparing’ and ‘Limbic Predominant’ when compared to ‘Typical AD’ 5 . Typical AD is described by a balanced distribution of NFTs in the cortical and limbic regions, whereas the other two subtypes are described at opposite ends of the spectrum, from low to high hippocampal NFT load relative to neocortical regions. Similar hypothesis-driven approaches have been applied to in vivo methods such as structural MRI, where atrophy patterns resembled prior neuropathologically-defined subtypes 6,7 . Furthermore, AD subtypes have been observed in data-driven studies using in vivo imaging 2,8 . A framework was built based on tau pathology and atrophy in AD subtype topography along two axes of ‘typicality’ and ‘severity’ 2 . ‘Typicality’ was proposed as a spectrum with ‘Limbic Predominant’ and ‘Hippocampal Sparing’ on opposite ends with ‘Typical AD’ in the middle. Whereas ‘severity’ reflects findings from atrophy-based studies with ‘Minimal’ (low atrophy) and ‘Diffuse’ (high atrophy) patterns as extremes compared to ‘Typical AD.’ It is not certain yet whether this framework could be applied to imaging modalities other than tau PET and MRI-based atrophy. Furthermore, longitudinal MRI studies of AD subtypes revealed that neurodegeneration progressed along either a cortical or a limbic pathway 9 . Similarly, modelling the spread of tau pathology in AD, two pathways were proposed starting either in the entorhinal cortex or in association cortices 10 . Molecular imaging complements MRI by measuring functional changes for the study of neuropathological hallmarks of AD. However, the main body of research in AD heterogeneity has used MRI scans, although research is branching out to other imaging modalities. There are few studies investigating AD heterogeneity using [ 18 F] fluorodeoxyglucose (FDG)-PET. Only one data-driven study has been published using a data-driven approach to FDG-PET to find different patterns of hypometabolism in AD 11 , which has been complemented by a recent study in amnestic mild cognitive impairment individuals (aMCI) 12 . Similar spatial subtypes to those found in MRI studies were identified, including a ‘Cortical Predominant’ hypometabolism subtype showing similarity to the Hippocampal Sparing atrophy subtype, and a ‘Limbic Predominant’ subtype. An important aspect to consider for linking the findings in these FDG-PET studies with MRI studies is the use of the term ‘Typical AD’ pattern, which differs across the two modalities. Prominent parieto-temporal hypometabolism reflects the typical AD pattern in FDG-PET, while widespread atrophy in cortex and as the most typical feature hippocampus reflects the typical AD pattern in MRI. 13–15 Comparisons of FDG-PET and MRI across the AD continuum have been performed 16–19 . These comparative studies found correlations across the two modalities, but not a full correspondence of regional neurodegeneration. Similarly, comparison of hypometabolism patterns in aMCI individuals was assessed relative to hippocampal atrophy 12 . However, the combination of FDG-PET and MRI techniques encompassing regions beyond the hippocampus in AD subtypes has not yet been performed. The aim of the current study was to simultaneously evaluate heterogeneity in measures of neurodegeneration (FDG-PET and MRI) to: (a) identify data-driven subtypes of AD, thereby testing the existence of corticolimbic pathways, (b) investigate the overlap and relationship of these corticolimbic subtypes across FDG-PET and MRI, (c) compare data-driven with hypothesis-driven methods of subtyping. Corresponding neurodegeneration patterns, and clinical and demographic information were assessed for each of the subtypes. Understanding the link between atrophy and metabolism could lead to better subtype classification in the context of a biological framework of AD. Materials and methods Participants The cohort consisted of 176 amyloid-beta negative (Aβ-) cognitively normal (CN) and 180 amyloid-positive (Aβ+) Alzheimer’s disease (AD) dementia individuals with both an MRI and an FDG-PET scan, from the Alzheimer’s Disease Neuroimaging Initiative (ADNI), including ADNI1, ADNI2/GO and ADNI3 phases (Table 1 ). The ADNI is a longitudinal multi-centre study aimed at investigating whether neuroimaging methods, together with genetic, clinical, and neuropsychological measures could be used to follow the progression of MCI and AD. The ADNI was launched in 2003 as a public-private partnership, led by principal investigator Michael W. Weiner, MD. For AD patients the inclusion criteria consisted of: Mini-Mental State Examination scores of 20–26, Clinical Dementia Rating scores of 0.5 or 1.0 and meeting the criteria for probable AD. More information of the inclusion and exclusion criteria for the participants can be found on the ADNI website ( https://adni.loni.usc.edu/methods/documents/ ). FDG-PET scans from the initial visit and corresponding MRI scans had a mean of 25 days interval and a maximum of 156 days interval ( Suppl. Figure 1 ). Individuals with AD were included in this study based on Aβ positivity and CNs on Aβ negativity. This was determined in first place by an AV45-PET standard uptake value ratio (SUVR) greater than 1.11 20 . If this data was not available for the individual, we used their Aβ ( 1 – 42 ) values from CSF, which were deemed Aβ positive if lower than 880pg/ml 21 . Table 1 Cohort Demographics and Clinical Characteristics Clinical and Demographic Characteristics Alzheimer’s Disease Cognitively Normal p values N 180 176 ― Women (%) 81 (45%) 84 (48%) 0.682 Age (years) 74 ( 8 ) 75 ( 7 ) 0.867 Disease Duration 3 ( 3 ) ― ― Education (years) 15 ( 3 ) 17 ( 3 ) < 0.001 APOE ε4 (%) 126 (70%) 34 (19%) < 0.001 MMSE 23 ( 2 ) 29 ( 1 ) < 0.001 Global CDR 0.8 ( 3 ) 0.03 (0.2) < 0.001 ADNI-EF -0.4 (0.7) 0.8 (0.5) < 0.001 ADNI-MEM -0.8 (0.3) 0.9 (0.5) < 0.001 ADNI-LAN -0.2 (0.6) 0.8 (0.5) < 0.001 CDF t-tau (pg/ml) 383 (149) 223 ( 70 ) < 0.001 CSF p-tau (pg/ml) 38 ( 16 ) 20 ( 6 ) < 0.001 CSF Aβ (pg/ml) 587 (191) 1204 (333) < 0.001 The values shown in the table are the means with standard deviations in brackets except for number of individuals, women and APOE ε4 carriership for which the percentages are provided. The p-values correspond to the Χ 2 tests conducted for categorical variables and Kruskal-Wallis for continuous variables and were corrected for multiple comparison with the Holm-Šidák method. Abbreviations: AD: Alzheimer’s disease individuals, CN: cognitively normal individuals, APOE ε4: apolipoprotein E ε4 allele, MMSE: Mini Mental State Examination, CDR: The Clinical Dementia Rating Scale, ADNI-EF Alzheimer’s Disease Neuroimaging Initiative executive function composite score, ADNI-MEM: Alzheimer’s Disease Neuroimaging Initiative memory composite score, ADNI-LAN: Alzheimer’s Disease Neuroimaging Initiative language composite score, CSF t-tau: total tau, CSF p-tau: phosphorylated tau, CSF Aβ: amyloid-beta 1-42 peptide. FDG-PET Multiple scanners were used for the FDG-PET scans and followed the appropriate protocols ( http://adni.loni.usc.edu/methods/documents/ ). Dynamic 3D scans made up of six frames of 5 minutes were retrieved 30 to 60 minutes after administration of [ 18 F] FDG. FDG-PET scans were co-registered to individuals’ MRI scans in MNI space using PETSurfer 22,23 and regional standard uptake ratio (SUVR) were extracted without partial volume correction 19,24 . We investigated 82 bilateral regions of interest (ROIs) defined on individual’s MRI including cortical and subcortical structures based on the standard atlases provided by FreeSurfer 25,26 . From these regions, SUVRs were extracted in MNI space. Intensity normalisation was carried out by dividing the regional values by the individual’s global mean 27 . This type of normalisation has been used in prior FDG-PET dementia and subtyping studies 11,28 . MRI T1-weighted magnetisation-prepared rapid gradient-echo (MPRAGE) MRI scans were preprocessed using FreeSurfer (version 6.0.0, https://freesurfer.net ). MRI scans were collected from various sites using multiple scanners (both 1.5 and 3T scanners), following the appropriate protocol ( https://adni.loni.usc.edu/methods/documents/ ). These data was preprocessed in-house through theHiveDB database 29 . Briefly, the preprocessing pipeline involved removal of artefacts, transformation to Talairach space and segmentation of cortical and subcortical regions. From the same 41 regions used for the FDG-PET, cortical and subcortical grey matter volumes were extracted. The raw MRI scans were assessed visually for quality control. Additionally, the estimated total intracranial volumes (ICV) from the FreeSurfer output were plotted against the regional volumes to identify any outliers where the values were grossly under- or over- estimated. The raw images were then checked for these cases to confirm exclusion. Further, scans that were not segmented or registered properly were excluded. In total, 29 individuals (6 CN, 23 AD) were excluded due to: poor PET scan quality, failed quality control, poor segmentation or under/over-estimated ICV volumes. The grey matter volumes were adjusted for head size per region by using a residual approach using the eICV values from the cognitively normal individuals 30,31 as shown below: Volume adj = Volume raw – β(eICV raw – eICV mean ) This adjustment was carried out for each group (CN, AD) in relation to the CNs. Volume raw is the uncorrected volume for the brain region of interest. The β value and mean ICV, eICV mean , are calculated by running a linear regression model per region of interest in relation to the ICV using data from the CNs. CSF biomarkers In addition to CSF Aβ ( 1 – 42 ) values, we assessed CSF measures of tau phosphorylated at threonine 181 (p-tau) and total tau (t-tau) taken from an automated Elecsys cobas e 601 analyser. Neuropsychological Testing We used the Mini-Mental State Examination (MMSE) to assess global cognition and composite scores for executive function (ADNI-EF), memory (ADNI-MEM), and language (ADNI-LAN) 32–34 to assess specific cognitive domains. For MMSE, lower scores indicate higher global cognitive impairment and lower scores for each of the composite scores indicate greater impairment. Data-driven Subtyping: Hierarchical Clustering Analysis To classify individuals into subtypes in FDG-PET and MRI, unsupervised random forest hierarchical clustering was performed ( Suppl. Figures 2, 3 ) to identify the linear and non-linear relationships from the regional values. This method has previously been used to investigate heterogeneity within neurodegenerative diseases using grey matter volumes 10,35,36 . In the current study, an identical clustering procedure was applied to regional glucose metabolism values and grey matter volumes using the same 41 bilateral cortical and subcortical ROIs. Two separate clustering models were performed, one using glucose metabolism values and another with grey matter volumes. First, a distance matrix was calculated based on the regional values using a random forest algorithm, which provides information on the similarities and dissimilarities across the brain regions. The optimisation of hyperparameters was performed by selecting the values with the lowest out-of-bag errors 37 for: number of variables randomly sampled at each split ( mtry ) and minimum size of the terminal nodes ( nodesize ). The number of trees was set to 20,000 for all models. Additionally, the stability of the chosen random forest model was tested by running the random forest algorithm 100 times. The differences between the chosen model and the simulated models were calculated ( Suppl. Figures 4, 5 ). The distance matrix was then reduced to three dimensions using classical multi-dimensional scaling to simplify the interpretation of the most important features that distinguish the clusters from each other. The first three dimensions from the multi-dimensional scaling were used for clustering as they explained the greatest differences between the groups, identified by plotting the eigenvalues of the dimensions. Agglomerative hierarchical clustering with average linkage was then run using the reduced matrix to identify clusters. The output of the clustering is a dendrogram and to group the individuals into subtypes the number of clusters needs to be chosen. This number was derived by using various cluster validation indices, namely the Calinski-Harabasz, Davies-Bouldin, Dunn and Silhouette indices from NbClust and fpc libraries in R. The Calinski-Harabasz index is calculated by comparing the between cluster variance with the within cluster variance. The Davies-Bouldin index evaluates cluster compactness and distinctness by comparing within-cluster distances to between-cluster distances. Similarly, the Dunn index assess cluster compactness and separation. The Silhouette index is the measure of how well the objects fit into its allocated cluster. Collectively, these four indices capture a well-rounded assessment for choosing the number of clusters for our models. Inter-modality Comparison of Subtypes Comparisons between FDG-PET and MRI subtypes were investigated by: 1) subtyping in one modality and mapping the corresponding atrophy/hypometabolism patterns in the other modality, 2) frequency of the derived FDG-PET and MRI subtypes, 3) crossover between individual subtype allocations. Regional atrophy patterns of the FDG-PET subtypes and regional hypometabolism patterns of the MRI subtypes were assessed using w-scores. W-scores 17 were calculated as z-scores by subtracting the mean and dividing by the standard deviation from the cognitively normal data and adjusted for covariates: age, sex, education and APOE ε4 allele carriership. Furthermore, the frequencies of the individuals belonging to a given subtype in both modalities, e.g., an individual being classified as Cortical Predominant in FDG-PET clustering and Cortical Predominant in MRI clustering, was calculated. The different modality-specific subtypes were compared in terms of the defining topographical patterns, which enabled us to visualise the similarities and differences across the subtypes and whether individuals had similar neurodegeneration patterns in the two modalities. To test whether the modality-specific subtype classifications overlapped at an individual level, the individuals’ subtype allocations were compared in an alluvial plot. Additionally, to numerically demonstrate the overlap, a ratio for each FDG-PET and MRI subtype pairing was calculated using the total number of cases for each FDG-PET subtype. For instance, the ratio was determined by dividing the number of Cortical Predominant MRI cases by the total number of Cortical Predominant FDG-PET cases. Data-driven versus Hypothesis-driven Subtypes This analysis was performed to compare our data-driven method with a previously described hypothesis-driven method designed to identify neuropathologically defined corticolimbic subtypes 5–7 . The aim was to investigate whether there would be an overlap in the subtype categorisation from these two methods. Given that hypothesis-driven subtyping has been applied to MRI data previously, this analysis additionally aimed to determine whether such a subtyping could be adapted using FDG-PET data, mirroring techniques previously applied to tau PET 38 . Prior hypothesis-driven studies have used the hippocampus-to-cortex ratio to group AD individuals into one of three subtypes (‘Limbic Predominant’, ‘Typical AD’, 'Hippocampal Sparing’) using a two-step process using the 75th and 25th percentiles and median hippocampal and cortical values 5–7 . Hypothesis-driven subtype labels in the current study were ‘Limbic Predominant’ (low hippocampus-to-cortex ratio) and ‘Cortical Predominant’ (high hippocampus-to-cortex ratio). The hippocampus-to-cortex ratio captures whether a subtype shows either a limbic or a cortical pattern and was conceptualized as the ‘typicality’ axis in a recent framework explaining the topography of AD subtypes 2 . A second axis of the subtypes was termed ‘severity’, which refers whether a subtype shows atypically low or high atrophy. Hence, in this analysis, we further compared the data-driven and hypothesis-driven subtyping in the context of the published framework. For MRI, the measure for ‘typicality’ was the hippocampus-to-cortex ratio in volume measures and the measure for ‘severity’ was the total grey matter volume. For FDG-PET, ‘typicality’ was hippocampus-to-cortex ratio in glucose metabolism and ‘severity’ was the total average cortical SUVR. All subtype classifications were then plotted along these two axes. These are orthogonal to each other, explaining the differing patterns in the five common patterns of atrophy (hippocampal-sparing, limbic-predominant, typical AD, minimal and diffuse). To explore this, Pearson correlations were calculated for measures of ‘typicality’ and ‘severity’ within and between the two imaging modalities. Typicality and severity measures were further used to generate a conceptual figure for the two modalities’ subtypes along these two axes. Values for each subtype were rescaled to range from values 0 to 1 using min-max normalisation. These normalised values where then averaged and plotted. Statistical Analysis RStudio and R version 4.2.0 were used for statistical analyses. W-scores were plotted to show the differences in the regional SUVRs and grey matter volumes in the subtypes compared to the cognitively normal (CN) individuals 17,39 . For FDG-PET, the w-scores were calculated using SUVRs scaled by the pons as the reference region. The pons was chosen because it has been compared with other reference regions across ageing and Alzheimer’s disease studies using FDG-PET and it has been shown to work well with both PVC and non-PVC data 40 . These values were averaged across the AD individuals respectively and reversed so that the higher w-scores correspond to greater neurodegeneration in the AD group. Brain maps were created using the ggseg library in R 41 . The demographic and clinical variables of the subtypes were compared using χ² and Kruskal-Wallis tests that were adjusted for multiple comparisons with the Holm-Šidák method. Significance was deemed if the p values were < 0.05 where the null hypothesis can be rejected when \(p<\text{α }\) with α = 0.001, this stringent alpha value was set to avoid Type 1 family-wise errors. Comparisons were made between the subtypes and the cognitively normal individuals, as well as pairwise comparisons between the groups. Results Cohort characteristics The basic clinical and demographic information of the cohort is reported in Table 1 . As expected, The AD group significantly differed from the CN group in terms of percentage of APOE ε4 carriers, cognitive measures, and CSF markers. There were no significant differences between our AD and CN groups in terms of sex and age. FDG-PET Clustering The overall AD group’s hypometabolism pattern showed neurodegeneration in posterior cingulate and frontal cortical regions and deeper structures such as the hippocampus (Fig. 1 a ) . The clustering model resulted in two main distinguishing patterns, neurodegeneration in cortical versus limbic pathways (Fig. 1 c). The Calinski-Harabasz index peaked at five clusters, for the Davies-Bouldin index was lower at five, the Dunn index plateaus at five and the Silhouette index was highest at 3–5 clusters ( Suppl. Figure 6 ). Therefore, we chose five clusters as the optimal solution for the FDG-PET model. The FDG-PET model was split into five distinct hypometabolism-based subtypes. Three subtypes showed cortical-predominant hypometabolism of differing severity and spatial distribution. The Cortical Predominant posterior subtype had cortical hypometabolism mainly in the posterior regions (9%), whereas the Cortical Predominant and Cortical Predominant + subtypes showed more widespread cortical hypometabolism (32% and 11% respectively). The Cortical Predominant + subtype had greater hypometabolism than the other two cortical predominant subtypes. Although all the subtypes showed some hypometabolism in the hippocampus, these subtypes had proportionally less involvement of this region compared to the cortical areas. Two subtypes displayed limbic hypometabolism, focal to the medial temporal and deeper structures (amygdala, hippocampus). Here, a principal Limbic Predominant subtype (36%) could be distinguished from and a Limbic Predominant frontal subtype (13%). In the clustering dendrogram (Fig. 1 ), the Limbic Predominant frontal cluster originates from its own branch, whereas the Limbic Predominant cluster comes from the same branch as the Cortical Predominant posterior cluster. Thus, the clustering separates these subtypes by a frontal versus a posterior hypometabolism pattern. By contrast, the Cortical Predominant and Cortical Predominant + clusters are separated on the opposite side of the dendrogram by the severity of their cortical hypometabolic patterns. For inter-modality comparisons, the corresponding atrophy patterns in these subtypes were plotted (Fig. 1 e ) . Based on visual comparison of the w-scores, the brain maps were topographically similar across FDG-PET subtypes, showing AD-typical atrophy in medial temporal, hippocampal, and some frontal areas. However, the atrophy pattern of the Cortical Predominant subtype was not as widespread in the cortical regions compared to the hypometabolism. Additional maps using PVC SUVRs from PETSurfer were plotted ( Suppl. Figure 8 ). These brain maps did not differ greatly from our maps in Fig. 1 topographically but did result in lower w-scores. Regarding demographic and clinical differences among the FDG-PET AD subtypes (Table 2 ), the Cortical Predominant + subtype was the youngest (67.5 years), had the earliest age at onset (64.7 years), more pronounced language impairment, and lowest executive function scores. This subtype also had the highest grey matter volume-based and SUVR-based hippocampus-to-cortex ratios. The other two cortical subtypes (Cortical Predominant, Cortical Predominant posterior) had a higher SUVR-based hippocampus-to-cortex ratio than the limbic subtypes. Cortical Predominant posterior also had a high hippocampus-to-cortex ratio using grey matter volumes compared to the limbic subtypes. Among the limbic subtypes, Limbic Predominant frontal was the oldest, had latest age at onset and worst language scores. There were no significant differences between the subtypes for the other variables: sex, disease duration, years of education, APOE ε4 carriers, MMSE, CDR, cognitive measures of memory, and CSF biomarkers. Although not statistically significant, two of the cortical subtypes, Cortical Predominant and Cortical Predominant+, had lower percentage of APOE ε4 carriers and Limbic Predominant frontal the highest percentage of APOE ε4 carriers. Table 2 Demographic and clinical characteristics of the FDG-PET AD subtypes Demographic & Clinical Characteristics Cortical Predominant Cortical Predominant+ Limbic Predominant Limbic Predominant frontal Cortical Predominant posterior Cognitively Normal p values N (%) 57 (32%) 20 (11%) 64 (36%) 23 (13%) 16 (9%) 176 ― Women (%) 22 (39%) 10 (50%) 32 (50%) 10 (43%) 7 (44%) 84 (48%) 0.768 Age (years) 75 (7.9) 67 (8.8) 74 (6.8) 78 (6.6) 73 (8.8) 75 (7.1) < 0.001 Disease Duration (years) 2.5 (2.7) 2.7 (2.1) 2.9 (2.5) 3.1 (2.9) 1.8 (2.1) - 0.274 Age at Onset (years) 72 (8.1) 65 (8.3) 71 ( 7 ) 75 (7.4) 71 (8.8) - < 0.001 ab Education (years) 15 (2.5) 16 (2.8) 16 (2.9) 15 (3.5) 15 (2.2) 17 (2.6) 0.361 APOE ɛ4 (%) 35 (61%) 11 (55%) 48 (75%) 20 (87%) 12 (75%) 34 (19%) 0.083 MMSE 23 (2.1) 22 (2.6) 24 ( 2 ) 23 (1.9) 23 (1.7) 29 (1.4) 0.167 Global CDR 0.77 (0.25) 0.88 (0.22) 0.73 (0.25) 0.77 (0.25) 0.89 (0.4) 0.025 (0.16) 0.225 ADNI-EF -0.43 (0.62) -0.99 (0.73) -0.19 (0.61) -0.6 (0.51) -0.58 (0.7) 0.8 (0.47) < 0.001 bc ADNI-MEM -0.85 (0.32) -0.89 (0.32) -0.66 (0.34) -0.93 (0.31) -0.8 (0.4) 0.89 (0.52) 0.007 ADNI-LAN -0.24 (0.53) -0.43 (0.52) 0.041 (0.54) -0.5 (0.61) -0.31 (0.56) 0.82 (0.52) < 0.001 bce CSF t-tau (pg/ml) 378 (145) 407 (143) 415 (167) 335 (98) 335 (152) 223 ( 70 ) 0.958 CSF p-tau (pg/ml) 38 ( 15 ) 41 ( 16 ) 43 ( 19 ) 33 ( 10 ) 33 ( 15 ) 20 (6.3) 0.961 CSF Aβ (pg/ml) 585 (248) 588 (173) 588 (169) 596 (150) 578 (139) 1204 (333) 0.537 SUVR-based Hippocampus-to-cortex ratio 0.35 (0.04) 0.38 (0.047) 0.3 (0.034) 0.32 (0.032) 0.33 (0.032) 0.32 (0.029) < 0.001 Grey Matter Volume-based Hippocampus-to-cortex ratio 0.16 (0.021) 0.19 (0.029) 0.16 (0.02) 0.16 (0.023) 0.17 (0.025) 0.18 (0.021) < 0.001 fg Total Grey Matter Volume (mm³) 558154 (33242) 544820 (35401) 554040 (38026) 537830 (47861) 539736 (36515) 580718 (40407) 0.129 Total Average Cortical Uptake (SUVR) 1.5 (0.16) 1.5 (0.12) 1.6 (0.18) 1.6 (0.11) 1.6 (0.2) 1.7 (0.19) 0.021 The values shown in the table are the means with standard deviations in brackets except for number of individuals, women and APOE ε4 for which the percentages are provided. The reported p-values correspond to Χ2 tests which were used for categorical variables and Kruskal-Wallis for continuous variables and were corrected for multiple comparison with the Holm-Šidák method. Footnotes indicate cases where p values were significant in the post hoc pairwise comparisons across AD subtypes, p < 0.05. The CN group data is displayed for reference. Abbreviations: CP: Cortical Predominant, CP+: Cortical Predominant+, LP: Limbic Predominant, LP fr.: Limbic Predominant frontal, CP post.: Cortical Predominant posterior, CN: cognitively normal individuals, APOE ε4: apolipoprotein E ε4 allele, MMSE: Mini Mental State Examination, CDR: The Clinical Dementia Rating Scale, ADNI-EF Alzheimer’s Disease Neuroimaging Initiative executive function composite score, ADNI-MEM: Alzheimer’s Disease Neuroimaging Initiative memory composite score, ADNI-LAN: Alzheimer’s Disease Neuroimaging Initiative language composite score, CSF t-tau: total tau, CSF p-tau: phosphorylated tau, CSF Aβ: amyloid-beta 1-42 peptide. Cortical Predominant+ < Limbic Predominant, p < 0.05 Cortical Predominant+ < Cortical Predominant, p < 0.05 Limbic Predominant < Limbic Predominant frontal, p < 0.05 Limbic Predominant frontal < Limbic Predominant, p < 0.05 Limbic Predominant < Cortical Predominant+, p < 0.05 Limbic Predominant frontal < Cortical Predominant+, p < 0.05 Limbic Predominant < Cortical Predominant, p < 0.05 Limbic Predominant < Cortical Predominant posterior, p < 0.05 Cortical Predominant posterior < Cortical Predominant+, p < 0.05 Cortical Predominant < Cortical Predominant+, p < 0.05 MRI Clustering The overall AD group showed atrophy in the expected medial temporal regions such as hippocampus and amygdala (Fig. 1 b), which will be referred to as the a ‘typical’ AD pattern for MRI. Similar to FDG-PET subtypes, clustering revealed a distinction between either a limbic or a cortical pathway in MRI (Fig. 1 d). The Calinski-Harabasz index peaked at three but was still high at five clusters, the Davies-Bouldin index was lower at five, the Dunn index was high for five albeit plateaued at six before a sharp increase after that and the Silhouette index was highest at three to six clusters ( Suppl. Figure 7 ). Therefore, we chose five clusters as the optimal solution for the MRI model based on these results and considering prior work identifying five biological subtypes. Another reason for choosing a higher cluster solution was based on the lack of sensitivity for finding atypical patterns when implementing a three- and four- cluster solutions. The MRI clustering model was split into five atrophy-based subtypes. In contrast to the FDG-PET subtypes which were limited to cortical and limbic subtypes, the MRI subtypes showed additional ‘minimal’ versus ‘diffuse’ atrophy patterns. Similar to the FDG-PET Cortical Predominant subtypes, a Cortical Predominant MRI subtype (19%) showed greater cortical atrophy relative to the hippocampus. The Limbic Predominant subtype (27%) had the opposite pattern with greater atrophy in the hippocampus relative to the cortex. The Minimal subtype (19%) had some atrophy in the hippocampus and amygdala, but very little atrophy compared to cognitively normal individuals in the cortical regions. There were two diffuse atrophy subtypes, one with greater overall atrophy, Diffuse+ (6%), and one with similarly diffuse but less severe atrophy (28%). Based on the inter-modality comparison, the atrophy-based subtypes displayed hypometabolism of differing severity in temporo-parietal and lateral temporal regions often described to be the ‘typical AD’ pattern in FDG-PET scans (Fig. 1 f). Compared to the corresponding atrophy maps of the FDG-PET subtypes (Fig. 1 e), the hypometabolism maps (Fig. 1 f, Suppl. Figure 9 ) were more topographically similar to the MRI subtypes when based on visual comparison of w-scores. These corresponding maps showed both more pronounced (higher w-scores) and more widespread hypometabolism in the Minimal and Cortical Predominant subtypes compared to their atrophy maps (Fig. 1 f, Suppl. Figure 9 ). Regarding demographic and clinical differences, among the MRI AD subtypes (Table 3 ), the Diffuse subtype had the lowest executive function scores compared to the Minimal and Limbic Predominant subtypes. Diffuse, Diffuse+, and Cortical Predominant had significantly worse executive function scores compared to the Minimal subtype. Minimal and Limbic Predominant subtypes had significantly lower SUVR-based hippocampus-to-cortex ratios to Cortical Predominant. Significant differences were also found in the grey matter volume-based hippocampus-to-cortex ratios: Cortical Predominant had the highest hippocampus-to-cortex ratio. There were no significant differences between the subtypes for the other variables: sex, age, disease duration, age at onset, years of education, APOE ε4 carriers, MMSE, CDR, cognitive measure of memory and language, and CSF biomarkers. Despite not showing a significant difference, Diffuse + had the highest proportion of APOE ε4 carriers (90%) and was the oldest group (79.3 years). Table 3: Demographic and clinical characteristics of the MRI AD subtypes. Demographic & Clinical Characteristics Cortical Predominant Diffuse Limbic Predominant Diffuse+ Minimal Cognitively Normal p values N (%) 35 (19%) 51 (28%) 49 (27%) 10 (6%) 35 (19%) 176 ― Women (%) 10 (29%) 30 (59%) 22 (45%) 6 (60%) 13 (37%) 84 (48%) 0.049 Age (years) 70 (9.5) 74 (8.4) 76 (6.3) 79 (4.9) 73 (6.7) 75 (7.1) 0.004 Disease Duration (years) 2.2 (2.3) 3.4 (2.5) 2.5 (2.7) 4.4 (3.2) 1.9 (2.1) ― 0.002 Age at Onset (years) 68 (9.4) 71 (8.6) 74 (6.7) 75 (7.3) 71 (6.7) ― 0.021 Education (years) 15 (3) 15 (2.9) 16 (2.6) 16 (2.7) 16 (2.6) 17 (2.6) 0.197 APOE ɛ4 (%) 22 (63%) 38 (75%) 31 (63%) 9 (90%) 26 (74%) 34 (19%) 0.329 MMSE 23 (2.4) 23 (2) 23 (2.1) 23 (2.1) 24 (2.1) 29 (1.4) 0.767 Global CDR 0.79 (0.25) 0.83 (0.24) 0.74 (0.32) 0.89 (0.22) 0.72 (0.25) 0.025 (0.16) 0.121 ADNI-EF -0.6 (0.62) -0.72 (0.6) -0.32 (0.5) -0.63 (0.72) -0.0096 (0.74) 0.8 (0.47) <0.001[a][b][c] ADNI-MEM -0.78 (0.32) -0.91 (0.33) -0.78 (0.29) -0.85 (0.43) -0.65 (0.4) 0.89 (0.52) 0.059 ADNI-LAN -0.22 (0.41) -0.4 (0.52) -0.22 (0.55) -0.13 (0.67) 0.11 (0.68) 0.82 (0.52) 0.007 CSF t-tau (pg/ml) 390 (164) 398 (145) 380 (141) 280 (77) 377 (161) 223 (70) 0.065 CSF p-tau (pg/ml) 39 (17) 40 (16) 37 (14) 26 (7.4) 39 (18) 20 (6.3) 0.041 CSF Aβ (pg/ml) 564 (168) 575 (159) 661 (245) 536 (128) 551 (181) 1204 (333) 0.425 SUVR-based Hippocampus-to-cortex ratio 0.35 (0.047) 0.34 (0.048) 0.32 (0.038) 0.34 (0.053) 0.31 (0.043) 0.32 (0.029) <0.001[d][e] Grey Matter Volume-based Hippocampus-to-cortex ratio 0.18 (0.021) 0.17 (0.026) 0.15 (0.018) 0.16 (0.027) 0.15 (0.022) 0.18 (0.021) <0.001 p [f][g] Total Grey Matter Volume (mm³) 561400 (17261) 518539 (12671) 552539 (15299) 478411 (11842) 606360 (16474) 580718 (40407) <0.001 m s [h] Total Average Cortical Uptake (SUVR) 1.6 (0.16) 1.5 (0.14) 1.6 (0.15) 1.4 (0.081) 1.7 (0.21) 1.7 (0.19) <0.001 a [i][j][k] The values shown in the table are the means with standard deviations in brackets except for number of individuals, women and APOE ε4 for which the percentages are provided. Χ 2 tests were used for categorical variables and Kruskal-Wallis for continuous variables and were corrected for multiple comparison with the Holm-Šidák method. Footnotes indicate cases where p values were significant in the post hoc pairwise comparisons across AD subtypes, p < 0.05. The CN group data is displayed for reference. Abbreviations: CP: Cortical Predominant, LP: Limbic Predominant, Min: Minimal atrophy, Dif: Diffuse, Dif+: Diffuse+, CN: cognitively normal individuals, APOE ε4: apolipoprotein E ε4 allele, MMSE: Mini Mental State Examination, CDR: The Clinical Dementia Rating Scale, ADNI-EF Alzheimer’s Disease Neuroimaging Initiative executive function composite score, ADNI-MEM: Alzheimer’s Disease Neuroimaging Initiative memory composite score, ADNI-LAN: Alzheimer’s Disease Neuroimaging Initiative language composite score, CSF t-tau: total tau, CSF p-tau: phosphorylated tau, CSF Aβ: amyloid-beta 1-42 peptide. Diffuse < Minimal, p < 0.05 Cortical Predominant < Minimal, p < 0.05 Diffuse < Limbic Predominant, p < 0.05 Minimal < Cortical Predominant, p < 0.05 Limbic Predominant < Cortical Predominant, p < 0.05 Diffuse < Cortical Predominant, p < 0.05 Minimal < Diffuse, p < 0.05 Limbic Predominant < Minimal, p < 0.05 Diffuse+ < Minimal, p < 0.05 Diffuse+ < Cortical Predominant, p < 0.05 Diffuse+ < Limbic Predominant, p < 0.05 Individual-level Subtype Allocations To assess the consistency between the two modalities in subtype assignments, the subtype categorizations for individuals were compared across both FDG-PET and MRI (Fig. 2 ). We propose that the cortical subtypes and limbic subtypes are most similar between the FDG-PET and MRI subtypes. Namely, the FDG-PET Cortical Predominant, Cortical Predominant posterior and Cortical Predominant + subtypes are equivalent with MRI Cortical Predominant, Diffuse or Diffuse + patterns topographically. Whereas FDG-PET Limbic Predominant and Limbic Predominant frontal are equivalent to MRI Limbic Predominant. As a Minimal pattern was only found in MRI, we do not think that this subtype has an equivalent in FDG-PET. The agreement between the FDG-PET and MRI subtype allocations was low as this was less than 50%. Although the compared subtypes showed similar topographies of neurodegeneration (i.e., cortical/limbic predominant hypometabolism and atrophy, respectively) they did not match at the individual-level. All possible combinations of allocated FDG-PET and MRI subtypes of varying percentages were found (Fig. 2 a, b). AD individuals classified into the FDG-PET Cortical Predominant subtype matched most with the MRI Limbic Predominant (33.3%) and MRI Cortical Predominant (24.6%) subtypes. By contrast, individuals classified as FDG-PET Cortical Predominant + matched best with MRI Cortical Predominant (35%) and Diffuse (35%). FDG-PET Limbic Predominant best matched with three MRI subtypes: Limbic Predominant (28.1%), Minimal (26.6%) and Diffuse (25%). FDG-PET Limbic Predominant frontal best matched with MRI Diffuse (52.2%). FDG-PET Cortical Predominant posterior matched best with MRI Limbic Predominant (37.5%). Data-driven versus Hypothesis-driven Subtyping Clustering-based and prior hypothesis-based subtypes were compared within the framework of typicality and severity 2 in both FDG-PET and MRI (Fig. 3 ). Within each modality, data-driven and hypothesis-driven subtypes overlapped with each other reasonably well for most subtypes (Fig. 3 a, b). The agreement between MRI data-driven and MRI hypothesis-driven limbic predominant subtypes (55.6%) was better than that between cortical predominant subtypes (30%) (Fig. 3 b). Contrarily, the agreement between FDG-PET data-driven and FDG-PET hypothesis-driven cortical predominant subtypes (90%) was better than that between limbic predominant subtypes (82%) (Fig. 3 a ) . Association between typicality and severity by modality differed. Correlation between typicality and severity (Table 4 ) was significant in FDG-PET (r 2 = 0.25), but not in MRI. The severity measures (r 2 = 0.15) in FDG-PET and MRI were more strongly associated with each other than typicality measures (r 2 = 0.02). Additionally, FDG-PET subtypes are more separable across the typicality axis than MRI, this is evident when comparing averaged normalised values of typicality and severity for each subtype (Fig. 4 ). Whereas for MRI subtypes, there was a clearer split of the along the severity axis. Table 4 Correlations between Typicality & Severity in FDG-PET and MRI. Model R² P value FDG-PET Typicality and FDG-PET Severity (Fig. 3 a) 0.25 < 0.001** MRI Typicality and MRI Severity (Fig. 3 b) 0.0011 0.53 FDG-PET Severity and MRI Severity 0.15 < 0.001** FDG-PET Typicality and MRI Typicality 0.02 0.0073* Pearson correlations between measures of ‘typicality’ and ‘severity’ in both modalities. * P < 0.05. ** P < 0.001. Discussion This study investigated cross-modality Alzheimer’s disease subtypes by applying data-driven subtyping models to regional FDG-PET and MRI data. To our knowledge, this is the first study to implement identical data-driven models to concurrent FDG-PET and MRI data to identify and compare modality-specific neurodegeneration-based AD subtypes. Despite FDG-PET and MRI both being interchangeable measures of neurodegeneration within the ATN framework 45,46 , our findings show that the respective neurodegeneration subtypes differed across modalities and within individuals. Using the same data-driven model, cortical and limbic AD subtypes were independently identified in both FDG-PET and MRI methodology. At a group-level, the expected pattern of temporo-parietal hypometabolism 42–46 was found in the AD group. However, FDG-PET subtypes showed distinct patterns of hypometabolism: Cortical Predominant, Cortical Predominant+, Cortical Predominant posterior, Limbic Predominant and Limbic Predominant frontal. Our findings resemble findings of the only previous study addressing FDG subtypes in AD 11 . In terms of topography, Cortical Predominant, Cortical Predominant posterior and Limbic Predominant subtypes in our study are closest to the Typical subtype (hypometabolism in cortical and limbic regions) identified by Levin et al. although with differing clinical characteristics 11 . The differences in findings could possibly be explained by methodological (clustering method, regions used for clustering, etc.) and sample differences. Two subtypes that were identified included the Cortical Predominant+ (younger, fewer APOE ε4 carriers, executive function impairment) and Limbic Predominant frontal (older age, more APOE ε4 carriers), which align well with the Cortical Predominant and Limbic Predominant subtypes, respectively, identified by Levin et al 11 . Together, these two studies indicate the clear presence of distinct cortical- and limbic-predominant profiles of hypometabolism in AD. The MRI pattern of the whole AD group showed the expected pattern of medial temporal and hippocampal atrophy 47–50 . However, the MRI AD subtypes showed five distinct patterns: Cortical Predominant, Limbic Predominant, Minimal atrophy, Diffuse, Diffuse+. Our data-driven MRI subtypes had similar percentages and clinical presentation to what has been found previously 2,9,35,51,52 . In accordance with previous studies, Limbic Predominant had focal limbic atrophy and later age of onset. Cortical Predominant resembles the ‘Hippocampal Sparing’ MRI subtype previously described, with greater atrophy in the cortical relative to the limbic regions, as well as greater executive function impairment, higher proportion of men, and younger age. The two Diffuse subtypes in this study resemble Typical and Diffuse subtypes identified previously (widespread atrophy, older, worse memory scores). At the other end of the severity dimension, Minimal atrophy subtype resembled prior Minimal subtypes topographically and clinically (shortest disease duration, highest MMSE). We did not observe an overlap in terms of topography nor demographics, despite identifying cortical and limbic subtypes in both FDG-PET and MRI. As the ‘typical’ AD pattern for each modality differs, this can be seen in the subtypes’ modality-specific neurodegeneration. Key regions show neurodegeneration at varying levels of severity, such as posterior cingulate cortex in FDG-PET ( Fig. 1c ) and medial temporal lobe and hippocampus for MRI ( Fig. 1d ). FDG-PET subtypes show a clearer cortical pathway compared to the MRI subtypes, which are more susceptible to more limbic neurodegeneration. These results are in line with the common AD patterns in these two modalities: neocortical hypometabolism in FDG-PET 43,45,53 and medial temporal atrophy in MRI 47,49 . While the regional atrophy does not mimic hypometabolism in FDG-PET subtypes ( Fig. 1c, e ), regional hypometabolism mirrors atrophy in MRI subtypes more closely ( Fig. 1d, f ). For example, cortical predominant FDG-PET subtypes with limited hippocampal hypometabolism were found to show considerable limbic (hippocampal) atrophy. This finding is congruent with the well-established evidence that higher cortical hypometabolism is closely associated with higher hippocampal atrophy 17,19,54 . Our study goes a step further to demonstrate that in contrast, a cortical-predominant MRI subtype with limited hippocampal atrophy also shows cortical-predominant hypometabolism with limited hippocampal hypometabolism. A previous study assessed FDG-PET severity within AD typical cortical regions in hypothesis-driven MRI subtypes using the so-called ‘Hypometabolic Convergence Index’ 7 . Hippocampal Sparing had highest values of this index compared to Limbic Predominant and Typical AD subtypes, highlighting the overlap in cortical hypometabolism and atrophy. Although our FDG-PET subtypes had similar patterns of cortical hypometabolism and atrophy, this was not the case when assessing deeper structures ( Fig 1 e, f) . Thus, FDG-PET and MRI do not necessarily share a bidirectional relationship in capturing hippocampal neurodegeneration. Interestingly when assessing how individuals were subtyped in both modalities, a remarkably low classification correspondence between equivalent subtypes (i.e., cortical/limbic predominant) in the different modalities was found ( Fig. 2a ). Individuals classified as Cortical hypometabolism subtypes were not always classified as cortical predominant atrophy subtypes and even matched with Limbic Predominant atrophy patterns. Limbic hypometabolism subtypes agreed with different levels of atrophy severity from Minimal to Diffuse. The ‘disconnection hypothesis’ has been proposed in AD to explain concurrent retrosplenial cortex hypometabolism and medial temporal atrophy 15,19,55 . It has been proposed that local and/or distant atrophy result in downstream hypometabolism in the parietal and retrosplenial cortex 19,56 . In our corresponding neurodegeneration maps ( Fig. 1e, f ), association between levels of isthmus cingulate (retrosplenial cortex) hypometabolism relative to medial temporal lobe atrophy can be appreciated. FDG-PET subtypes; Cortical Predominant and Cortical Predominant posterior had greatest isthmus cingulate hypometabolism and consequently displayed greater medial temporal atrophy than the other subtypes ( Fig. 1c, e ). Conversely, Cortical Predominant+ hypometabolism subtype had greatest atrophy in cingulate/precuneus and least in hippocampus compared to the other hypometabolism subtypes. This is in concordance with prior findings in early onset AD patients who typically show more posterior cortical atrophy compared to atrophy in the hippocampus 53,57 . MRI subtypes’ corresponding hypometabolism maps all showed isthmus cingulate involvement even in the Minimal atrophy subtype ( Fig. 1d, f ). This finding resembles prior findings of a minimal hippocampal and posterior cingulate atrophy AD pattern that showed significant hypometabolism 53 . This could be indicative of FDG-PET being an earlier marker of neurodegeneration than MRI 58–61 . Moreover, neuropathological data and its association with specific neurodegeneration patterns could provide the link between FDG-PET and MRI. Potential copathologies are often present in combination with AD 62 therefore more analysis is needed especially within subtypes 52 and across different imaging modalities 63 . Different pathologies have been linked to specific patterns of atrophy and hypometabolism 64–66 . These copathologies could also explain the divergence between atrophy and metabolism patterns. Measures such as the hippocampus-to-cortex ratio based on neuropathological data 5 have been applied successfully using grey matter volumes from MRI 6,7 . In this study, we compared such measures with findings from our data-driven approach 2 . Our findings showed that there was agreement between hypothesis-driven and data-driven approaches. Interestingly, FDG-PET Cortical Predominant+ also had higher hippocampus-to-cortex ratio when using grey matter volumes similar to the finding in Levin et al., 2021 11 . The data-driven subtypes additionally capture the second axis of ‘severity’, as Minimal and Diffuse subtypes were found in the MRI. It can also be argued that our FDG-PET subtypes show different levels of ‘severity’ as Limbic Predominant frontal and Cortical Predominant+ show greater overall hypometabolism than the other limbic and cortical subtypes. The contrast between the data-driven and hypothesis-based methods highlights the issue with harmonisation of methods for subtyping in AD 4 . In general, hypothesis-driven hippocampus-to-cortex ratios work well using FDG-PET SUVRs and match well with data-driven subtypes. However, as this is the first study to test these two methods using FDG-PET SUVRs this would need to be further validated. Typicality and severity were correlated in FDG-PET indicating that cortical subtypes are the more severe compared to the limbic subtypes. Whereas, typicality and severity were not correlated in MRI, showing that the link is different across the two modalities ( Table 4 ). These findings could indicate that there are more complex mechanisms at play, such as copathologies 64,66 as neither of these measures are specific to AD. Conceptually, FDG-PET and MRI subtypes found in our study can be defined along the ‘typicality’ and ‘severity' axes ( Fig. 4 ). Typicality splits the two modality-specific subtypes into either a limbic or cortical pattern. Severity splits the extremes found in the MRI subtypes of limited to widespread atrophy. In contrast, there was a greater split along the typicality axis for FDG-PET subtypes between cortical and limbic hypometabolism. In this study, we propose that subtypes do not always lie orthogonal along these two axes and are often a combination of both. Namely, we found that FDG-PET Limbic Predominant subtype is left of the centre towards limbic along the typicality axis but is positioned lower on the severity axis compared to Limbic Predominant frontal. The same applies to the Cortical Predominant+ and Cortical Predominant subtypes which have widespread hypometabolism. The severity for the current FDG-PET subtypes is backed by frontal hypometabolism being indicative of a later stage of AD 67 . The complexity of the relationship of the two modalities is reflected in this figure and underlying mismatches between the two modalities have been highlighted throughout this study. This study has various limitations, one is its cross-sectional design. To concretise the current findings, longitudinal studies are needed to be able to label these patterns as subtypes. Longitudinal clustering has been performed in AD to investigate heterogeneity of topographical differences in MRI 9,68 . In addition, tau- and amyloid-PET studies have used cross-sectional data with a probabilistic model for predicting disease subtype and stage 69,70 . Future research should explore the progression of glucose metabolism across these subtypes in relation to other imaging modalities. Additionally, testing the models and classification of the subtypes in clinical cohorts for external validity is important. Another limitation is that the individuals in this study likely have mixed pathologies that cannot be detected through neuroimaging alone. Utilising neuropathological data in combination with in vivo data should be a focus in future studies. Ultimately, the current study identified data-driven cortical and limbic AD subtypes from FDG-PET and MRI scans. These data-driven subtypes overlapped well with hypothesis-driven methods, validating our findings. Although, the main finding was that structure does not always reflect function when assessing corresponding patterns of neurodegeneration in these subtypes. Cortical and limbic subtypes did not overlap in the two modalities in terms of individual subtype allocation. These subtypes lie along ‘severity’ and ‘typicality’ axes distinctly across modalities as shown in our conceptual figure ( Fig. 4 ). Copathologies may contribute to the divergence in FDG-PET and MRI subtype patterns. In conclusion, the current findings highlight the need for a multimodal perspective for understanding the complex biological AD mechanisms. Declarations Data availability The data used in this study are available from the corresponding authors on request. Raw data are available from the LONI database (https://ida.loni.usc.edu). Acknowledgements We would like to thank the Strategic Research Programme in Neuroscience at Karolinska Institutet (StratNeuro), the Swedish Research Council (VR) 2016-02282, 2021-01861, the Center for Innovative Medicine (CIMED) FoUI-954459, FoUI-975174, the regional agreement on medical training and clinical research (ALF) between Stockholm County Council and Karolinska Institutet FoUI-952838, FoUI-954893, The Swedish Brain Foundation (Hjärnfonden) FO2021-0119, FO2022-0084, The Swedish Alzheimer's Foundation (Alzheimerfonden) AF-967495, AF-980387, The Swedish Parkinson's foundation (Parkinsonfonden) 1443/2022, 1521/23 King Gustaf V:s and Queen Victorias Foundation 20221213_064354, and Olle Engkvists Foundation (Olle Engkvists Stiftelse) 186-0660, 224-0069 as well as Birgitta and Sten Westerberg for additional financial support. Funding DF receives funding from the Swedish Research Council (Vetenskapsrådet, grant 2022-00916), the Center for Innovative Medicine (CIMED, grants 20200505 and FoUI-988826), the regional agreement on medical training and clinical research of Stockholm Region (ALF Medicine, grants FoUI-962240 and FoUI-987534), the Swedish Brain Foundation (Hjärnfonden FO2023-0261, FO2022-0175, FO2021-0131), the Swedish Alzheimer Foundation (Alzheimerfonden AF-968032, AF-980580, AF-994058), the Swedish Dementia Foundation (Demensfonden), the Gamla Tjänarinnor Foundation, the Gun och Bertil Stohnes Foundation, Funding for Research from Karolinska Institutet, Neurofonden, and the Foundation for Geriatric Diseases at Karolinska Institutet, as well as contributions from private bequests. MJG is supported by the ‘Miguel Servet’ program (CP19/00031) of the Instituto de Salud Carlos III – Fondo Europeo de Desarrollo Regional (ISCIII-FEDER). The ADNI is funded by the National Institute on Aging, by the National Institute of Biomedical Imaging and Bioengineering, and through generous contributions from the following: Alzheimer’s Association; Alzheimer’s Drug Discovery Foundation; BioClinica, Inc.; Biogen Idec Inc.; Bristol-Myers Squibb Company; Eisai Inc.; Elan Pharmaceuticals, Inc.; Eli Lilly and Company; F. Hoffmann-La Roche Ltd and its affiliated company Genentech, Inc.; GE Healthcare; Innogenetics, N.V.; IXICO Ltd.; Janssen Alzheimer Immunotherapy Research & Development, LLC; Johnson & Johnson Pharmaceutical Re- search & Development LLC; Medpace, Inc.; Merck & Co., Inc.; Meso Scale Diagnostics, LLC; NeuroRx Research; Novartis Pharmaceuticals Corporation; Pfizer Inc.; Piramal Imaging; Servier; Synarc Inc.; and Takeda Pharmaceutical Company. The Canadian Institutes of Health Research is providing funds to support ADNI clinical sites in Canada. Private sector contributions are facilitated by the Foundation for the NIH (fnih.org). The grantee organization is the Northern California Institute for Research and Education, and the study is coordinated by the Alzheimer’s Disease Cooperative Study at the University of California, San Diego. ADNI data are disseminated by the Laboratory of Neuro Imaging at the University of California, Los Angeles. Competing interests DF consults for BioArctic and has received honoraria from Esteve. References 2023 Alzheimer’s disease facts and figures (2023) Alzheimers Dement Published online March 14:alz13016. 10.1002/alz.13016 Ferreira D, Nordberg A, Westman E (2020) Biological subtypes of Alzheimer disease: A systematic review and meta-analysis. Neurology 94(10):436–448. 10.1212/WNL.0000000000009058 Jellinger KA (2021) Pathobiological Subtypes of Alzheimer Disease. Dement Geriatr Cogn Disord 49(4):321–333. 10.1159/000508625 Mohanty R, Mårtensson G, Poulakis K et al (2020) Comparison of subtyping methods for neuroimaging studies in Alzheimer’s disease: a call for harmonization. Brain Commun 2(2):fcaa192. 10.1093/braincomms/fcaa192 Murray ME, Graff-Radford NR, Ross OA, Petersen RC, Duara R, Dickson DW (2011) Neuropathologically defined subtypes of Alzheimer’s disease with distinct clinical characteristics: A retrospective study. Lancet Neurol 10(9):785–796. 10.1016/S1474-4422(11)70156-9 Whitwell JL, Dickson DW, Murray ME et al (2012) Neuroimaging correlates of pathologically defined subtypes of Alzheimer’s disease: a case-control study. Lancet Neurol 11(10):868–877. 10.1016/S1474-4422(12)70200-4 Risacher SL, Anderson WH, Charil A et al (2017) Alzheimer disease brain atrophy subtypes are associated with cognition and rate of decline. Neurology 89(21):2176–2186. 10.1212/WNL.0000000000004670 Jellinger KA (2012) Neuropathological subtypes of Alzheimer’s disease. Acta Neuropathol (Berl) 123(1):153–154. 10.1007/s00401-011-0889-9 Poulakis K, Pereira JB, Muehlboeck JS et al (2022) Multi-cohort and longitudinal Bayesian clustering study of stage and subtype in Alzheimer’s disease. Nat Commun 13(1):4566. 10.1038/s41467-022-32202-6 Ferreira D, Mohanty R, Murray ME, Nordberg A, Kantarci K, Westman E (2022) The hippocampal sparing subtype of Alzheimer’s disease assessed in neuropathology and in vivo tau positron emission tomography: a systematic review. Acta Neuropathol Commun 10(1):166. 10.1186/s40478-022-01471-z Levin F, Ferreira D, Lange C et al (2021) Data-driven FDG-PET subtypes of Alzheimer’s disease-related neurodegeneration. Alzheimers Res Ther 13(1):49. 10.1186/s13195-021-00785-9 Caminiti SP, De Francesco S, Tondo G et al FDG-PET markers of heterogeneity and different risk of progression in amnestic MCI. Alzheimers Dement . 2023;n/a(n/a). 10.1002/alz.13385 Ishii K, Kawachi T, Sasaki H et al (2005) Voxel-Based Morphometric Comparison Between Early- and Late-Onset Mild Alzheimer’s Disease and Assessment of Diagnostic Performance of Z Score Images. AJNR Am J Neuroradiol 26(2):333–340 Teune LK, Bartels AL, de Jong BM et al (2010) Typical cerebral metabolic patterns in neurodegenerative brain diseases. Mov Disord 25(14):2395–2404. 10.1002/mds.23291 Grothe MJ, Teipel SJ, for the Alzheimer’s Disease Neuroimaging Initiative (2016) Spatial patterns of atrophy, hypometabolism, and amyloid deposition in Alzheimer’s disease correspond to dissociable functional brain networks: Network-Specificity of AD Pathology. Hum Brain Mapp 37(1):35–53. 10.1002/hbm.23018 Chételat G, Desgranges B, Landeau B et al (2008) Direct voxel-based comparison between grey matter hypometabolism and atrophy in Alzheimer’s disease. Brain 131(1):60–71. 10.1093/brain/awm288 La Joie R, Perrotin A, Barré L et al (2012) Region-Specific Hierarchy between Atrophy, Hypometabolism, and β-Amyloid (Aβ) Load in Alzheimer’s Disease Dementia. J Neurosci 32(46):16265–16273. 10.1523/JNEUROSCI.2170-12.2012 Dukart J, Kherif F, Mueller K et al (2013) Generative FDG-PET and MRI Model of Aging and Disease Progression in Alzheimer’s Disease. PLoS Comput Biol 9(4):e1002987. 10.1371/journal.pcbi.1002987 Strom A, Iaccarino L, Edwards L et al (2021) Cortical hypometabolism reflects local atrophy and tau pathology in symptomatic Alzheimer’s disease. Brain 145(2):713–728. 10.1093/brain/awab294 Landau SM, Thomas BA, Thurfjell L et al (2014) Amyloid PET imaging in Alzheimer’s disease: a comparison of three radiotracers. Eur J Nucl Med Mol Imaging 41(7):1398–1407. 10.1007/s00259-014-2753-3 Hansson O, Seibyl J, Stomrud E et al (2018) CSF biomarkers of Alzheimer’s disease concord with amyloid-β PETand predict clinical progression: A study of fully automated immunoassays in BioFINDER and ADNI cohorts. Alzheimers Dement J Alzheimers Assoc 14(11):1470–1481. 10.1016/j.jalz.2018.01.010 Greve DN, Svarer C, Fisher PM et al (2014) Cortical surface-based analysis reduces bias and variance in kinetic modeling of brain PET data. NeuroImage 92:225–236. 10.1016/j.neuroimage.2013.12.021 Greve DN, Salat DH, Bowen SL et al (2016) Different Partial Volume Correction Methods Lead to Different Conclusions: an 18F-FDG PET Study of Aging. NeuroImage 132:334–343. 10.1016/j.neuroimage.2016.02.042 Landau SM, Harvey D, Madison CM et al (2011) Associations between cognitive, functional, and FDG-PET measures of decline in AD and MCI. Neurobiol Aging 32(7):1207–1218. 10.1016/j.neurobiolaging.2009.07.002 Desikan RS, Ségonne F, Fischl B et al (2006) An automated labeling system for subdividing the human cerebral cortex on MRI scans into gyral based regions of interest. NeuroImage 31(3):968–980. 10.1016/j.neuroimage.2006.01.021 Fischl B, Salat DH, Busa E et al (2002) Whole brain segmentation: automated labeling of neuroanatomical structures in the human brain. Neuron 33(3):341–355. 10.1016/s0896-6273(02)00569-x Fox PT, Mintun MA, Reiman EM, Raichle ME (1988) Enhanced Detection of Focal Brain Responses Using Intersubject Averaging and Change-Distribution Analysis of Subtracted PET Images. J Cereb Blood Flow Metab 8(5):642–653. 10.1038/jcbfm.1988.111 Cerami C, Della Rosa PA, Magnani G et al (2015) Brain metabolic maps in Mild Cognitive Impairment predict heterogeneity of progression to dementia. NeuroImage Clin 7:187–194. 10.1016/j.nicl.2014.12.004 Muehlboeck JS, Westman E, Simmons A (2014) TheHiveDB image data management and analysis framework. Front Neuroinformatics 7:49. 10.3389/fninf.2013.00049 Voevodskaya O, Simmons A, Nordenskjöld R et al (2014) The effects of intracranial volume adjustment approaches on multiple regional MRI volumes in healthy aging and Alzheimer’s disease. Front Aging Neurosci . ;6. Accessed September 5, 2023. https://www.frontiersin.org/articles/ 10.3389/fnagi.2014.00264 Falahati F, Ferreira D, Soininen H et al (2016) The Effect of Age Correction on Multivariate Classification in Alzheimer’s Disease, with a Focus on the Characteristics of Incorrectly and Correctly Classified Subjects. Brain Topogr 29(2):296–307. 10.1007/s10548-015-0455-1 Crane PK, Carle A, Gibbons LE et al (2012) Development and assessment of a composite score for memory in the Alzheimer’s Disease Neuroimaging Initiative (ADNI). Brain Imaging Behav 6(4):502–516. 10.1007/s11682-012-9186-z Gibbons LE, Carle AC, Mackin RS et al (2012) A composite score for executive functioning, validated in Alzheimer’s Disease Neuroimaging Initiative (ADNI) participants with baseline mild cognitive impairment. Brain Imaging Behav 6(4):517–527. 10.1007/s11682-012-9176-1 Choi SE, Mukherjee S, Gibbons LE et al (2020) Development and validation of language and visuospatial composite scores in ADNI. Alzheimers Dement N Y N 6(1):e12072. 10.1002/trc2.12072 Poulakis K, Pereira JB, Mecocci P et al (2018) Heterogeneous patterns of brain atrophy in Alzheimer’s disease. Neurobiol Aging 65:98–108. 10.1016/j.neurobiolaging.2018.01.009 Inguanzo A, Poulakis K, Mohanty R et al (2021) Dementia with Lewy bodies subtypes identified by cluster analysis on structural MRI. Alzheimers Dement 17(S4):e053573. 10.1002/alz.053573 Probst P, Wright MN, Boulesteix AL (2019) Hyperparameters and tuning strategies for random forest. WIREs Data Min Knowl Discov 9(3):e1301. 10.1002/widm.1301 Charil A, Shcherbinin S, Southekal S et al (2019) Tau Subtypes of Alzheimer’s Disease Determined in vivo Using Flortaucipir PET Imaging. J Alzheimers Dis 71(3):1037–1048. 10.3233/JAD-190264 Jack CR, Petersen RC, Xu YC et al (1997) Medial temporal atrophy on MRI in normal aging and very mild Alzheimer’s disease. Neurology 49(3):786–794. 10.1212/wnl.49.3.786 Nugent S, Croteau E, Potvin O et al (2020) Selection of the optimal intensity normalization region for FDG-PET studies of normal aging and Alzheimer’s disease. Sci Rep 10:9261. 10.1038/s41598-020-65957-3 Visualization of Brain Statistics With R Packages ggseg and ggseg3d - Athanasia M. Mowinckel, Didac Vidal-Piñeiro, 2020. Accessed August 31 (2023) https://journals.sagepub.com/doi/full/10.1177/2515245920928009 Minoshima S (1994) Posterior cingulate cortex in Alzheimer’s disease. Lancet 344(8926):895. 10.1016/S0140-6736(94)92871-1 Minoshima S, Giordani B, Berent S, Frey KA, Foster NL, Kuhl DE (1997) Metabolic reduction in the posterior cingulate cortex in very early Alzheimer’s disease. Published online July. 10.1002/ana.410420114 Hoffman JM, Welsh-Bohmer KA, Hanson M et al (2000) FDG PET imaging in patients with pathologically verified dementia. J Nucl Med Off Publ Soc Nucl Med 41(11):1920–1928 Herholz K, Salmon E, Perani D et al (2002) Discrimination between Alzheimer dementia and controls by automated analysis of multicenter FDG PET. NeuroImage 17(1):302–316. 10.1006/nimg.2002.1208 Mosconi L, Tsui WH, Herholz K et al (2008) Multicenter Standardized 18F-FDG PET Diagnosis of Mild Cognitive Impairment, Alzheimer’s Disease, and Other Dementias. J Nucl Med Off Publ Soc Nucl Med 49(3):390–398. 10.2967/jnumed.107.045385 Wahlund LO, Julin P, Johansson SE, Scheltens P (2000) Visual rating and volumetry of the medial temporal lobe on magnetic resonance imaging in dementia: a comparative study. J Neurol Neurosurg Psychiatry 69(5):630–635. 10.1136/jnnp.69.5.630 Westman E, Cavallin L, Muehlboeck JS et al (2011) Sensitivity and Specificity of Medial Temporal Lobe Visual Ratings and Multivariate Regional MRI Classification in Alzheimer’s Disease. PLoS ONE 6(7):e22506. 10.1371/journal.pone.0022506 Scheltens P, Leys D, Barkhof F et al (1992) Atrophy of medial temporal lobes on MRI in probable Alzheimer’s disease and normal ageing: diagnostic value and neuropsychological correlates. J Neurol Neurosurg Psychiatry 55(10):967–972. 10.1136/jnnp.55.10.967 Jack CR, Petersen RC, Xu Y et al (1998) Rate of medial temporal lobe atrophy in typical aging and Alzheimer’s disease. Neurology 51(4):993–999. 10.1212/WNL.51.4.993 Byun MS, Kim SE, Park J et al (2015) Heterogeneity of Regional Brain Atrophy Patterns Associated with Distinct Progression Rates in Alzheimer’s Disease. PLoS ONE 10(11):e0142756. 10.1371/journal.pone.0142756 Boon BDC, Labuzan S, Peng Z et al (2023) Retrospective Evaluation of Neuropathologic Proxies of the Minimal Atrophy Subtype Compared to Corticolimbic Alzheimer Disease Subtypes. Neurol Published online August 14. 10.1212/WNL.0000000000207685 Shima K, Matsunari I, Samuraki M et al (2012) Posterior cingulate atrophy and metabolic decline in early stage Alzheimer’s disease. Neurobiol Aging 33(9):2006–2017. 10.1016/j.neurobiolaging.2011.07.009 Meguro K, LeMestric C, Landeau B, Desgranges B, Eustache F, Baron JC (2001) Relations between hypometabolism in the posterior association neocortex and hippocampal atrophy in Alzheimer’s disease: a PET/MRI correlative study. J Neurol Neurosurg Psychiatry 71(3):315–321. 10.1136/jnnp.71.3.315 Delbeuck X, Van der Linden M, Collette F (2003) Alzheimer’ Disease as a Disconnection Syndrome? Neuropsychol Rev 13(2):79–92. 10.1023/A:1023832305702 Minoshima S, Cross D, Thientunyakit T, Foster NL, Drzezga A (2022) 18F-FDG PET Imaging in Neurodegenerative Dementing Disorders: Insights into Subtype Classification, Emerging Disease Categories, and Mixed Dementia with Copathologies. J Nucl Med 63(Supplement 1):2S–12S. 10.2967/jnumed.121.263194 Shiino A, Watanabe T, Maeda K, Kotani E, Akiguchi I, Matsuda M (2006) Four subgroups of Alzheimer’s disease based on patterns of atrophy using VBM and a unique pattern for early onset disease. NeuroImage 33(1):17–26. 10.1016/j.neuroimage.2006.06.010 Ewers M, Brendel M, Rizk-Jackson A et al (2013) Reduced FDG-PET brain metabolism and executive function predict clinical progression in elderly healthy subjects. NeuroImage Clin 4:45–52. 10.1016/j.nicl.2013.10.018 Mosconi L (2013) Glucose metabolism in normal aging and Alzheimer’s disease: methodological and physiological considerations for PET studies. Clin Transl Imaging 1(4):217–233. 10.1007/s40336-013-0026-y Benvenutto A, Giusiano B, Koric L et al (2018) Imaging Biomarkers of Neurodegeneration in Alzheimer’s Disease: Distinct Contributions of Cortical MRI Atrophy and FDG-PET Hypometabolism. J Alzheimers Dis 65(4):1147–1157. 10.3233/JAD-180292 Perini G, Rodriguez-Vieitez E, Kadir A, Sala A, Savitcheva I, Nordberg A (2021) Clinical impact of 18F-FDG-PET among memory clinic patients with uncertain diagnosis. Eur J Nucl Med Mol Imaging 48(2):612–622. 10.1007/s00259-020-04969-7 Kapasi A, DeCarli C, Schneider JA (2017) Impact of multiple pathologies on the threshold for clinically overt dementia. Acta Neuropathol (Berl) 134(2):171–186. 10.1007/s00401-017-1717-7 Mohanty R, Ferreira D, Frerich S et al (2022) Neuropathologic Features of Antemortem Atrophy-Based Subtypes of Alzheimer Disease. Neurology 99(4):e323–e333. 10.1212/WNL.0000000000200573 Rabinovici GD, Seeley WW, Kim EJ et al (2007) Distinct MRI atrophy patterns in autopsy-proven Alzheimer’s disease and frontotemporal lobar degeneration. Am J Alzheimers Dis Other Demen 22(6):474–488. 10.1177/1533317507308779 Silva-Rodríguez J, Labrador-Espinosa MA, Moscoso A et al (2023) Characteristics of amnestic patients with hypometabolism patterns suggestive of Lewy body pathology. Brain J Neurol Published online June 7:awad194. 10.1093/brain/awad194 Grothe MJ, Moscoso A, Silva-Rodríguez J et al (2023) Differential diagnosis of amnestic dementia patients based on an FDG-PET signature of autopsy-confirmed LATE-NC. Alzheimers Dement 19(4):1234–1244. 10.1002/alz.12763 Herholz K, Nordberg A, Salmon E et al (1999) Impairment of Neocortical Metabolism Predicts Progression in Alzheimer’s Disease. Dement Geriatr Cogn Disord 10(6):494–504. 10.1159/000017196 Poulakis K, Ferreira D, Pereira JB, Smedby Ö, Vemuri P, Westman E (2020) Fully bayesian longitudinal unsupervised learning for the assessment and visualization of AD heterogeneity and progression. Aging 12(13):12622–12647. 10.18632/aging.103623 Vogel JW, Young AL, Oxtoby NP et al (2021) Four distinct trajectories of tau deposition identified in Alzheimer’s disease. Nat Med 27(5):871–881. 10.1038/s41591-021-01309-6 Collij LE, Salvadó G, Wottschel V et al (2022) Spatial-Temporal Patterns of β-Amyloid Accumulation: A Subtype and Stage Inference Model Analysis. Neurology 98(17):e1692–e1703. 10.1212/WNL.0000000000200148 Additional Declarations The authors declare potential competing interests as follows: DF consults for BioArctic and has received honoraria from Esteve. Supplementary Files FDGMRISubtypesSupplMaterial.docx Cite Share Download PDF Status: Published Journal Publication published 22 Nov, 2024 Read the published version in Brain Communications → Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-4454593","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":305009955,"identity":"6c7cf518-6dae-4dd3-acd0-d0afd5e892be","order_by":0,"name":"Sophia H. Wheatley","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA+klEQVRIiWNgGAWjYHACNoYEIGkAYj5gYJCTYCZBC2MDkGFMnBYGJC2JMwipl2/vPfbgQQ1D4nb25ucPEioOp89sZ2D+8AGPFoMz59INEo4xJO7sOWbYkHDmcO5sZgY2SXxWGUjkmEkksDHkbriRYNiQ2HY4dx5QCzMPPofNAGn5B9Ry//nHhsR/h9PlmBmYP//B55kbQC2JbSBbeIC2NBxOkAaGmDQ+HUC/pEkk9knUbziTUzgj4Vi64cxmxjbJHnwOA4aY5I9vNsYGx49v+PChxlpe4vzhwx9+4LOGAexTCRivmQEUP3g1QLXAQR0B1aNgFIyCUTASAQCOMlE+y5Vp3wAAAABJRU5ErkJggg==","orcid":"https://orcid.org/0000-0003-3041-1954","institution":"Karolinska Institutet, Stockholm, Sweden","correspondingAuthor":true,"prefix":"","firstName":"Sophia","middleName":"H.","lastName":"Wheatley","suffix":""},{"id":305010999,"identity":"f1048085-0aa3-4c5f-b161-3fd32037209a","order_by":1,"name":"Rosaleena Mohanty","email":"","orcid":"https://orcid.org/0000-0001-6499-1251","institution":"Karolinska Institutet, Stockholm, Sweden","correspondingAuthor":false,"prefix":"","firstName":"Rosaleena","middleName":"","lastName":"Mohanty","suffix":""},{"id":305011001,"identity":"217b0e03-7244-413a-9a8b-3a7447cf9750","order_by":2,"name":"Konstantinos Poulakis","email":"","orcid":"https://orcid.org/0000-0002-6600-8086","institution":"Karolinska Institutet, Stockholm, Sweden","correspondingAuthor":false,"prefix":"","firstName":"Konstantinos","middleName":"","lastName":"Poulakis","suffix":""},{"id":305011003,"identity":"0eeafce4-c91e-4ccd-a6c0-fcfa65f51fbc","order_by":3,"name":"Fedor Levin","email":"","orcid":"","institution":"Deutsches Zentrum für Neurodegenerative Erkrankungen (DZNE), Rostock, Germany","correspondingAuthor":false,"prefix":"","firstName":"Fedor","middleName":"","lastName":"Levin","suffix":""},{"id":305012147,"identity":"f1992e83-05df-490b-a88c-2aa93f34ebe2","order_by":4,"name":"J-Sebastian Muehlboeck","email":"","orcid":"","institution":"Karolinska Institutet, Stockholm, Sweden","correspondingAuthor":false,"prefix":"","firstName":"J-Sebastian","middleName":"","lastName":"Muehlboeck","suffix":""},{"id":305012148,"identity":"f775a9ce-aa88-462f-8206-515d8610ffef","order_by":5,"name":"Agneta Nordberg","email":"","orcid":"https://orcid.org/0000-0001-7345-5151","institution":"Karolinska Institutet, Stockholm, Sweden","correspondingAuthor":false,"prefix":"","firstName":"Agneta","middleName":"","lastName":"Nordberg","suffix":""},{"id":305012149,"identity":"c23cad3b-0bed-4aff-bbd1-62ed4859a40c","order_by":6,"name":"Michel J. Grothe","email":"","orcid":"https://orcid.org/0000-0003-2600-9022","institution":"CIEN Foundation, Madrid, Spain","correspondingAuthor":false,"prefix":"","firstName":"Michel","middleName":"J.","lastName":"Grothe","suffix":""},{"id":305012829,"identity":"bdd9bddd-b75e-45c0-9e6e-78a74f8a9b68","order_by":7,"name":"Daniel Ferreira","email":"","orcid":"https://orcid.org/0000-0001-9522-4338","institution":"Karolinska Institutet, Stockholm, Sweden","correspondingAuthor":false,"prefix":"","firstName":"Daniel","middleName":"","lastName":"Ferreira","suffix":""},{"id":305012830,"identity":"1aafde92-5975-4457-ab0f-52bb4d1fdc75","order_by":8,"name":"Eric Westman","email":"","orcid":"https://orcid.org/0000-0002-3115-2977","institution":"Karolinska Institutet, Stockholm, Sweden","correspondingAuthor":false,"prefix":"","firstName":"Eric","middleName":"","lastName":"Westman","suffix":""}],"badges":[],"createdAt":"2024-05-21 11:50:18","currentVersionCode":1,"declarations":{"humanSubjects":true,"vertebrateSubjects":false,"conflictsOfInterestStatement":true,"humanSubjectEthicalGuidelines":true,"humanSubjectConsent":true,"humanSubjectClinicalTrial":false,"humanSubjectCaseReport":false,"vertebrateSubjectEthicalGuidelines":false},"doi":"10.21203/rs.3.rs-4454593/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-4454593/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1093/braincomms/fcae426","type":"published","date":"2024-11-23T00:00:00+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":57518328,"identity":"0bb4d96d-ec71-4634-b963-3b3d898b20c1","added_by":"auto","created_at":"2024-05-31 20:31:00","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":704856,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eData-driven AD subtypes based on FDG-PET and MRI.\u003c/strong\u003eOverall patterns of neurodegeneration in Aβ+ AD versus Aβ- CNs visualised for \u003cstrong\u003ea)\u003c/strong\u003ehypometabolism in FDG-PET and \u003cstrong\u003eb)\u003c/strong\u003e atrophy in MRI. Clustering identified five subtypes of each \u003cstrong\u003ec)\u003c/strong\u003e hypometabolism and \u003cstrong\u003ed)\u003c/strong\u003e atrophy, which are shown by dendrograms and brain maps in each modality. Corresponding patterns of \u003cstrong\u003ee)\u003c/strong\u003e atrophy patterns in FDG-PET subtypes and \u003cstrong\u003ef)\u003c/strong\u003e hypometabolism patterns in MRI subtypes were visualized. All brain maps are represented as w-scores where regional values (pons scaled SUVR in FDG-PET and volumes in MRI) are adjusted for age, sex, education and \u003cem\u003eAPOE\u003c/em\u003eε4 carriership.\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-4454593/v1/3b7e70b4d52431f65bbfabbd.png"},{"id":57518330,"identity":"df3c3f32-0cf1-4290-8862-2b9b6e3e72ad","added_by":"auto","created_at":"2024-05-31 20:31:00","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":265879,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eRelationship between AD subtypes in FDG-PET and MRI at an individual-level.\u003c/strong\u003e \u003cstrong\u003ea)\u003c/strong\u003e Alluvial plot showing the individual-level allocation of the FDG-PET and MRI subtypes (frequency on y axis). Colour coding is used in Figure 4, cortical subtypes in red and limbic subtypes in blue.\u003cstrong\u003e b) \u003c/strong\u003ePercentage combination of FDG-PET (y axis) and MRI (x axis) subtypes calculated using total number of each FDG-PET subtype (total percentage sum up row-wise). Colour coding shows lower (red) to higher percentage (green) agreement.\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-4454593/v1/e5403539a9bf8f09cc278154.png"},{"id":57518332,"identity":"04d6d749-bb30-44ce-8b69-8aa6f79c6d18","added_by":"auto","created_at":"2024-05-31 20:31:00","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":275035,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eData-driven versus hypothesis-driven AD subtypes in FDG-PET and MRI along typicality and severity axes. \u003c/strong\u003eHypothesis-driven subtypes are defined along the ‘typicality’ axis using prior hippocampus-to-cortex ratio classification using grey matter volumes and FDG-PET SUVRs. ‘Typicality’ and ‘severity’ in data-driven (colours) and hypothesis-driven (shapes) subtypes. For agreement across these two subtyping measures, blue triangles and red diamonds are shown under ‘Subtype Agreement’. \u003cstrong\u003ea)\u003c/strong\u003e Data-driven FDG-PET subtypes with FDG-PEThypothesis-driven subtypes along MRI typicality and severity, \u003cstrong\u003eb)\u003c/strong\u003e Data-driven MRI subtypes with MRI hypothesis-driven subtypes along MRI typicality and severity. \u0026nbsp;Abbreviations: CP: Cortical Predominant, CP+: Cortical Predominant+, LP: Limbic Predominant, LP frontal: Limbic Predominant frontal, Min: Minimal atrophy.\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-4454593/v1/70748fcfcfcfbb7110882503.png"},{"id":57518331,"identity":"1ffad176-1b28-41ab-a013-3047842d8508","added_by":"auto","created_at":"2024-05-31 20:31:00","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":203403,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eData-driven neurodegeneration-based AD subtypes along axes of typicality and severity\u003c/strong\u003e. The current study’s neurodegeneration-based AD subtypes were defined along the typicality-severity framework (Ferreira et al., 2020). The placement of the subtypes was calculated using the mean min-max normalised ‘typicality’ and ‘severity’ values for the subtypes (SUVR-based values for FDG-PET subtypes, volume-based for MRI subtypes). The colours correspond to a neurodegeneration AD subtype identified in the current study. Abbreviations: CP: Cortical Predominant, CP+: Cortical Predominant+, LP: Limbic Predominant, LP frontal: Limbic Predominant frontal, Min: Minimal atrophy.\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-4454593/v1/103a3125bc5494812daf6571.png"},{"id":69919999,"identity":"6f1a51e7-adac-4fbd-b91b-b178c0310fce","added_by":"auto","created_at":"2024-11-26 15:12:24","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2397549,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4454593/v1/8f700abf-a2fe-4517-8768-198c6648bb7f.pdf"},{"id":57518329,"identity":"273c090d-3cbb-4a55-9cdd-400f50413869","added_by":"auto","created_at":"2024-05-31 20:31:00","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":2706097,"visible":true,"origin":"","legend":"","description":"","filename":"FDGMRISubtypesSupplMaterial.docx","url":"https://assets-eu.researchsquare.com/files/rs-4454593/v1/2edb950d636bf9563b1cae3d.docx"}],"financialInterests":"The authors declare potential competing interests as follows: DF consults for BioArctic and has received honoraria from Esteve.","formattedTitle":"\u003cp\u003e­­­Divergent Neurodegenerative Patterns: Comparison of FDG-PET- and MRI-based Alzheimer’s Disease Subtypes\u003c/p\u003e","fulltext":[{"header":"Introduction","content":"\u003cp\u003eAlzheimer\u0026rsquo;s disease (AD) is the most common neurodegenerative disease\u003csup\u003e1\u003c/sup\u003e and is heterogeneous in nature\u003csup\u003e2\u003c/sup\u003e. Over the last years, distinct AD subtypes displaying different clinical and biological features have been identified\u003csup\u003e2,3\u003c/sup\u003e. Despite this, there are discrepancies in findings across the studies, especially at an individual-level\u003csup\u003e4\u003c/sup\u003e. The common consensus from AD subtyping research is the presence of distinct cortical-predominant versus limbic-predominant profiles of neurodegeneration, which primarily reflect differential distributions of tau neurofibrillary tangle (NFTs) pathology in neuropathological findings.\u003c/p\u003e \u003cp\u003eCorticolimbic subtypes were first identified from neuropathological data by distributions of NFTs in cortical and limbic regions. Such AD subtypes were coined as \u0026lsquo;Hippocampal Sparing\u0026rsquo; and \u0026lsquo;Limbic Predominant\u0026rsquo; when compared to \u0026lsquo;Typical AD\u0026rsquo;\u003csup\u003e5\u003c/sup\u003e. Typical AD is described by a balanced distribution of NFTs in the cortical and limbic regions, whereas the other two subtypes are described at opposite ends of the spectrum, from low to high hippocampal NFT load relative to neocortical regions. Similar hypothesis-driven approaches have been applied to \u003cem\u003ein vivo\u003c/em\u003e methods such as structural MRI, where atrophy patterns resembled prior neuropathologically-defined subtypes\u003csup\u003e6,7\u003c/sup\u003e. Furthermore, AD subtypes have been observed in data-driven studies using in vivo imaging \u003csup\u003e2,8\u003c/sup\u003e. A framework was built based on tau pathology and atrophy in AD subtype topography along two axes of \u0026lsquo;typicality\u0026rsquo; and \u0026lsquo;severity\u0026rsquo;\u003csup\u003e2\u003c/sup\u003e. \u0026lsquo;Typicality\u0026rsquo; was proposed as a spectrum with \u0026lsquo;Limbic Predominant\u0026rsquo; and \u0026lsquo;Hippocampal Sparing\u0026rsquo; on opposite ends with \u0026lsquo;Typical AD\u0026rsquo; in the middle. Whereas \u0026lsquo;severity\u0026rsquo; reflects findings from atrophy-based studies with \u0026lsquo;Minimal\u0026rsquo; (low atrophy) and \u0026lsquo;Diffuse\u0026rsquo; (high atrophy) patterns as extremes compared to \u0026lsquo;Typical AD.\u0026rsquo; It is not certain yet whether this framework could be applied to imaging modalities other than tau PET and MRI-based atrophy. Furthermore, longitudinal MRI studies of AD subtypes revealed that neurodegeneration progressed along either a cortical or a limbic pathway\u003csup\u003e9\u003c/sup\u003e. Similarly, modelling the spread of tau pathology in AD, two pathways were proposed starting either in the entorhinal cortex or in association cortices\u003csup\u003e10\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eMolecular imaging complements MRI by measuring functional changes for the study of neuropathological hallmarks of AD. However, the main body of research in AD heterogeneity has used MRI scans, although research is branching out to other imaging modalities. There are few studies investigating AD heterogeneity using [\u003csup\u003e18\u003c/sup\u003eF] fluorodeoxyglucose (FDG)-PET. Only one data-driven study has been published using a data-driven approach to FDG-PET to find different patterns of hypometabolism in AD\u003csup\u003e11\u003c/sup\u003e, which has been complemented by a recent study in amnestic mild cognitive impairment individuals (aMCI)\u003csup\u003e12\u003c/sup\u003e. Similar spatial subtypes to those found in MRI studies were identified, including a \u0026lsquo;Cortical Predominant\u0026rsquo; hypometabolism subtype showing similarity to the Hippocampal Sparing atrophy subtype, and a \u0026lsquo;Limbic Predominant\u0026rsquo; subtype. An important aspect to consider for linking the findings in these FDG-PET studies with MRI studies is the use of the term \u0026lsquo;Typical AD\u0026rsquo; pattern, which differs across the two modalities. Prominent parieto-temporal hypometabolism reflects the typical AD pattern in FDG-PET, while widespread atrophy in cortex and as the most typical feature hippocampus reflects the typical AD pattern in MRI.\u003csup\u003e13\u0026ndash;15\u003c/sup\u003e\u003c/p\u003e \u003cp\u003eComparisons of FDG-PET and MRI across the AD continuum have been performed\u003csup\u003e16\u0026ndash;19\u003c/sup\u003e. These comparative studies found correlations across the two modalities, but not a full correspondence of regional neurodegeneration. Similarly, comparison of hypometabolism patterns in aMCI individuals was assessed relative to hippocampal atrophy\u003csup\u003e12\u003c/sup\u003e. However, the combination of FDG-PET and MRI techniques encompassing regions beyond the hippocampus in AD subtypes has not yet been performed.\u003c/p\u003e \u003cp\u003eThe aim of the current study was to simultaneously evaluate heterogeneity in measures of neurodegeneration (FDG-PET and MRI) to: (a) identify data-driven subtypes of AD, thereby testing the existence of corticolimbic pathways, (b) investigate the overlap and relationship of these corticolimbic subtypes across FDG-PET and MRI, (c) compare data-driven with hypothesis-driven methods of subtyping. Corresponding neurodegeneration patterns, and clinical and demographic information were assessed for each of the subtypes. Understanding the link between atrophy and metabolism could lead to better subtype classification in the context of a biological framework of AD.\u003c/p\u003e"},{"header":"Materials and methods","content":"\u003cp\u003eParticipants\u003c/p\u003e\n\u003cp\u003eThe cohort consisted of 176 amyloid-beta negative (A\u0026beta;-) cognitively normal (CN) and 180 amyloid-positive (A\u0026beta;+) Alzheimer\u0026rsquo;s disease (AD) dementia individuals with both an MRI and an FDG-PET scan, from the Alzheimer\u0026rsquo;s Disease Neuroimaging Initiative (ADNI), including ADNI1, ADNI2/GO and ADNI3 phases (Table \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e). The ADNI is a longitudinal multi-centre study aimed at investigating whether neuroimaging methods, together with genetic, clinical, and neuropsychological measures could be used to follow the progression of MCI and AD. The ADNI was launched in 2003 as a public-private partnership, led by principal investigator Michael W. Weiner, MD. For AD patients the inclusion criteria consisted of: Mini-Mental State Examination scores of 20\u0026ndash;26, Clinical Dementia Rating scores of 0.5 or 1.0 and meeting the criteria for probable AD. More information of the inclusion and exclusion criteria for the participants can be found on the ADNI website (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://adni.loni.usc.edu/methods/documents/\u003c/span\u003e\u003c/span\u003e). FDG-PET scans from the initial visit and corresponding MRI scans had a mean of 25 days interval and a maximum of 156 days interval (\u003cem\u003eSuppl. Figure\u0026nbsp;1\u003c/em\u003e). Individuals with AD were included in this study based on A\u0026beta; positivity and CNs on A\u0026beta; negativity. This was determined in first place by an AV45-PET standard uptake value ratio (SUVR) greater than 1.11\u003csup\u003e20\u003c/sup\u003e. If this data was not available for the individual, we used their A\u0026beta; (\u003cspan class=\"CitationRef\"\u003e1\u003c/span\u003e\u0026ndash;\u003cspan class=\"CitationRef\"\u003e42\u003c/span\u003e) values from CSF, which were deemed A\u0026beta; positive if lower than 880pg/ml\u003csup\u003e21\u003c/sup\u003e.\u003c/p\u003e\n\u003cdiv class=\"gridtable\"\u003e\u0026nbsp;\u003ctable id=\"Tab1\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eCohort Demographics and Clinical Characteristics\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003ccolgroup cols=\"4\"\u003e\u003c/colgroup\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eClinical and Demographic Characteristics\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eAlzheimer\u0026rsquo;s Disease\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eCognitively Normal\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003ep values\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eN\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e180\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e176\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e―\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eWomen (%)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e81 (45%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e84 (48%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.682\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eAge (years)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e74 (\u003cspan class=\"CitationRef\"\u003e8\u003c/span\u003e)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e75 (\u003cspan class=\"CitationRef\"\u003e7\u003c/span\u003e)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.867\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eDisease Duration\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e3 (\u003cspan class=\"CitationRef\"\u003e3\u003c/span\u003e)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e―\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e―\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eEducation (years)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e15 (\u003cspan class=\"CitationRef\"\u003e3\u003c/span\u003e)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e17 (\u003cspan class=\"CitationRef\"\u003e3\u003c/span\u003e)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eAPOE \u0026epsilon;4 (%)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e126 (70%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e34 (19%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eMMSE\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e23 (\u003cspan class=\"CitationRef\"\u003e2\u003c/span\u003e)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e29 (\u003cspan class=\"CitationRef\"\u003e1\u003c/span\u003e)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eGlobal CDR\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.8 (\u003cspan class=\"CitationRef\"\u003e3\u003c/span\u003e)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.03 (0.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eADNI-EF\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e-0.4 (0.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.8 (0.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eADNI-MEM\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e-0.8 (0.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.9 (0.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eADNI-LAN\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e-0.2 (0.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.8 (0.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eCDF t-tau (pg/ml)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e383 (149)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e223 (\u003cspan class=\"CitationRef\"\u003e70\u003c/span\u003e)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eCSF p-tau (pg/ml)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e38 (\u003cspan class=\"CitationRef\"\u003e16\u003c/span\u003e)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e20 (\u003cspan class=\"CitationRef\"\u003e6\u003c/span\u003e)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eCSF A\u0026beta; (pg/ml)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e587 (191)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1204 (333)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n\u003c/div\u003e\n\u003cp\u003eThe values shown in the table are the means with standard deviations in brackets except for number of individuals, women and \u003cem\u003eAPOE\u003c/em\u003e \u0026epsilon;4 carriership for which the percentages are provided. The p-values correspond to the \u0026Chi;\u003csup\u003e2\u0026nbsp;\u003c/sup\u003etests conducted for categorical variables and Kruskal-Wallis for continuous variables and were corrected for multiple comparison with the Holm-\u0026Scaron;id\u0026aacute;k method. Abbreviations: AD: Alzheimer\u0026rsquo;s disease individuals, CN: cognitively normal individuals, \u003cem\u003eAPOE\u003c/em\u003e \u0026epsilon;4: apolipoprotein E \u0026epsilon;4 allele, MMSE: Mini Mental State Examination, CDR: The Clinical Dementia Rating Scale, ADNI-EF Alzheimer\u0026rsquo;s Disease Neuroimaging Initiative executive function composite score, ADNI-MEM: Alzheimer\u0026rsquo;s Disease Neuroimaging Initiative memory composite score, ADNI-LAN: Alzheimer\u0026rsquo;s Disease Neuroimaging Initiative language composite score, CSF t-tau: total tau, CSF p-tau: phosphorylated tau, CSF A\u0026beta;: amyloid-beta 1-42 peptide.\u003c/p\u003e\n\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\n \u003ch2\u003eFDG-PET\u003c/h2\u003e\n \u003cp\u003eMultiple scanners were used for the FDG-PET scans and followed the appropriate protocols (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://adni.loni.usc.edu/methods/documents/\u003c/span\u003e\u003c/span\u003e). Dynamic 3D scans made up of six frames of 5 minutes were retrieved 30 to 60 minutes after administration of [\u003csup\u003e18\u003c/sup\u003eF] FDG. FDG-PET scans were co-registered to individuals\u0026rsquo; MRI scans in MNI space using PETSurfer\u003csup\u003e22,23\u003c/sup\u003e and regional standard uptake ratio (SUVR) were extracted without partial volume correction\u003csup\u003e19,24\u003c/sup\u003e. We investigated 82 bilateral regions of interest (ROIs) defined on individual\u0026rsquo;s MRI including cortical and subcortical structures based on the standard atlases provided by FreeSurfer\u003csup\u003e25,26\u003c/sup\u003e. From these regions, SUVRs were extracted in MNI space. Intensity normalisation was carried out by dividing the regional values by the individual\u0026rsquo;s global mean\u003csup\u003e27\u003c/sup\u003e. This type of normalisation has been used in prior FDG-PET dementia and subtyping studies\u003csup\u003e11,28\u003c/sup\u003e.\u003c/p\u003e\n\u003c/div\u003e\n\u003ch3\u003eMRI\u003c/h3\u003e\n\u003cp\u003eT1-weighted magnetisation-prepared rapid gradient-echo (MPRAGE) MRI scans were preprocessed using FreeSurfer (version 6.0.0, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://freesurfer.net\u003c/span\u003e\u003c/span\u003e). MRI scans were collected from various sites using multiple scanners (both 1.5 and 3T scanners), following the appropriate protocol (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://adni.loni.usc.edu/methods/documents/\u003c/span\u003e\u003c/span\u003e). These data was preprocessed in-house through theHiveDB database\u003csup\u003e29\u003c/sup\u003e. Briefly, the preprocessing pipeline involved removal of artefacts, transformation to Talairach space and segmentation of cortical and subcortical regions. From the same 41 regions used for the FDG-PET, cortical and subcortical grey matter volumes were extracted.\u003c/p\u003e\n\u003cp\u003eThe raw MRI scans were assessed visually for quality control. Additionally, the estimated total intracranial volumes (ICV) from the FreeSurfer output were plotted against the regional volumes to identify any outliers where the values were grossly under- or over- estimated. The raw images were then checked for these cases to confirm exclusion. Further, scans that were not segmented or registered properly were excluded. In total, 29 individuals (6 CN, 23 AD) were excluded due to: poor PET scan quality, failed quality control, poor segmentation or under/over-estimated ICV volumes.\u003c/p\u003e\n\u003cp\u003eThe grey matter volumes were adjusted for head size per region by using a residual approach using the eICV values from the cognitively normal individuals \u003csup\u003e30,31\u003c/sup\u003eas shown below:\u003c/p\u003e\n\u003cdiv id=\"Sec5\" class=\"Section2\"\u003e\n \u003cp\u003eVolume\u003csub\u003eadj\u003c/sub\u003e = Volume\u003csub\u003eraw\u003c/sub\u003e \u0026ndash; \u0026beta;(eICV\u003csub\u003eraw\u003c/sub\u003e \u0026ndash; eICV\u003csub\u003emean\u003c/sub\u003e)\u003c/p\u003e\n \u003cp\u003eThis adjustment was carried out for each group (CN, AD) in relation to the CNs. Volume\u003csub\u003eraw\u003c/sub\u003e is the uncorrected volume for the brain region of interest. The \u0026beta; value and mean ICV, \u003cem\u003eeICV\u003c/em\u003e\u003csub\u003e\u003cem\u003emean\u003c/em\u003e\u003c/sub\u003e, are calculated by running a linear regression model per region of interest in relation to the ICV using data from the CNs.\u003c/p\u003e\n \u003cp\u003eCSF biomarkers\u003c/p\u003e\n \u003cp\u003eIn addition to CSF A\u0026beta; (\u003cspan class=\"CitationRef\"\u003e1\u003c/span\u003e\u0026ndash;\u003cspan class=\"CitationRef\"\u003e42\u003c/span\u003e) values, we assessed CSF measures of tau phosphorylated at threonine 181 (p-tau) and total tau (t-tau) taken from an automated Elecsys cobas e 601 analyser.\u003c/p\u003e\n \u003cp\u003eNeuropsychological Testing\u003c/p\u003e\n \u003cp\u003eWe used the Mini-Mental State Examination (MMSE) to assess global cognition and composite scores for executive function (ADNI-EF), memory (ADNI-MEM), and language (ADNI-LAN)\u003csup\u003e32\u0026ndash;34\u003c/sup\u003e to assess specific cognitive domains. For MMSE, lower scores indicate higher global cognitive impairment and lower scores for each of the composite scores indicate greater impairment.\u003c/p\u003e\n \u003cp\u003eData-driven Subtyping: Hierarchical Clustering Analysis\u003c/p\u003e\n \u003cp\u003eTo classify individuals into subtypes in FDG-PET and MRI, unsupervised random forest hierarchical clustering was performed (\u003cem\u003eSuppl. Figures\u0026nbsp;2, 3\u003c/em\u003e) to identify the linear and non-linear relationships from the regional values. This method has previously been used to investigate heterogeneity within neurodegenerative diseases using grey matter volumes\u003csup\u003e10,35,36\u003c/sup\u003e. In the current study, an identical clustering procedure was applied to regional glucose metabolism values and grey matter volumes using the same 41 bilateral cortical and subcortical ROIs. Two separate clustering models were performed, one using glucose metabolism values and another with grey matter volumes.\u003c/p\u003e\n \u003cp\u003eFirst, a distance matrix was calculated based on the regional values using a random forest algorithm, which provides information on the similarities and dissimilarities across the brain regions. The optimisation of hyperparameters was performed by selecting the values with the lowest out-of-bag errors\u003csup\u003e37\u003c/sup\u003e for: number of variables randomly sampled at each split (\u003cem\u003emtry\u003c/em\u003e) and minimum size of the terminal nodes (\u003cem\u003enodesize\u003c/em\u003e). The number of trees was set to 20,000 for all models. Additionally, the stability of the chosen random forest model was tested by running the random forest algorithm 100 times. The differences between the chosen model and the simulated models were calculated (\u003cem\u003eSuppl. Figures\u0026nbsp;4, 5\u003c/em\u003e).\u003c/p\u003e\n \u003cp\u003eThe distance matrix was then reduced to three dimensions using classical multi-dimensional scaling to simplify the interpretation of the most important features that distinguish the clusters from each other. The first three dimensions from the multi-dimensional scaling were used for clustering as they explained the greatest differences between the groups, identified by plotting the eigenvalues of the dimensions. Agglomerative hierarchical clustering with average linkage was then run using the reduced matrix to identify clusters. The output of the clustering is a dendrogram and to group the individuals into subtypes the number of clusters needs to be chosen. This number was derived by using various cluster validation indices, namely the Calinski-Harabasz, Davies-Bouldin, Dunn and Silhouette indices from \u003cem\u003eNbClust\u003c/em\u003e and \u003cem\u003efpc\u003c/em\u003e libraries in R. The Calinski-Harabasz index is calculated by comparing the between cluster variance with the within cluster variance. The Davies-Bouldin index evaluates cluster compactness and distinctness by comparing within-cluster distances to between-cluster distances. Similarly, the Dunn index assess cluster compactness and separation. The Silhouette index is the measure of how well the objects fit into its allocated cluster. Collectively, these four indices capture a well-rounded assessment for choosing the number of clusters for our models.\u003c/p\u003e\n \u003cp\u003eInter-modality Comparison of Subtypes\u003c/p\u003e\n \u003cp\u003eComparisons between FDG-PET and MRI subtypes were investigated by: 1) subtyping in one modality and mapping the corresponding atrophy/hypometabolism patterns in the other modality, 2) frequency of the derived FDG-PET and MRI subtypes, 3) crossover between individual subtype allocations.\u003c/p\u003e\n \u003cp\u003eRegional atrophy patterns of the FDG-PET subtypes and regional hypometabolism patterns of the MRI subtypes were assessed using w-scores. W-scores\u003csup\u003e17\u003c/sup\u003e were calculated as z-scores by subtracting the mean and dividing by the standard deviation from the cognitively normal data and adjusted for covariates: age, sex, education and \u003cem\u003eAPOE \u0026epsilon;4\u003c/em\u003e allele carriership. Furthermore, the frequencies of the individuals belonging to a given subtype in both modalities, e.g., an individual being classified as Cortical Predominant in FDG-PET clustering and Cortical Predominant in MRI clustering, was calculated. The different modality-specific subtypes were compared in terms of the defining topographical patterns, which enabled us to visualise the similarities and differences across the subtypes and whether individuals had similar neurodegeneration patterns in the two modalities. To test whether the modality-specific subtype classifications overlapped at an individual level, the individuals\u0026rsquo; subtype allocations were compared in an alluvial plot. Additionally, to numerically demonstrate the overlap, a ratio for each FDG-PET and MRI subtype pairing was calculated using the total number of cases for each FDG-PET subtype. For instance, the ratio was determined by dividing the number of Cortical Predominant MRI cases by the total number of Cortical Predominant FDG-PET cases.\u003c/p\u003e\n \u003cp\u003eData-driven versus Hypothesis-driven Subtypes\u003c/p\u003e\n \u003cp\u003eThis analysis was performed to compare our data-driven method with a previously described hypothesis-driven method designed to identify neuropathologically defined corticolimbic subtypes\u003csup\u003e5\u0026ndash;7\u003c/sup\u003e. The aim was to investigate whether there would be an overlap in the subtype categorisation from these two methods. Given that hypothesis-driven subtyping has been applied to MRI data previously, this analysis additionally aimed to determine whether such a subtyping could be adapted using FDG-PET data, mirroring techniques previously applied to tau PET\u003csup\u003e38\u003c/sup\u003e. Prior hypothesis-driven studies have used the hippocampus-to-cortex ratio to group AD individuals into one of three subtypes (\u0026lsquo;Limbic Predominant\u0026rsquo;, \u0026lsquo;Typical AD\u0026rsquo;, \u0026apos;Hippocampal Sparing\u0026rsquo;) using a two-step process using the 75th and 25th percentiles and median hippocampal and cortical values\u003csup\u003e5\u0026ndash;7\u003c/sup\u003e. Hypothesis-driven subtype labels in the current study were \u0026lsquo;Limbic Predominant\u0026rsquo; (low hippocampus-to-cortex ratio) and \u0026lsquo;Cortical Predominant\u0026rsquo; (high hippocampus-to-cortex ratio).\u003c/p\u003e\n \u003cp\u003eThe hippocampus-to-cortex ratio captures whether a subtype shows either a limbic or a cortical pattern and was conceptualized as the \u0026lsquo;typicality\u0026rsquo; axis in a recent framework explaining the topography of AD subtypes\u003csup\u003e2\u003c/sup\u003e. A second axis of the subtypes was termed \u0026lsquo;severity\u0026rsquo;, which refers whether a subtype shows atypically low or high atrophy. Hence, in this analysis, we further compared the data-driven and hypothesis-driven subtyping in the context of the published framework. For MRI, the measure for \u0026lsquo;typicality\u0026rsquo; was the hippocampus-to-cortex ratio in volume measures and the measure for \u0026lsquo;severity\u0026rsquo; was the total grey matter volume. For FDG-PET, \u0026lsquo;typicality\u0026rsquo; was hippocampus-to-cortex ratio in glucose metabolism and \u0026lsquo;severity\u0026rsquo; was the total average cortical SUVR. All subtype classifications were then plotted along these two axes. These are orthogonal to each other, explaining the differing patterns in the five common patterns of atrophy (hippocampal-sparing, limbic-predominant, typical AD, minimal and diffuse). To explore this, Pearson correlations were calculated for measures of \u0026lsquo;typicality\u0026rsquo; and \u0026lsquo;severity\u0026rsquo; within and between the two imaging modalities. Typicality and severity measures were further used to generate a conceptual figure for the two modalities\u0026rsquo; subtypes along these two axes. Values for each subtype were rescaled to range from values 0 to 1 using min-max normalisation. These normalised values where then averaged and plotted.\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec6\" class=\"Section2\"\u003e\n \u003ch2\u003eStatistical Analysis\u003c/h2\u003e\n \u003cp\u003eRStudio and R version 4.2.0 were used for statistical analyses. W-scores were plotted to show the differences in the regional SUVRs and grey matter volumes in the subtypes compared to the cognitively normal (CN) individuals \u003csup\u003e17,39\u003c/sup\u003e. For FDG-PET, the w-scores were calculated using SUVRs scaled by the pons as the reference region. The pons was chosen because it has been compared with other reference regions across ageing and Alzheimer\u0026rsquo;s disease studies using FDG-PET and it has been shown to work well with both PVC and non-PVC data\u003csup\u003e40\u003c/sup\u003e. These values were averaged across the AD individuals respectively and reversed so that the higher w-scores correspond to greater neurodegeneration in the AD group. Brain maps were created using the \u003cem\u003eggseg\u003c/em\u003e library in R\u003csup\u003e41\u003c/sup\u003e.\u003c/p\u003e\n \u003cp\u003eThe demographic and clinical variables of the subtypes were compared using \u0026chi;\u0026sup2; and Kruskal-Wallis tests that were adjusted for multiple comparisons with the Holm-\u0026Scaron;id\u0026aacute;k method. Significance was deemed if the p values were \u0026lt;\u0026thinsp;0.05 where the null hypothesis can be rejected when \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(p\u0026lt;\\text{\u0026alpha; }\\)\u003c/span\u003e\u003c/span\u003ewith \u0026alpha;\u0026thinsp;=\u0026thinsp;0.001, this stringent alpha value was set to avoid Type 1 family-wise errors. Comparisons were made between the subtypes and the cognitively normal individuals, as well as pairwise comparisons between the groups.\u003c/p\u003e\n\u003c/div\u003e"},{"header":"Results","content":"\u003cp\u003eCohort characteristics\u003c/p\u003e\n\u003cp\u003eThe basic clinical and demographic information of the cohort is reported in Table \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e. As expected, The AD group significantly differed from the CN group in terms of percentage of \u003cem\u003eAPOE\u003c/em\u003e \u0026epsilon;4 carriers, cognitive measures, and CSF markers. There were no significant differences between our AD and CN groups in terms of sex and age.\u003c/p\u003e\n\u003cp\u003eFDG-PET Clustering\u003c/p\u003e\n\u003cp\u003eThe overall AD group\u0026rsquo;s hypometabolism pattern showed neurodegeneration in posterior cingulate and frontal cortical regions and deeper structures such as the hippocampus (Fig. \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003ea\u003cem\u003e)\u003c/em\u003e. The clustering model resulted in two main distinguishing patterns, neurodegeneration in cortical versus limbic pathways (Fig. \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003ec). The Calinski-Harabasz index peaked at five clusters, for the Davies-Bouldin index was lower at five, the Dunn index plateaus at five and the Silhouette index was highest at 3\u0026ndash;5 clusters (\u003cem\u003eSuppl. Figure\u0026nbsp;6\u003c/em\u003e). Therefore, we chose five clusters as the optimal solution for the FDG-PET model. The FDG-PET model was split into five distinct hypometabolism-based subtypes. Three subtypes showed cortical-predominant hypometabolism of differing severity and spatial distribution. The Cortical Predominant posterior subtype had cortical hypometabolism mainly in the posterior regions (9%), whereas the Cortical Predominant and Cortical Predominant\u0026thinsp;+\u0026thinsp;subtypes showed more widespread cortical hypometabolism (32% and 11% respectively). The Cortical Predominant\u0026thinsp;+\u0026thinsp;subtype had greater hypometabolism than the other two cortical predominant subtypes. Although all the subtypes showed some hypometabolism in the hippocampus, these subtypes had proportionally less involvement of this region compared to the cortical areas. Two subtypes displayed limbic hypometabolism, focal to the medial temporal and deeper structures (amygdala, hippocampus). Here, a principal Limbic Predominant subtype (36%) could be distinguished from and a Limbic Predominant frontal subtype (13%). In the clustering dendrogram (Fig. \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e), the Limbic Predominant frontal cluster originates from its own branch, whereas the Limbic Predominant cluster comes from the same branch as the Cortical Predominant posterior cluster. Thus, the clustering separates these subtypes by a frontal versus a posterior hypometabolism pattern. By contrast, the Cortical Predominant and Cortical Predominant\u0026thinsp;+\u0026thinsp;clusters are separated on the opposite side of the dendrogram by the severity of their cortical hypometabolic patterns.\u003c/p\u003e\n\u003cp\u003eFor inter-modality comparisons, the corresponding atrophy patterns in these subtypes were plotted (Fig. \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003ee\u003cem\u003e)\u003c/em\u003e. Based on visual comparison of the w-scores, the brain maps were topographically similar across FDG-PET subtypes, showing AD-typical atrophy in medial temporal, hippocampal, and some frontal areas. However, the atrophy pattern of the Cortical Predominant subtype was not as widespread in the cortical regions compared to the hypometabolism. Additional maps using PVC SUVRs from PETSurfer were plotted (\u003cem\u003eSuppl. Figure\u0026nbsp;8\u003c/em\u003e). These brain maps did not differ greatly from our maps in Fig. \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e topographically but did result in lower w-scores.\u003c/p\u003e\n\u003cp\u003eRegarding demographic and clinical differences among the FDG-PET AD subtypes (Table \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003e), the Cortical Predominant\u0026thinsp;+\u0026thinsp;subtype was the youngest (67.5 years), had the earliest age at onset (64.7 years), more pronounced language impairment, and lowest executive function scores. This subtype also had the highest grey matter volume-based and SUVR-based hippocampus-to-cortex ratios. The other two cortical subtypes (Cortical Predominant, Cortical Predominant posterior) had a higher SUVR-based hippocampus-to-cortex ratio than the limbic subtypes. Cortical Predominant posterior also had a high hippocampus-to-cortex ratio using grey matter volumes compared to the limbic subtypes. Among the limbic subtypes, Limbic Predominant frontal was the oldest, had latest age at onset and worst language scores. There were no significant differences between the subtypes for the other variables: sex, disease duration, years of education, \u003cem\u003eAPOE \u0026epsilon;4\u003c/em\u003e carriers, MMSE, CDR, cognitive measures of memory, and CSF biomarkers. Although not statistically significant, two of the cortical subtypes, Cortical Predominant and Cortical Predominant+, had lower percentage of \u003cem\u003eAPOE \u0026epsilon;4\u003c/em\u003e carriers and Limbic Predominant frontal the highest percentage of \u003cem\u003eAPOE \u0026epsilon;4\u003c/em\u003ecarriers.\u0026nbsp;\u003c/p\u003e\n\u003ctable id=\"Tab2\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eDemographic and clinical characteristics of the FDG-PET AD subtypes\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eDemographic \u0026amp; Clinical Characteristics\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eCortical Predominant\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eCortical Predominant+\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eLimbic Predominant\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eLimbic Predominant frontal\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eCortical Predominant posterior\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eCognitively Normal\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003ep values\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eN (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e57 (32%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e20 (11%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e64 (36%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e23 (13%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e16 (9%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e176\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e―\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eWomen (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e22 (39%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e10 (50%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e32 (50%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e10 (43%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e7 (44%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e84 (48%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.768\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eAge (years)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e75 (7.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e67 (8.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e74 (6.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e78 (6.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e73 (8.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e75 (7.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eDisease Duration (years)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e2.5 (2.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e2.7 (2.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e2.9 (2.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e3.1 (2.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.8 (2.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.274\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eAge at Onset (years)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e72 (8.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e65 (8.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e71 (\u003cspan class=\"CitationRef\"\u003e7\u003c/span\u003e)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e75 (7.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e71 (8.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003csup\u003eab\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eEducation (years)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e15 (2.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e16 (2.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e16 (2.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e15 (3.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e15 (2.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e17 (2.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.361\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cem\u003eAPOE\u003c/em\u003e ɛ4 (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e35 (61%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e11 (55%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e48 (75%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e20 (87%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e12 (75%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e34 (19%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.083\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eMMSE\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e23 (2.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e22 (2.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e24 (\u003cspan class=\"CitationRef\"\u003e2\u003c/span\u003e)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e23 (1.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e23 (1.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e29 (1.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.167\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eGlobal CDR\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.77 (0.25)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.88 (0.22)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.73 (0.25)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.77 (0.25)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.89 (0.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.025 (0.16)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.225\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eADNI-EF\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e-0.43 (0.62)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e-0.99 (0.73)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e-0.19 (0.61)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e-0.6 (0.51)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e-0.58 (0.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.8 (0.47)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003csup\u003ebc\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eADNI-MEM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e-0.85 (0.32)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e-0.89 (0.32)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e-0.66 (0.34)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e-0.93 (0.31)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e-0.8 (0.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.89 (0.52)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.007\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eADNI-LAN\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e-0.24 (0.53)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e-0.43 (0.52)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.041 (0.54)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e-0.5 (0.61)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e-0.31 (0.56)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.82 (0.52)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003csup\u003ebce\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eCSF t-tau (pg/ml)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e378 (145)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e407 (143)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e415 (167)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e335 (98)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e335 (152)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e223 (\u003cspan class=\"CitationRef\"\u003e70\u003c/span\u003e)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.958\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eCSF p-tau (pg/ml)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e38 (\u003cspan class=\"CitationRef\"\u003e15\u003c/span\u003e)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e41 (\u003cspan class=\"CitationRef\"\u003e16\u003c/span\u003e)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e43 (\u003cspan class=\"CitationRef\"\u003e19\u003c/span\u003e)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e33 (\u003cspan class=\"CitationRef\"\u003e10\u003c/span\u003e)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e33 (\u003cspan class=\"CitationRef\"\u003e15\u003c/span\u003e)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e20 (6.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.961\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eCSF A\u0026beta; (pg/ml)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e585 (248)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e588 (173)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e588 (169)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e596 (150)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e578 (139)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1204 (333)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.537\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eSUVR-based Hippocampus-to-cortex ratio\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.35 (0.04)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.38 (0.047)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.3 (0.034)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.32 (0.032)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.33 (0.032)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.32 (0.029)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eGrey Matter Volume-based Hippocampus-to-cortex ratio\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.16 (0.021)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.19 (0.029)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.16 (0.02)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.16 (0.023)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.17 (0.025)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.18 (0.021)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003csup\u003efg\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eTotal Grey Matter Volume (mm\u0026sup3;)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e558154 (33242)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e544820 (35401)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e554040 (38026)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e537830 (47861)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e539736 (36515)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e580718 (40407)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.129\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eTotal Average Cortical Uptake (SUVR)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.5 (0.16)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.5 (0.12)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.6 (0.18)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.6 (0.11)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.6 (0.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.7 (0.19)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.021\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003eThe values shown in the table are the means with standard deviations in brackets except for number of individuals, women and \u003cem\u003eAPOE\u003c/em\u003e \u0026epsilon;4 for which the percentages are provided. The reported p-values correspond to \u0026Chi;2 tests which were used for categorical variables and Kruskal-Wallis for continuous variables and were corrected for multiple comparison with the Holm-\u0026Scaron;id\u0026aacute;k method. Footnotes indicate cases where p values were significant in the post hoc pairwise comparisons across AD subtypes, p \u0026lt; 0.05. The CN group data is displayed for reference. Abbreviations: CP: Cortical Predominant, CP+: Cortical Predominant+, LP: Limbic Predominant, LP fr.: Limbic Predominant frontal, CP post.: Cortical Predominant posterior, CN: cognitively normal individuals, \u003cem\u003eAPOE\u003c/em\u003e \u0026epsilon;4: apolipoprotein E \u0026epsilon;4 allele, MMSE: Mini Mental State Examination, CDR: The Clinical Dementia Rating Scale, ADNI-EF Alzheimer\u0026rsquo;s Disease Neuroimaging Initiative executive function composite score, ADNI-MEM: Alzheimer\u0026rsquo;s Disease Neuroimaging Initiative memory composite score, ADNI-LAN: Alzheimer\u0026rsquo;s Disease Neuroimaging Initiative language composite score, CSF t-tau: total tau, CSF p-tau: phosphorylated tau, CSF A\u0026beta;: amyloid-beta 1-42 peptide.\u003c/p\u003e\n\u003col style=\"list-style-type: lower-alpha;\"\u003e\n \u003cli\u003eCortical Predominant+ \u0026lt; Limbic Predominant, p \u0026lt; 0.05\u003c/li\u003e\n \u003cli\u003eCortical Predominant+ \u0026lt; Cortical Predominant, p \u0026lt; 0.05\u003c/li\u003e\n \u003cli\u003eLimbic Predominant \u0026lt; Limbic Predominant frontal, p \u0026lt; 0.05\u003c/li\u003e\n \u003cli\u003eLimbic Predominant frontal \u0026lt; Limbic Predominant, p \u0026lt; 0.05\u003c/li\u003e\n \u003cli\u003eLimbic Predominant \u0026lt; Cortical Predominant+, p \u0026lt; 0.05\u003c/li\u003e\n \u003cli\u003eLimbic Predominant frontal \u0026lt; Cortical Predominant+, p \u0026lt; 0.05\u003c/li\u003e\n \u003cli\u003eLimbic Predominant \u0026lt; Cortical Predominant, p \u0026lt; 0.05\u003c/li\u003e\n \u003cli\u003eLimbic Predominant \u0026lt; Cortical Predominant posterior, p \u0026lt; 0.05\u003c/li\u003e\n \u003cli\u003eCortical Predominant posterior \u0026lt; Cortical Predominant+, p \u0026lt; 0.05\u003c/li\u003e\n \u003cli\u003eCortical Predominant \u0026lt; Cortical Predominant+, p \u0026lt; 0.05\u003c/li\u003e\n\u003c/ol\u003e\n\u003cp\u003eMRI Clustering\u003c/p\u003e\n\u003cp\u003eThe overall AD group showed atrophy in the expected medial temporal regions such as hippocampus and amygdala (Fig. \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003eb), which will be referred to as the a \u0026lsquo;typical\u0026rsquo; AD pattern for MRI. Similar to FDG-PET subtypes, clustering revealed a distinction between either a limbic or a cortical pathway in MRI (Fig. \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003ed). The Calinski-Harabasz index peaked at three but was still high at five clusters, the Davies-Bouldin index was lower at five, the Dunn index was high for five albeit plateaued at six before a sharp increase after that and the Silhouette index was highest at three to six clusters (\u003cem\u003eSuppl. Figure\u0026nbsp;7\u003c/em\u003e). Therefore, we chose five clusters as the optimal solution for the MRI model based on these results and considering prior work identifying five biological subtypes. Another reason for choosing a higher cluster solution was based on the lack of sensitivity for finding atypical patterns when implementing a three- and four- cluster solutions. The MRI clustering model was split into five atrophy-based subtypes. In contrast to the FDG-PET subtypes which were limited to cortical and limbic subtypes, the MRI subtypes showed additional \u0026lsquo;minimal\u0026rsquo; versus \u0026lsquo;diffuse\u0026rsquo; atrophy patterns. Similar to the FDG-PET Cortical Predominant subtypes, a Cortical Predominant MRI subtype (19%) showed greater cortical atrophy relative to the hippocampus. The Limbic Predominant subtype (27%) had the opposite pattern with greater atrophy in the hippocampus relative to the cortex. The Minimal subtype (19%) had some atrophy in the hippocampus and amygdala, but very little atrophy compared to cognitively normal individuals in the cortical regions. There were two diffuse atrophy subtypes, one with greater overall atrophy, Diffuse+ (6%), and one with similarly diffuse but less severe atrophy (28%).\u003c/p\u003e\n\u003cp\u003eBased on the inter-modality comparison, the atrophy-based subtypes displayed hypometabolism of differing severity in temporo-parietal and lateral temporal regions often described to be the \u0026lsquo;typical AD\u0026rsquo; pattern in FDG-PET scans (Fig. \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003ef). Compared to the corresponding atrophy maps of the FDG-PET subtypes (Fig. \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003ee), the hypometabolism maps (Fig. \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003ef, \u003cem\u003eSuppl. Figure\u0026nbsp;9\u003c/em\u003e) were more topographically similar to the MRI subtypes when based on visual comparison of w-scores. These corresponding maps showed both more pronounced (higher w-scores) and more widespread hypometabolism in the Minimal and Cortical Predominant subtypes compared to their atrophy maps (Fig. \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003ef, \u003cem\u003eSuppl. Figure\u0026nbsp;9\u003c/em\u003e).\u003c/p\u003e\n\u003cp\u003eRegarding demographic and clinical differences, among the MRI AD subtypes (Table \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003e), the Diffuse subtype had the lowest executive function scores compared to the Minimal and Limbic Predominant subtypes. Diffuse, Diffuse+, and Cortical Predominant had significantly worse executive function scores compared to the Minimal subtype. Minimal and Limbic Predominant subtypes had significantly lower SUVR-based hippocampus-to-cortex ratios to Cortical Predominant. Significant differences were also found in the grey matter volume-based hippocampus-to-cortex ratios: Cortical Predominant had the highest hippocampus-to-cortex ratio. There were no significant differences between the subtypes for the other variables: sex, age, disease duration, age at onset, years of education, \u003cem\u003eAPOE\u003c/em\u003e \u0026epsilon;4 carriers, MMSE, CDR, cognitive measure of memory and language, and CSF biomarkers. Despite not showing a significant difference, Diffuse\u0026thinsp;+\u0026thinsp;had the highest proportion of \u003cem\u003eAPOE\u003c/em\u003e \u0026epsilon;4 carriers (90%) and was the oldest group (79.3 years).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 3: Demographic and clinical characteristics of the MRI AD subtypes.\u003c/strong\u003e\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\" width=\"684\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd width=\"20.437956204379564%\" valign=\"bottom\"\u003e\n \u003cp\u003e\u003cstrong\u003eDemographic \u0026amp; Clinical Characteristics\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"19.416058394160583%\" valign=\"bottom\"\u003e\n \u003cp\u003e\u003cstrong\u003eCortical Predominant\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.94890510948905%\" valign=\"bottom\"\u003e\n \u003cp\u003e\u003cstrong\u003eDiffuse\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.824817518248175%\" valign=\"bottom\"\u003e\n \u003cp\u003e\u003cstrong\u003eLimbic Predominant\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.824817518248175%\" valign=\"bottom\"\u003e\n \u003cp\u003e\u003cstrong\u003eDiffuse+\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.927007299270073%\" valign=\"bottom\"\u003e\n \u003cp\u003e\u003cstrong\u003eMinimal\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.197080291970803%\" valign=\"bottom\"\u003e\n \u003cp\u003e\u003cstrong\u003eCognitively Normal\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"6.423357664233577%\" valign=\"bottom\"\u003e\n \u003cp\u003e\u003cstrong\u003ep values\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"20.437956204379564%\" valign=\"bottom\"\u003e\n \u003cp\u003eN (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"19.416058394160583%\" valign=\"bottom\"\u003e\n \u003cp\u003e35 (19%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.94890510948905%\" valign=\"bottom\"\u003e\n \u003cp\u003e51 (28%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.824817518248175%\" valign=\"bottom\"\u003e\n \u003cp\u003e49 (27%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.824817518248175%\" valign=\"bottom\"\u003e\n \u003cp\u003e10 (6%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.927007299270073%\" valign=\"bottom\"\u003e\n \u003cp\u003e35 (19%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.197080291970803%\" valign=\"bottom\"\u003e\n \u003cp\u003e176\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"6.423357664233577%\" valign=\"bottom\"\u003e\n \u003cp\u003e―\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"20.437956204379564%\" valign=\"bottom\"\u003e\n \u003cp\u003eWomen (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"19.416058394160583%\" valign=\"bottom\"\u003e\n \u003cp\u003e10 (29%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.94890510948905%\" valign=\"bottom\"\u003e\n \u003cp\u003e30 (59%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.824817518248175%\" valign=\"bottom\"\u003e\n \u003cp\u003e22 (45%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.824817518248175%\" valign=\"bottom\"\u003e\n \u003cp\u003e6 (60%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.927007299270073%\" valign=\"bottom\"\u003e\n \u003cp\u003e13 (37%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.197080291970803%\" valign=\"bottom\"\u003e\n \u003cp\u003e84 (48%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"6.423357664233577%\" valign=\"bottom\"\u003e\n \u003cp\u003e0.049\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"20.437956204379564%\" valign=\"bottom\"\u003e\n \u003cp\u003eAge (years)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"19.416058394160583%\" valign=\"bottom\"\u003e\n \u003cp\u003e70 (9.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.94890510948905%\" valign=\"bottom\"\u003e\n \u003cp\u003e74 (8.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.824817518248175%\" valign=\"bottom\"\u003e\n \u003cp\u003e76 (6.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.824817518248175%\" valign=\"bottom\"\u003e\n \u003cp\u003e79 (4.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.927007299270073%\" valign=\"bottom\"\u003e\n \u003cp\u003e73 (6.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.197080291970803%\" valign=\"bottom\"\u003e\n \u003cp\u003e75 (7.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"6.423357664233577%\" valign=\"bottom\"\u003e\n \u003cp\u003e0.004\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"20.437956204379564%\" valign=\"bottom\"\u003e\n \u003cp\u003eDisease Duration (years)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"19.416058394160583%\" valign=\"bottom\"\u003e\n \u003cp\u003e2.2 (2.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.94890510948905%\" valign=\"bottom\"\u003e\n \u003cp\u003e3.4 (2.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.824817518248175%\" valign=\"bottom\"\u003e\n \u003cp\u003e2.5 (2.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.824817518248175%\" valign=\"bottom\"\u003e\n \u003cp\u003e4.4 (3.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.927007299270073%\" valign=\"bottom\"\u003e\n \u003cp\u003e1.9 (2.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.197080291970803%\" valign=\"bottom\"\u003e\n \u003cp\u003e―\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"6.423357664233577%\" valign=\"bottom\"\u003e\n \u003cp\u003e0.002\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"20.437956204379564%\" valign=\"bottom\"\u003e\n \u003cp\u003eAge at Onset (years)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"19.416058394160583%\" valign=\"bottom\"\u003e\n \u003cp\u003e68 (9.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.94890510948905%\" valign=\"bottom\"\u003e\n \u003cp\u003e71 (8.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.824817518248175%\" valign=\"bottom\"\u003e\n \u003cp\u003e74 (6.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.824817518248175%\" valign=\"bottom\"\u003e\n \u003cp\u003e75 (7.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.927007299270073%\" valign=\"bottom\"\u003e\n \u003cp\u003e71 (6.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.197080291970803%\" valign=\"bottom\"\u003e\n \u003cp\u003e―\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"6.423357664233577%\" valign=\"bottom\"\u003e\n \u003cp\u003e0.021\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"20.437956204379564%\" valign=\"bottom\"\u003e\n \u003cp\u003eEducation (years)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"19.416058394160583%\" valign=\"bottom\"\u003e\n \u003cp\u003e15 (3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.94890510948905%\" valign=\"bottom\"\u003e\n \u003cp\u003e15 (2.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.824817518248175%\" valign=\"bottom\"\u003e\n \u003cp\u003e16 (2.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.824817518248175%\" valign=\"bottom\"\u003e\n \u003cp\u003e16 (2.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.927007299270073%\" valign=\"bottom\"\u003e\n \u003cp\u003e16 (2.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.197080291970803%\" valign=\"bottom\"\u003e\n \u003cp\u003e17 (2.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"6.423357664233577%\" valign=\"bottom\"\u003e\n \u003cp\u003e0.197\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"20.437956204379564%\" valign=\"bottom\"\u003e\n \u003cp\u003e\u003cem\u003eAPOE\u003c/em\u003e ɛ4 (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"19.416058394160583%\" valign=\"bottom\"\u003e\n \u003cp\u003e22 (63%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.94890510948905%\" valign=\"bottom\"\u003e\n \u003cp\u003e38 (75%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.824817518248175%\" valign=\"bottom\"\u003e\n \u003cp\u003e31 (63%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.824817518248175%\" valign=\"bottom\"\u003e\n \u003cp\u003e9 (90%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.927007299270073%\" valign=\"bottom\"\u003e\n \u003cp\u003e26 (74%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.197080291970803%\" valign=\"bottom\"\u003e\n \u003cp\u003e34 (19%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"6.423357664233577%\" valign=\"bottom\"\u003e\n \u003cp\u003e0.329\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"20.437956204379564%\" valign=\"bottom\"\u003e\n \u003cp\u003eMMSE\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"19.416058394160583%\" valign=\"bottom\"\u003e\n \u003cp\u003e23 (2.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.94890510948905%\" valign=\"bottom\"\u003e\n \u003cp\u003e23 (2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.824817518248175%\" valign=\"bottom\"\u003e\n \u003cp\u003e23 (2.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.824817518248175%\" valign=\"bottom\"\u003e\n \u003cp\u003e23 (2.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.927007299270073%\" valign=\"bottom\"\u003e\n \u003cp\u003e24 (2.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.197080291970803%\" valign=\"bottom\"\u003e\n \u003cp\u003e29 (1.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"6.423357664233577%\" valign=\"bottom\"\u003e\n \u003cp\u003e0.767\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"20.437956204379564%\" valign=\"bottom\"\u003e\n \u003cp\u003eGlobal CDR\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"19.416058394160583%\" valign=\"bottom\"\u003e\n \u003cp\u003e0.79 (0.25)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.94890510948905%\" valign=\"bottom\"\u003e\n \u003cp\u003e0.83 (0.24)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.824817518248175%\" valign=\"bottom\"\u003e\n \u003cp\u003e0.74 (0.32)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.824817518248175%\" valign=\"bottom\"\u003e\n \u003cp\u003e0.89 (0.22)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.927007299270073%\" valign=\"bottom\"\u003e\n \u003cp\u003e0.72 (0.25)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.197080291970803%\" valign=\"bottom\"\u003e\n \u003cp\u003e0.025 (0.16)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"6.423357664233577%\" valign=\"bottom\"\u003e\n \u003cp\u003e0.121\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"20.437956204379564%\" valign=\"bottom\"\u003e\n \u003cp\u003eADNI-EF\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"19.416058394160583%\" valign=\"bottom\"\u003e\n \u003cp\u003e-0.6 (0.62)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.94890510948905%\" valign=\"bottom\"\u003e\n \u003cp\u003e-0.72 (0.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.824817518248175%\" valign=\"bottom\"\u003e\n \u003cp\u003e-0.32 (0.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.824817518248175%\" valign=\"bottom\"\u003e\n \u003cp\u003e-0.63 (0.72)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.927007299270073%\" valign=\"bottom\"\u003e\n \u003cp\u003e-0.0096 (0.74)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.197080291970803%\" valign=\"bottom\"\u003e\n \u003cp\u003e0.8 (0.47)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"6.423357664233577%\" valign=\"bottom\"\u003e\n \u003cp\u003e\u0026lt;0.001[a][b][c]\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"20.437956204379564%\" valign=\"bottom\"\u003e\n \u003cp\u003eADNI-MEM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"19.416058394160583%\" valign=\"bottom\"\u003e\n \u003cp\u003e-0.78 (0.32)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.94890510948905%\" valign=\"bottom\"\u003e\n \u003cp\u003e-0.91 (0.33)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.824817518248175%\" valign=\"bottom\"\u003e\n \u003cp\u003e-0.78 (0.29)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.824817518248175%\" valign=\"bottom\"\u003e\n \u003cp\u003e-0.85 (0.43)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.927007299270073%\" valign=\"bottom\"\u003e\n \u003cp\u003e-0.65 (0.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.197080291970803%\" valign=\"bottom\"\u003e\n \u003cp\u003e0.89 (0.52)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"6.423357664233577%\" valign=\"bottom\"\u003e\n \u003cp\u003e0.059\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"20.437956204379564%\" valign=\"bottom\"\u003e\n \u003cp\u003eADNI-LAN\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"19.416058394160583%\" valign=\"bottom\"\u003e\n \u003cp\u003e-0.22 (0.41)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.94890510948905%\" valign=\"bottom\"\u003e\n \u003cp\u003e-0.4 (0.52)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.824817518248175%\" valign=\"bottom\"\u003e\n \u003cp\u003e-0.22 (0.55)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.824817518248175%\" valign=\"bottom\"\u003e\n \u003cp\u003e-0.13 (0.67)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.927007299270073%\" valign=\"bottom\"\u003e\n \u003cp\u003e0.11 (0.68)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.197080291970803%\" valign=\"bottom\"\u003e\n \u003cp\u003e0.82 (0.52)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"6.423357664233577%\" valign=\"bottom\"\u003e\n \u003cp\u003e0.007\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"20.437956204379564%\" valign=\"bottom\"\u003e\n \u003cp\u003eCSF t-tau (pg/ml)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"19.416058394160583%\" valign=\"bottom\"\u003e\n \u003cp\u003e390 (164)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.94890510948905%\" valign=\"bottom\"\u003e\n \u003cp\u003e398 (145)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.824817518248175%\" valign=\"bottom\"\u003e\n \u003cp\u003e380 (141)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.824817518248175%\" valign=\"bottom\"\u003e\n \u003cp\u003e280 (77)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.927007299270073%\" valign=\"bottom\"\u003e\n \u003cp\u003e377 (161)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.197080291970803%\" valign=\"bottom\"\u003e\n \u003cp\u003e223 (70)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"6.423357664233577%\" valign=\"bottom\"\u003e\n \u003cp\u003e0.065\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"20.437956204379564%\" valign=\"bottom\"\u003e\n \u003cp\u003eCSF p-tau (pg/ml)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"19.416058394160583%\" valign=\"bottom\"\u003e\n \u003cp\u003e39 (17)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.94890510948905%\" valign=\"bottom\"\u003e\n \u003cp\u003e40 (16)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.824817518248175%\" valign=\"bottom\"\u003e\n \u003cp\u003e37 (14)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.824817518248175%\" valign=\"bottom\"\u003e\n \u003cp\u003e26 (7.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.927007299270073%\" valign=\"bottom\"\u003e\n \u003cp\u003e39 (18)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.197080291970803%\" valign=\"bottom\"\u003e\n \u003cp\u003e20 (6.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"6.423357664233577%\" valign=\"bottom\"\u003e\n \u003cp\u003e0.041\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"20.437956204379564%\" valign=\"bottom\"\u003e\n \u003cp\u003eCSF A\u0026beta;\u0026nbsp;(pg/ml)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"19.416058394160583%\" valign=\"bottom\"\u003e\n \u003cp\u003e564 (168)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.94890510948905%\" valign=\"bottom\"\u003e\n \u003cp\u003e575 (159)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.824817518248175%\" valign=\"bottom\"\u003e\n \u003cp\u003e661 (245)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.824817518248175%\" valign=\"bottom\"\u003e\n \u003cp\u003e536 (128)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.927007299270073%\" valign=\"bottom\"\u003e\n \u003cp\u003e551 (181)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.197080291970803%\" valign=\"bottom\"\u003e\n \u003cp\u003e1204 (333)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"6.423357664233577%\" valign=\"bottom\"\u003e\n \u003cp\u003e0.425\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"20.437956204379564%\" valign=\"bottom\"\u003e\n \u003cp\u003eSUVR-based Hippocampus-to-cortex ratio\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"19.416058394160583%\" valign=\"bottom\"\u003e\n \u003cp\u003e0.35 (0.047)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.94890510948905%\" valign=\"bottom\"\u003e\n \u003cp\u003e0.34 (0.048)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.824817518248175%\" valign=\"bottom\"\u003e\n \u003cp\u003e0.32 (0.038)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.824817518248175%\" valign=\"bottom\"\u003e\n \u003cp\u003e0.34 (0.053)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.927007299270073%\" valign=\"bottom\"\u003e\n \u003cp\u003e0.31 (0.043)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.197080291970803%\" valign=\"bottom\"\u003e\n \u003cp\u003e0.32 (0.029)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"6.423357664233577%\" valign=\"bottom\"\u003e\n \u003cp\u003e\u0026lt;0.001[d][e]\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"20.437956204379564%\" valign=\"bottom\"\u003e\n \u003cp\u003eGrey Matter Volume-based Hippocampus-to-cortex ratio\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"19.416058394160583%\" valign=\"bottom\"\u003e\n \u003cp\u003e0.18 (0.021)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.94890510948905%\" valign=\"bottom\"\u003e\n \u003cp\u003e0.17 (0.026)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.824817518248175%\" valign=\"bottom\"\u003e\n \u003cp\u003e0.15 (0.018)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.824817518248175%\" valign=\"bottom\"\u003e\n \u003cp\u003e0.16 (0.027)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.927007299270073%\" valign=\"bottom\"\u003e\n \u003cp\u003e0.15 (0.022)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.197080291970803%\" valign=\"bottom\"\u003e\n \u003cp\u003e0.18 (0.021)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"6.423357664233577%\" valign=\"bottom\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003csup\u003ep\u003c/sup\u003e\u003csup\u003e[f][g]\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"20.437956204379564%\" valign=\"bottom\"\u003e\n \u003cp\u003eTotal Grey Matter Volume (mm\u0026sup3;)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"19.416058394160583%\" valign=\"bottom\"\u003e\n \u003cp\u003e561400 (17261)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.94890510948905%\" valign=\"bottom\"\u003e\n \u003cp\u003e518539 (12671)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.824817518248175%\" valign=\"bottom\"\u003e\n \u003cp\u003e552539 (15299)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.824817518248175%\" valign=\"bottom\"\u003e\n \u003cp\u003e478411 (11842)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.927007299270073%\" valign=\"bottom\"\u003e\n \u003cp\u003e606360 (16474)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.197080291970803%\" valign=\"bottom\"\u003e\n \u003cp\u003e580718 (40407)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"6.423357664233577%\" valign=\"bottom\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003csup\u003em\u003c/sup\u003e\u003csup\u003es\u003c/sup\u003e\u003csup\u003e[h]\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"20.437956204379564%\" valign=\"bottom\"\u003e\n \u003cp\u003eTotal Average Cortical Uptake (SUVR)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"19.416058394160583%\" valign=\"bottom\"\u003e\n \u003cp\u003e1.6 (0.16)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.94890510948905%\" valign=\"bottom\"\u003e\n \u003cp\u003e1.5 (0.14)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.824817518248175%\" valign=\"bottom\"\u003e\n \u003cp\u003e1.6 (0.15)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.824817518248175%\" valign=\"bottom\"\u003e\n \u003cp\u003e1.4 (0.081)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.927007299270073%\" valign=\"bottom\"\u003e\n \u003cp\u003e1.7 (0.21)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.197080291970803%\" valign=\"bottom\"\u003e\n \u003cp\u003e1.7 (0.19)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"6.423357664233577%\" valign=\"bottom\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003csup\u003ea\u003c/sup\u003e\u003csup\u003e[i][j][k]\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u0026nbsp;The values shown in the table are the means with standard deviations in brackets except for number of individuals, women and \u003cem\u003eAPOE\u003c/em\u003e \u0026epsilon;4 for which the percentages are provided. \u0026Chi;\u003csup\u003e2\u003c/sup\u003e tests were used for categorical variables and Kruskal-Wallis for continuous variables and were corrected for multiple comparison with the Holm-\u0026Scaron;id\u0026aacute;k method. Footnotes indicate cases where p values were significant in the post hoc pairwise comparisons across AD subtypes, p \u0026lt; 0.05. The CN group data is displayed for reference. Abbreviations: CP: Cortical Predominant, LP: Limbic Predominant, Min: Minimal atrophy, Dif: Diffuse, Dif+: Diffuse+, CN: cognitively normal individuals, \u003cem\u003eAPOE\u003c/em\u003e \u0026epsilon;4: apolipoprotein E \u0026epsilon;4 allele, MMSE: Mini Mental State Examination, CDR: The Clinical Dementia Rating Scale, ADNI-EF Alzheimer\u0026rsquo;s Disease Neuroimaging Initiative executive function composite score, ADNI-MEM: Alzheimer\u0026rsquo;s Disease Neuroimaging Initiative memory composite score, ADNI-LAN: Alzheimer\u0026rsquo;s Disease Neuroimaging Initiative language composite score, CSF t-tau: total tau, CSF p-tau: phosphorylated tau, CSF A\u0026beta;: amyloid-beta 1-42 peptide.\u003c/p\u003e\n\u003col style=\"list-style-type: lower-alpha;\"\u003e\n \u003cli\u003eDiffuse \u0026lt; Minimal, p \u0026lt; 0.05\u003c/li\u003e\n \u003cli\u003eCortical Predominant \u0026lt; Minimal, p \u0026lt; 0.05\u003c/li\u003e\n \u003cli\u003eDiffuse \u0026lt; Limbic Predominant, p \u0026lt; 0.05\u003c/li\u003e\n \u003cli\u003eMinimal \u0026lt; Cortical Predominant, p \u0026lt; 0.05\u003c/li\u003e\n \u003cli\u003eLimbic Predominant \u0026lt; Cortical Predominant, p \u0026lt; 0.05\u003c/li\u003e\n \u003cli\u003eDiffuse \u0026lt; Cortical Predominant, p \u0026lt; 0.05\u003c/li\u003e\n \u003cli\u003eMinimal \u0026lt; Diffuse, p \u0026lt; 0.05\u003c/li\u003e\n \u003cli\u003eLimbic Predominant \u0026lt; Minimal, p \u0026lt; 0.05\u003c/li\u003e\n \u003cli\u003eDiffuse+ \u0026lt; Minimal, p \u0026lt; 0.05\u003c/li\u003e\n \u003cli\u003eDiffuse+ \u0026lt; Cortical Predominant, p \u0026lt; 0.05\u003c/li\u003e\n \u003cli\u003eDiffuse+ \u0026lt; Limbic Predominant, p \u0026lt; 0.05\u003c/li\u003e\n\u003c/ol\u003e\n\u003cp\u003eIndividual-level Subtype Allocations\u003c/p\u003e\n\u003cp\u003eTo assess the consistency between the two modalities in subtype assignments, the subtype categorizations for individuals were compared across both FDG-PET and MRI (Fig. \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003e). We propose that the cortical subtypes and limbic subtypes are most similar between the FDG-PET and MRI subtypes. Namely, the FDG-PET Cortical Predominant, Cortical Predominant posterior and Cortical Predominant\u0026thinsp;+\u0026thinsp;subtypes are equivalent with MRI Cortical Predominant, Diffuse or Diffuse\u0026thinsp;+\u0026thinsp;patterns topographically. Whereas FDG-PET Limbic Predominant and Limbic Predominant frontal are equivalent to MRI Limbic Predominant. As a Minimal pattern was only found in MRI, we do not think that this subtype has an equivalent in FDG-PET.\u003c/p\u003e\n\u003cp\u003eThe agreement between the FDG-PET and MRI subtype allocations was low as this was less than 50%. Although the compared subtypes showed similar topographies of neurodegeneration (i.e., cortical/limbic predominant hypometabolism and atrophy, respectively) they did not match at the individual-level. All possible combinations of allocated FDG-PET and MRI subtypes of varying percentages were found (Fig. \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003ea, b). AD individuals classified into the FDG-PET Cortical Predominant subtype matched most with the MRI Limbic Predominant (33.3%) and MRI Cortical Predominant (24.6%) subtypes. By contrast, individuals classified as FDG-PET Cortical Predominant\u0026thinsp;+\u0026thinsp;matched best with MRI Cortical Predominant (35%) and Diffuse (35%). FDG-PET Limbic Predominant best matched with three MRI subtypes: Limbic Predominant (28.1%), Minimal (26.6%) and Diffuse (25%). FDG-PET Limbic Predominant frontal best matched with MRI Diffuse (52.2%). FDG-PET Cortical Predominant posterior matched best with MRI Limbic Predominant (37.5%).\u003c/p\u003e\n\u003cp\u003eData-driven versus Hypothesis-driven Subtyping\u003c/p\u003e\n\u003cp\u003eClustering-based and prior hypothesis-based subtypes were compared within the framework of typicality and severity\u003csup\u003e2\u003c/sup\u003e in both FDG-PET and MRI (Fig. \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003e). Within each modality, data-driven and hypothesis-driven subtypes overlapped with each other reasonably well for most subtypes (Fig. \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003ea, b). The agreement between MRI data-driven and MRI hypothesis-driven limbic predominant subtypes (55.6%) was better than that between cortical predominant subtypes (30%) (Fig. \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003eb). Contrarily, the agreement between FDG-PET data-driven and FDG-PET hypothesis-driven cortical predominant subtypes (90%) was better than that between limbic predominant subtypes (82%) (Fig. \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003ea\u003cem\u003e)\u003c/em\u003e.\u003c/p\u003e\n\u003cp\u003eAssociation between typicality and severity by modality differed. Correlation between typicality and severity (Table \u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003e) was significant in FDG-PET (r\u003csup\u003e2\u003c/sup\u003e\u0026thinsp;=\u0026thinsp;0.25), but not in MRI. The severity measures (r\u003csup\u003e2\u003c/sup\u003e\u0026thinsp;=\u0026thinsp;0.15) in FDG-PET and MRI were more strongly associated with each other than typicality measures (r\u003csup\u003e2\u003c/sup\u003e\u0026thinsp;=\u0026thinsp;0.02). Additionally, FDG-PET subtypes are more separable across the typicality axis than MRI, this is evident when comparing averaged normalised values of typicality and severity for each subtype (Fig. \u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003e). Whereas for MRI subtypes, there was a clearer split of the along the severity axis.\u003c/p\u003e\n\u003cdiv class=\"gridtable\"\u003e\u0026nbsp;\u003ctable id=\"Tab4\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 4\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eCorrelations between Typicality \u0026amp; Severity in FDG-PET and MRI.\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eModel\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eR\u0026sup2;\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eP value\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eFDG-PET Typicality and FDG-PET Severity\u003c/p\u003e\n \u003cp\u003e(Fig. \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003ea)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.25\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001**\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eMRI Typicality and MRI Severity (Fig. \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003eb)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.0011\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.53\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eFDG-PET Severity and MRI Severity\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.15\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001**\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eFDG-PET Typicality and MRI Typicality\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.02\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.0073*\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003ctfoot\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\"\u003ePearson correlations between measures of \u0026lsquo;typicality\u0026rsquo; and \u0026lsquo;severity\u0026rsquo; in both modalities.\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\"\u003e* P\u0026thinsp;\u0026lt;\u0026thinsp;0.05.\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\"\u003e** \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001.\u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tfoot\u003e\n \u003c/table\u003e\n\u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eThis study investigated cross-modality Alzheimer\u0026rsquo;s disease subtypes by applying data-driven subtyping models to regional FDG-PET and MRI data. To our knowledge, this is the first study to implement identical data-driven models to concurrent FDG-PET and MRI data to identify and compare modality-specific neurodegeneration-based AD subtypes. Despite FDG-PET and MRI both being interchangeable measures of neurodegeneration within the ATN framework\u003csup\u003e45,46\u003c/sup\u003e, our findings show that the respective neurodegeneration subtypes differed across modalities and within individuals.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eUsing the same data-driven model, cortical and limbic AD subtypes were independently identified in both FDG-PET and MRI methodology. At a group-level, the expected pattern of temporo-parietal hypometabolism\u003csup\u003e42\u0026ndash;46\u003c/sup\u003e was found in the AD group. However, FDG-PET subtypes showed distinct patterns of hypometabolism: Cortical Predominant, Cortical Predominant+, Cortical Predominant posterior, Limbic Predominant and Limbic Predominant frontal. Our findings resemble findings of the only previous study addressing FDG subtypes in AD\u003csup\u003e11\u003c/sup\u003e. In terms of topography, Cortical Predominant, Cortical Predominant posterior and Limbic Predominant subtypes in our study are closest to the Typical subtype (hypometabolism in cortical and limbic regions) identified by Levin et al. although with differing clinical characteristics\u003csup\u003e11\u003c/sup\u003e. The differences in findings could possibly be explained by methodological (clustering method, regions used for clustering, etc.) and sample differences. Two subtypes that were identified included the Cortical Predominant+ (younger, fewer \u003cem\u003eAPOE\u003c/em\u003e \u0026epsilon;4 carriers, executive function impairment) and Limbic Predominant frontal (older age, more \u003cem\u003eAPOE\u003c/em\u003e \u0026epsilon;4 carriers), which align well with the Cortical Predominant and Limbic Predominant subtypes, respectively, identified by Levin et al\u003csup\u003e11\u003c/sup\u003e. Together, these two studies indicate the clear presence of distinct cortical- and limbic-predominant profiles of hypometabolism in AD.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe MRI pattern of the whole AD group showed the expected pattern of medial temporal and hippocampal atrophy\u003csup\u003e47\u0026ndash;50\u003c/sup\u003e. However, the MRI AD subtypes showed five distinct patterns: Cortical Predominant, Limbic Predominant, Minimal atrophy, Diffuse, Diffuse+. Our data-driven MRI subtypes had similar\u0026nbsp;percentages\u0026nbsp;and clinical presentation to what has been found previously\u003csup\u003e2,9,35,51,52\u003c/sup\u003e. In accordance with previous studies, Limbic Predominant had focal limbic atrophy and later age of onset. Cortical Predominant resembles the \u0026lsquo;Hippocampal Sparing\u0026rsquo; MRI subtype previously described, with greater atrophy in the cortical relative to the limbic regions, as well as greater executive function impairment, higher proportion of men, and younger age. The two Diffuse subtypes in this study resemble Typical and Diffuse subtypes identified previously (widespread atrophy, older, worse memory scores). At the other end of the severity dimension, Minimal atrophy subtype resembled prior Minimal subtypes topographically and clinically (shortest disease duration, highest MMSE).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWe did not observe an overlap in terms of topography nor demographics, despite identifying cortical and limbic subtypes in both FDG-PET and MRI. As the \u0026lsquo;typical\u0026rsquo; AD pattern for each modality differs, this can be seen in the subtypes\u0026rsquo; modality-specific neurodegeneration. Key regions show neurodegeneration at varying levels of severity, such as posterior cingulate cortex in FDG-PET (\u003cem\u003eFig. 1c\u003c/em\u003e) and medial temporal lobe and hippocampus for MRI (\u003cem\u003eFig. 1d\u003c/em\u003e). FDG-PET subtypes show a clearer cortical pathway compared to the MRI subtypes, which are more susceptible to more limbic neurodegeneration. These results are in line with the common AD patterns in these two modalities: neocortical hypometabolism in FDG-PET\u003csup\u003e43,45,53\u003c/sup\u003e and medial temporal atrophy in MRI\u003csup\u003e47,49\u003c/sup\u003e. While the regional atrophy does not mimic hypometabolism in FDG-PET subtypes (\u003cem\u003eFig. 1c, e\u003c/em\u003e), regional hypometabolism mirrors atrophy in MRI subtypes more closely (\u003cem\u003eFig. 1d, f\u003c/em\u003e). For example, cortical predominant FDG-PET subtypes with limited hippocampal hypometabolism were found to show considerable limbic (hippocampal) atrophy. This finding is congruent with the well-established evidence that higher cortical hypometabolism is closely associated with higher hippocampal atrophy\u003csup\u003e17,19,54\u003c/sup\u003e. \u0026nbsp;Our study goes a step further to demonstrate that in contrast, a cortical-predominant MRI subtype with limited hippocampal atrophy also shows cortical-predominant hypometabolism with limited hippocampal hypometabolism. A previous study assessed FDG-PET severity within AD typical cortical regions in hypothesis-driven MRI subtypes using the so-called \u0026lsquo;Hypometabolic Convergence Index\u0026rsquo;\u003csup\u003e7\u003c/sup\u003e. Hippocampal Sparing had highest values of this index compared to Limbic Predominant and Typical AD subtypes, highlighting the overlap in cortical hypometabolism and atrophy. Although our FDG-PET subtypes had similar patterns of cortical hypometabolism and atrophy, this was not the case when assessing deeper structures (\u003cem\u003eFig 1 e, f)\u003c/em\u003e. Thus, FDG-PET and MRI do not necessarily share a bidirectional relationship in capturing hippocampal neurodegeneration.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eInterestingly when assessing how individuals were subtyped in both modalities, a remarkably low classification correspondence between equivalent subtypes (i.e., cortical/limbic predominant) in the different modalities was found (\u003cem\u003eFig. 2a\u003c/em\u003e). Individuals classified as Cortical hypometabolism subtypes were not always classified as cortical predominant atrophy subtypes and even matched with Limbic Predominant atrophy patterns. Limbic hypometabolism subtypes agreed with different levels of atrophy severity from Minimal to Diffuse. The \u0026lsquo;disconnection hypothesis\u0026rsquo; has been proposed in AD to explain concurrent retrosplenial cortex hypometabolism and medial temporal atrophy\u003csup\u003e15,19,55\u003c/sup\u003e. It has been proposed that local and/or distant atrophy result in downstream hypometabolism in the parietal and retrosplenial cortex\u003csup\u003e19,56\u003c/sup\u003e. In our corresponding neurodegeneration maps (\u003cem\u003eFig. 1e, f\u003c/em\u003e), association between levels of isthmus cingulate (retrosplenial cortex) hypometabolism relative to medial temporal lobe atrophy can be appreciated. FDG-PET subtypes; Cortical Predominant and Cortical Predominant posterior had greatest isthmus cingulate hypometabolism and consequently displayed greater medial temporal atrophy than the other subtypes (\u003cem\u003eFig. 1c, e\u003c/em\u003e). Conversely, Cortical Predominant+ hypometabolism subtype had greatest atrophy in cingulate/precuneus and least in hippocampus compared to the other hypometabolism subtypes. This is in concordance with prior findings in early onset AD patients who typically show more posterior cortical atrophy compared to atrophy in the hippocampus\u003csup\u003e53,57\u003c/sup\u003e. MRI subtypes\u0026rsquo; corresponding hypometabolism maps all showed isthmus cingulate involvement even in the Minimal atrophy subtype (\u003cem\u003eFig. 1d, f\u003c/em\u003e). This finding resembles prior findings of a minimal hippocampal and posterior cingulate atrophy AD pattern that showed significant hypometabolism\u003csup\u003e53\u003c/sup\u003e. This could be indicative of FDG-PET being an earlier marker of neurodegeneration than MRI\u003csup\u003e58\u0026ndash;61\u003c/sup\u003e. \u0026nbsp;Moreover, neuropathological data and its association with specific neurodegeneration patterns could provide the link between FDG-PET and MRI. Potential copathologies are often present in combination with AD\u003csup\u003e62\u003c/sup\u003e therefore more analysis is needed especially within subtypes\u003csup\u003e52\u003c/sup\u003e and across different imaging modalities\u003csup\u003e63\u003c/sup\u003e. Different pathologies have been linked to specific patterns of atrophy and hypometabolism\u003csup\u003e64\u0026ndash;66\u003c/sup\u003e. These copathologies could also explain the divergence between atrophy and metabolism patterns.\u003c/p\u003e\n\u003cp\u003eMeasures such as the hippocampus-to-cortex ratio based on neuropathological data\u003csup\u003e5\u003c/sup\u003e have been applied successfully using grey matter volumes from MRI\u003csup\u003e6,7\u003c/sup\u003e. In this study, we compared such measures with findings from our data-driven approach\u003csup\u003e2\u003c/sup\u003e. Our findings showed that there was agreement between hypothesis-driven and data-driven approaches. Interestingly, FDG-PET Cortical Predominant+ also had higher hippocampus-to-cortex ratio when using grey matter volumes similar to the finding in Levin et al., 2021\u003csup\u003e11\u003c/sup\u003e. The data-driven subtypes additionally capture the second axis of \u0026lsquo;severity\u0026rsquo;, as Minimal and Diffuse subtypes were found in the MRI. It can also be argued that our FDG-PET subtypes show different levels of \u0026lsquo;severity\u0026rsquo; as Limbic Predominant frontal and Cortical Predominant+ show greater overall hypometabolism than the other limbic and cortical subtypes. The contrast between the data-driven and hypothesis-based methods highlights the issue with harmonisation of methods for subtyping in AD\u003csup\u003e4\u003c/sup\u003e. In general, hypothesis-driven hippocampus-to-cortex ratios work well using FDG-PET SUVRs and match well with data-driven subtypes. However, as this is the first study to test these two methods using FDG-PET SUVRs this would need to be further validated. Typicality and severity were correlated in FDG-PET indicating that cortical subtypes are the more severe compared to the limbic subtypes. Whereas, typicality and severity were not correlated in MRI, showing that the link is different across the two modalities (\u003cem\u003eTable 4\u003c/em\u003e). These findings could indicate that there are more complex mechanisms at play, such as copathologies\u003csup\u003e64,66\u003c/sup\u003e as neither of these measures are specific to AD.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eConceptually, FDG-PET and MRI subtypes found in our study can be defined along the \u0026lsquo;typicality\u0026rsquo; and \u0026lsquo;severity\u0026apos; axes (\u003cem\u003eFig. 4\u003c/em\u003e). Typicality splits the two modality-specific subtypes into either a limbic or cortical pattern. Severity splits the extremes found in the MRI subtypes of limited to widespread atrophy. In contrast, there was a greater split along the typicality axis for FDG-PET subtypes between cortical and limbic hypometabolism. In this study, we propose that subtypes do not always lie orthogonal along these two axes and are often a combination of both. Namely, we found that FDG-PET Limbic Predominant subtype is left of the centre towards limbic along the typicality axis but is positioned lower on the severity axis compared to Limbic Predominant frontal. The same applies to the Cortical Predominant+ and Cortical Predominant subtypes which have widespread hypometabolism. The severity for the current FDG-PET subtypes is backed by frontal hypometabolism being indicative of a later stage of AD\u003csup\u003e67\u003c/sup\u003e. The complexity of the relationship of the two modalities is reflected in this figure and underlying mismatches between the two modalities have been highlighted throughout this study.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThis study has various limitations, one is its cross-sectional design. To concretise the current findings, longitudinal studies are needed to be able to label these patterns as subtypes. Longitudinal clustering has been performed in AD to investigate heterogeneity of topographical differences in MRI\u003csup\u003e9,68\u003c/sup\u003e. In addition, tau- and amyloid-PET studies have used cross-sectional data with a probabilistic model for predicting disease subtype and stage\u003csup\u003e69,70\u003c/sup\u003e. Future research should explore the progression of glucose metabolism across these subtypes in relation to other imaging modalities. Additionally, testing the models and classification of the subtypes in clinical cohorts for external validity is important. Another limitation is that the individuals in this study likely have mixed pathologies that cannot be detected through neuroimaging alone. Utilising neuropathological data in combination with in vivo data should be a focus in future studies.\u003c/p\u003e\n\u003cp\u003eUltimately, the current study identified data-driven cortical and limbic AD subtypes from FDG-PET and MRI scans. These data-driven subtypes overlapped well with hypothesis-driven methods, validating our findings. Although, the main finding was that structure does not always reflect function when assessing corresponding patterns of neurodegeneration in these subtypes. Cortical and limbic subtypes did not overlap in the two modalities in terms of individual subtype allocation. These subtypes lie along \u0026lsquo;severity\u0026rsquo; and \u0026lsquo;typicality\u0026rsquo; axes distinctly across modalities as shown in our conceptual figure (\u003cem\u003eFig. 4\u003c/em\u003e). Copathologies may contribute to the divergence in FDG-PET and MRI subtype patterns. In conclusion, the current findings highlight the need for a multimodal perspective for understanding the complex biological AD mechanisms.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eData availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe data used in this study are available from the corresponding authors on request. Raw data are available from the LONI database (https://ida.loni.usc.edu).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe would like to thank the Strategic Research Programme in Neuroscience at Karolinska Institutet (StratNeuro), the Swedish Research Council (VR) 2016-02282, 2021-01861, the Center for Innovative Medicine (CIMED) FoUI-954459, FoUI-975174, the regional agreement on medical training and clinical research (ALF) between Stockholm County Council and Karolinska Institutet FoUI-952838, FoUI-954893, The Swedish Brain Foundation (Hjärnfonden) FO2021-0119, FO2022-0084, The Swedish Alzheimer's Foundation (Alzheimerfonden) AF-967495, AF-980387, The Swedish Parkinson's foundation (Parkinsonfonden) 1443/2022, 1521/23 King Gustaf V:s and Queen Victorias Foundation 20221213_064354, and Olle Engkvists Foundation (Olle Engkvists Stiftelse) \u003cstrong\u003e186-0660, 224-0069\u0026nbsp;\u003c/strong\u003eas well as Birgitta and Sten Westerberg for additional financial support.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eDF receives funding from the Swedish Research Council (Vetenskapsrådet, grant 2022-00916), the Center for Innovative Medicine (CIMED, grants 20200505 and FoUI-988826), the regional agreement on medical training and clinical research of Stockholm Region (ALF Medicine, grants FoUI-962240 and FoUI-987534), the Swedish Brain Foundation (Hjärnfonden FO2023-0261, FO2022-0175, FO2021-0131), the Swedish Alzheimer Foundation (Alzheimerfonden AF-968032, AF-980580, AF-994058), the Swedish Dementia Foundation (Demensfonden), the Gamla Tjänarinnor Foundation, the Gun och Bertil Stohnes Foundation, Funding for Research from Karolinska Institutet, Neurofonden, and the Foundation for Geriatric Diseases at Karolinska Institutet, as well as contributions from private bequests. MJG is supported by the ‘Miguel Servet’ program (CP19/00031) of the Instituto de Salud Carlos III – Fondo Europeo de Desarrollo Regional (ISCIII-FEDER).\u0026nbsp;The ADNI is funded by the National Institute on Aging, by the National Institute of Biomedical Imaging and Bioengineering, and through generous contributions from the following: Alzheimer’s Association; Alzheimer’s Drug Discovery Foundation; BioClinica, Inc.; Biogen Idec Inc.; Bristol-Myers Squibb Company; Eisai Inc.; Elan Pharmaceuticals, Inc.; Eli Lilly and Company; F. Hoffmann-La Roche Ltd and its affiliated company Genentech, Inc.; GE Healthcare; Innogenetics, N.V.; IXICO Ltd.; Janssen Alzheimer Immunotherapy Research \u0026amp; Development, LLC; Johnson \u0026amp; Johnson Pharmaceutical Re- search \u0026amp; Development LLC; Medpace, Inc.; Merck \u0026amp; Co., Inc.; Meso Scale Diagnostics, LLC; NeuroRx Research; Novartis Pharmaceuticals Corporation; Pfizer Inc.; Piramal Imaging; Servier; Synarc Inc.; and Takeda Pharmaceutical Company. The Canadian Institutes of Health Research is providing funds to support ADNI clinical sites in Canada. Private sector contributions are facilitated by the Foundation for the NIH (fnih.org). The grantee organization is the Northern California Institute for Research and Education, and the study is coordinated by the Alzheimer’s Disease Cooperative Study at the University of California, San Diego. ADNI data are disseminated by the Laboratory of Neuro Imaging at the University of California, Los Angeles.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eDF consults for BioArctic and has received honoraria from Esteve.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003e2023 Alzheimer\u0026rsquo;s disease facts and figures (2023) Alzheimers Dement Published online March 14:alz13016. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/alz.13016\u003c/span\u003e\u003cspan address=\"10.1002/alz.13016\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFerreira D, Nordberg A, Westman E (2020) Biological subtypes of Alzheimer disease: A systematic review and meta-analysis. Neurology 94(10):436\u0026ndash;448. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1212/WNL.0000000000009058\u003c/span\u003e\u003cspan address=\"10.1212/WNL.0000000000009058\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJellinger KA (2021) Pathobiological Subtypes of Alzheimer Disease. Dement Geriatr Cogn Disord 49(4):321\u0026ndash;333. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1159/000508625\u003c/span\u003e\u003cspan address=\"10.1159/000508625\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMohanty R, M\u0026aring;rtensson G, Poulakis K et al (2020) Comparison of subtyping methods for neuroimaging studies in Alzheimer\u0026rsquo;s disease: a call for harmonization. Brain Commun 2(2):fcaa192. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1093/braincomms/fcaa192\u003c/span\u003e\u003cspan address=\"10.1093/braincomms/fcaa192\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMurray ME, Graff-Radford NR, Ross OA, Petersen RC, Duara R, Dickson DW (2011) Neuropathologically defined subtypes of Alzheimer\u0026rsquo;s disease with distinct clinical characteristics: A retrospective study. Lancet Neurol 10(9):785\u0026ndash;796. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/S1474-4422(11)70156-9\u003c/span\u003e\u003cspan address=\"10.1016/S1474-4422(11)70156-9\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWhitwell JL, Dickson DW, Murray ME et al (2012) Neuroimaging correlates of pathologically defined subtypes of Alzheimer\u0026rsquo;s disease: a case-control study. Lancet Neurol 11(10):868\u0026ndash;877. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/S1474-4422(12)70200-4\u003c/span\u003e\u003cspan address=\"10.1016/S1474-4422(12)70200-4\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRisacher SL, Anderson WH, Charil A et al (2017) Alzheimer disease brain atrophy subtypes are associated with cognition and rate of decline. Neurology 89(21):2176\u0026ndash;2186. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1212/WNL.0000000000004670\u003c/span\u003e\u003cspan address=\"10.1212/WNL.0000000000004670\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJellinger KA (2012) Neuropathological subtypes of Alzheimer\u0026rsquo;s disease. Acta Neuropathol (Berl) 123(1):153\u0026ndash;154. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s00401-011-0889-9\u003c/span\u003e\u003cspan address=\"10.1007/s00401-011-0889-9\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePoulakis K, Pereira JB, Muehlboeck JS et al (2022) Multi-cohort and longitudinal Bayesian clustering study of stage and subtype in Alzheimer\u0026rsquo;s disease. Nat Commun 13(1):4566. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41467-022-32202-6\u003c/span\u003e\u003cspan address=\"10.1038/s41467-022-32202-6\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFerreira D, Mohanty R, Murray ME, Nordberg A, Kantarci K, Westman E (2022) The hippocampal sparing subtype of Alzheimer\u0026rsquo;s disease assessed in neuropathology and in vivo tau positron emission tomography: a systematic review. Acta Neuropathol Commun 10(1):166. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/s40478-022-01471-z\u003c/span\u003e\u003cspan address=\"10.1186/s40478-022-01471-z\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLevin F, Ferreira D, Lange C et al (2021) Data-driven FDG-PET subtypes of Alzheimer\u0026rsquo;s disease-related neurodegeneration. Alzheimers Res Ther 13(1):49. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/s13195-021-00785-9\u003c/span\u003e\u003cspan address=\"10.1186/s13195-021-00785-9\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCaminiti SP, De Francesco S, Tondo G et al FDG-PET markers of heterogeneity and different risk of progression in amnestic MCI. \u003cem\u003eAlzheimers Dement\u003c/em\u003e. 2023;n/a(n/a). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/alz.13385\u003c/span\u003e\u003cspan address=\"10.1002/alz.13385\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eIshii K, Kawachi T, Sasaki H et al (2005) Voxel-Based Morphometric Comparison Between Early- and Late-Onset Mild Alzheimer\u0026rsquo;s Disease and Assessment of Diagnostic Performance of Z Score Images. AJNR Am J Neuroradiol 26(2):333\u0026ndash;340\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTeune LK, Bartels AL, de Jong BM et al (2010) Typical cerebral metabolic patterns in neurodegenerative brain diseases. Mov Disord 25(14):2395\u0026ndash;2404. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/mds.23291\u003c/span\u003e\u003cspan address=\"10.1002/mds.23291\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGrothe MJ, Teipel SJ, for the Alzheimer\u0026rsquo;s Disease Neuroimaging Initiative (2016) Spatial patterns of atrophy, hypometabolism, and amyloid deposition in Alzheimer\u0026rsquo;s disease correspond to dissociable functional brain networks: Network-Specificity of AD Pathology. Hum Brain Mapp 37(1):35\u0026ndash;53. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/hbm.23018\u003c/span\u003e\u003cspan address=\"10.1002/hbm.23018\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCh\u0026eacute;telat G, Desgranges B, Landeau B et al (2008) Direct voxel-based comparison between grey matter hypometabolism and atrophy in Alzheimer\u0026rsquo;s disease. Brain 131(1):60\u0026ndash;71. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1093/brain/awm288\u003c/span\u003e\u003cspan address=\"10.1093/brain/awm288\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLa Joie R, Perrotin A, Barr\u0026eacute; L et al (2012) Region-Specific Hierarchy between Atrophy, Hypometabolism, and β-Amyloid (Aβ) Load in Alzheimer\u0026rsquo;s Disease Dementia. J Neurosci 32(46):16265\u0026ndash;16273. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1523/JNEUROSCI.2170-12.2012\u003c/span\u003e\u003cspan address=\"10.1523/JNEUROSCI.2170-12.2012\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDukart J, Kherif F, Mueller K et al (2013) Generative FDG-PET and MRI Model of Aging and Disease Progression in Alzheimer\u0026rsquo;s Disease. PLoS Comput Biol 9(4):e1002987. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1371/journal.pcbi.1002987\u003c/span\u003e\u003cspan address=\"10.1371/journal.pcbi.1002987\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eStrom A, Iaccarino L, Edwards L et al (2021) Cortical hypometabolism reflects local atrophy and tau pathology in symptomatic Alzheimer\u0026rsquo;s disease. Brain 145(2):713\u0026ndash;728. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1093/brain/awab294\u003c/span\u003e\u003cspan address=\"10.1093/brain/awab294\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLandau SM, Thomas BA, Thurfjell L et al (2014) Amyloid PET imaging in Alzheimer\u0026rsquo;s disease: a comparison of three radiotracers. Eur J Nucl Med Mol Imaging 41(7):1398\u0026ndash;1407. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s00259-014-2753-3\u003c/span\u003e\u003cspan address=\"10.1007/s00259-014-2753-3\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHansson O, Seibyl J, Stomrud E et al (2018) CSF biomarkers of Alzheimer\u0026rsquo;s disease concord with amyloid-β PETand predict clinical progression: A study of fully automated immunoassays in BioFINDER and ADNI cohorts. Alzheimers Dement J Alzheimers Assoc 14(11):1470\u0026ndash;1481. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.jalz.2018.01.010\u003c/span\u003e\u003cspan address=\"10.1016/j.jalz.2018.01.010\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGreve DN, Svarer C, Fisher PM et al (2014) Cortical surface-based analysis reduces bias and variance in kinetic modeling of brain PET data. NeuroImage 92:225\u0026ndash;236. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.neuroimage.2013.12.021\u003c/span\u003e\u003cspan address=\"10.1016/j.neuroimage.2013.12.021\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGreve DN, Salat DH, Bowen SL et al (2016) Different Partial Volume Correction Methods Lead to Different Conclusions: an 18F-FDG PET Study of Aging. NeuroImage 132:334\u0026ndash;343. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.neuroimage.2016.02.042\u003c/span\u003e\u003cspan address=\"10.1016/j.neuroimage.2016.02.042\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLandau SM, Harvey D, Madison CM et al (2011) Associations between cognitive, functional, and FDG-PET measures of decline in AD and MCI. Neurobiol Aging 32(7):1207\u0026ndash;1218. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.neurobiolaging.2009.07.002\u003c/span\u003e\u003cspan address=\"10.1016/j.neurobiolaging.2009.07.002\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDesikan RS, S\u0026eacute;gonne F, Fischl B et al (2006) An automated labeling system for subdividing the human cerebral cortex on MRI scans into gyral based regions of interest. NeuroImage 31(3):968\u0026ndash;980. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.neuroimage.2006.01.021\u003c/span\u003e\u003cspan address=\"10.1016/j.neuroimage.2006.01.021\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFischl B, Salat DH, Busa E et al (2002) Whole brain segmentation: automated labeling of neuroanatomical structures in the human brain. Neuron 33(3):341\u0026ndash;355. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/s0896-6273(02)00569-x\u003c/span\u003e\u003cspan address=\"10.1016/s0896-6273(02)00569-x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFox PT, Mintun MA, Reiman EM, Raichle ME (1988) Enhanced Detection of Focal Brain Responses Using Intersubject Averaging and Change-Distribution Analysis of Subtracted PET Images. J Cereb Blood Flow Metab 8(5):642\u0026ndash;653. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/jcbfm.1988.111\u003c/span\u003e\u003cspan address=\"10.1038/jcbfm.1988.111\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCerami C, Della Rosa PA, Magnani G et al (2015) Brain metabolic maps in Mild Cognitive Impairment predict heterogeneity of progression to dementia. NeuroImage Clin 7:187\u0026ndash;194. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.nicl.2014.12.004\u003c/span\u003e\u003cspan address=\"10.1016/j.nicl.2014.12.004\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMuehlboeck JS, Westman E, Simmons A (2014) TheHiveDB image data management and analysis framework. Front Neuroinformatics 7:49. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3389/fninf.2013.00049\u003c/span\u003e\u003cspan address=\"10.3389/fninf.2013.00049\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVoevodskaya O, Simmons A, Nordenskj\u0026ouml;ld R et al (2014) The effects of intracranial volume adjustment approaches on multiple regional MRI volumes in healthy aging and Alzheimer\u0026rsquo;s disease. \u003cem\u003eFront Aging Neurosci\u003c/em\u003e. ;6. Accessed September 5, 2023. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.frontiersin.org/articles/\u003c/span\u003e\u003cspan address=\"https://www.frontiersin.org/articles/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3389/fnagi.2014.00264\u003c/span\u003e\u003cspan address=\"10.3389/fnagi.2014.00264\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFalahati F, Ferreira D, Soininen H et al (2016) The Effect of Age Correction on Multivariate Classification in Alzheimer\u0026rsquo;s Disease, with a Focus on the Characteristics of Incorrectly and Correctly Classified Subjects. Brain Topogr 29(2):296\u0026ndash;307. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s10548-015-0455-1\u003c/span\u003e\u003cspan address=\"10.1007/s10548-015-0455-1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCrane PK, Carle A, Gibbons LE et al (2012) Development and assessment of a composite score for memory in the Alzheimer\u0026rsquo;s Disease Neuroimaging Initiative (ADNI). Brain Imaging Behav 6(4):502\u0026ndash;516. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s11682-012-9186-z\u003c/span\u003e\u003cspan address=\"10.1007/s11682-012-9186-z\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGibbons LE, Carle AC, Mackin RS et al (2012) A composite score for executive functioning, validated in Alzheimer\u0026rsquo;s Disease Neuroimaging Initiative (ADNI) participants with baseline mild cognitive impairment. Brain Imaging Behav 6(4):517\u0026ndash;527. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s11682-012-9176-1\u003c/span\u003e\u003cspan address=\"10.1007/s11682-012-9176-1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChoi SE, Mukherjee S, Gibbons LE et al (2020) Development and validation of language and visuospatial composite scores in ADNI. Alzheimers Dement N Y N 6(1):e12072. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/trc2.12072\u003c/span\u003e\u003cspan address=\"10.1002/trc2.12072\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePoulakis K, Pereira JB, Mecocci P et al (2018) Heterogeneous patterns of brain atrophy in Alzheimer\u0026rsquo;s disease. Neurobiol Aging 65:98\u0026ndash;108. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.neurobiolaging.2018.01.009\u003c/span\u003e\u003cspan address=\"10.1016/j.neurobiolaging.2018.01.009\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eInguanzo A, Poulakis K, Mohanty R et al (2021) Dementia with Lewy bodies subtypes identified by cluster analysis on structural MRI. Alzheimers Dement 17(S4):e053573. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/alz.053573\u003c/span\u003e\u003cspan address=\"10.1002/alz.053573\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eProbst P, Wright MN, Boulesteix AL (2019) Hyperparameters and tuning strategies for random forest. WIREs Data Min Knowl Discov 9(3):e1301. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/widm.1301\u003c/span\u003e\u003cspan address=\"10.1002/widm.1301\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCharil A, Shcherbinin S, Southekal S et al (2019) Tau Subtypes of Alzheimer\u0026rsquo;s Disease Determined in vivo Using Flortaucipir PET Imaging. J Alzheimers Dis 71(3):1037\u0026ndash;1048. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3233/JAD-190264\u003c/span\u003e\u003cspan address=\"10.3233/JAD-190264\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJack CR, Petersen RC, Xu YC et al (1997) Medial temporal atrophy on MRI in normal aging and very mild Alzheimer\u0026rsquo;s disease. Neurology 49(3):786\u0026ndash;794. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1212/wnl.49.3.786\u003c/span\u003e\u003cspan address=\"10.1212/wnl.49.3.786\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNugent S, Croteau E, Potvin O et al (2020) Selection of the optimal intensity normalization region for FDG-PET studies of normal aging and Alzheimer\u0026rsquo;s disease. Sci Rep 10:9261. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41598-020-65957-3\u003c/span\u003e\u003cspan address=\"10.1038/s41598-020-65957-3\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVisualization of Brain Statistics With R Packages ggseg and ggseg3d - Athanasia M. Mowinckel, Didac Vidal-Pi\u0026ntilde;eiro, 2020. Accessed August 31 (2023) \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://journals.sagepub.com/doi/full/10.1177/2515245920928009\u003c/span\u003e\u003cspan address=\"https://journals.sagepub.com/doi/full/10.1177/2515245920928009\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMinoshima S (1994) Posterior cingulate cortex in Alzheimer\u0026rsquo;s disease. Lancet 344(8926):895. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/S0140-6736(94)92871-1\u003c/span\u003e\u003cspan address=\"10.1016/S0140-6736(94)92871-1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMinoshima S, Giordani B, Berent S, Frey KA, Foster NL, Kuhl DE (1997) Metabolic reduction in the posterior cingulate cortex in very early Alzheimer\u0026rsquo;s disease. Published online July. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/ana.410420114\u003c/span\u003e\u003cspan address=\"10.1002/ana.410420114\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHoffman JM, Welsh-Bohmer KA, Hanson M et al (2000) FDG PET imaging in patients with pathologically verified dementia. J Nucl Med Off Publ Soc Nucl Med 41(11):1920\u0026ndash;1928\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHerholz K, Salmon E, Perani D et al (2002) Discrimination between Alzheimer dementia and controls by automated analysis of multicenter FDG PET. NeuroImage 17(1):302\u0026ndash;316. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1006/nimg.2002.1208\u003c/span\u003e\u003cspan address=\"10.1006/nimg.2002.1208\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMosconi L, Tsui WH, Herholz K et al (2008) Multicenter Standardized 18F-FDG PET Diagnosis of Mild Cognitive Impairment, Alzheimer\u0026rsquo;s Disease, and Other Dementias. J Nucl Med Off Publ Soc Nucl Med 49(3):390\u0026ndash;398. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.2967/jnumed.107.045385\u003c/span\u003e\u003cspan address=\"10.2967/jnumed.107.045385\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWahlund LO, Julin P, Johansson SE, Scheltens P (2000) Visual rating and volumetry of the medial temporal lobe on magnetic resonance imaging in dementia: a comparative study. J Neurol Neurosurg Psychiatry 69(5):630\u0026ndash;635. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1136/jnnp.69.5.630\u003c/span\u003e\u003cspan address=\"10.1136/jnnp.69.5.630\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWestman E, Cavallin L, Muehlboeck JS et al (2011) Sensitivity and Specificity of Medial Temporal Lobe Visual Ratings and Multivariate Regional MRI Classification in Alzheimer\u0026rsquo;s Disease. PLoS ONE 6(7):e22506. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1371/journal.pone.0022506\u003c/span\u003e\u003cspan address=\"10.1371/journal.pone.0022506\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eScheltens P, Leys D, Barkhof F et al (1992) Atrophy of medial temporal lobes on MRI in probable Alzheimer\u0026rsquo;s disease and normal ageing: diagnostic value and neuropsychological correlates. J Neurol Neurosurg Psychiatry 55(10):967\u0026ndash;972. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1136/jnnp.55.10.967\u003c/span\u003e\u003cspan address=\"10.1136/jnnp.55.10.967\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJack CR, Petersen RC, Xu Y et al (1998) Rate of medial temporal lobe atrophy in typical aging and Alzheimer\u0026rsquo;s disease. Neurology 51(4):993\u0026ndash;999. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1212/WNL.51.4.993\u003c/span\u003e\u003cspan address=\"10.1212/WNL.51.4.993\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eByun MS, Kim SE, Park J et al (2015) Heterogeneity of Regional Brain Atrophy Patterns Associated with Distinct Progression Rates in Alzheimer\u0026rsquo;s Disease. PLoS ONE 10(11):e0142756. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1371/journal.pone.0142756\u003c/span\u003e\u003cspan address=\"10.1371/journal.pone.0142756\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBoon BDC, Labuzan S, Peng Z et al (2023) Retrospective Evaluation of Neuropathologic Proxies of the Minimal Atrophy Subtype Compared to Corticolimbic Alzheimer Disease Subtypes. Neurol Published online August 14. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1212/WNL.0000000000207685\u003c/span\u003e\u003cspan address=\"10.1212/WNL.0000000000207685\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShima K, Matsunari I, Samuraki M et al (2012) Posterior cingulate atrophy and metabolic decline in early stage Alzheimer\u0026rsquo;s disease. Neurobiol Aging 33(9):2006\u0026ndash;2017. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.neurobiolaging.2011.07.009\u003c/span\u003e\u003cspan address=\"10.1016/j.neurobiolaging.2011.07.009\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMeguro K, LeMestric C, Landeau B, Desgranges B, Eustache F, Baron JC (2001) Relations between hypometabolism in the posterior association neocortex and hippocampal atrophy in Alzheimer\u0026rsquo;s disease: a PET/MRI correlative study. J Neurol Neurosurg Psychiatry 71(3):315\u0026ndash;321. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1136/jnnp.71.3.315\u003c/span\u003e\u003cspan address=\"10.1136/jnnp.71.3.315\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDelbeuck X, Van der Linden M, Collette F (2003) Alzheimer\u0026rsquo; Disease as a Disconnection Syndrome? Neuropsychol Rev 13(2):79\u0026ndash;92. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1023/A:1023832305702\u003c/span\u003e\u003cspan address=\"10.1023/A:1023832305702\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMinoshima S, Cross D, Thientunyakit T, Foster NL, Drzezga A (2022) 18F-FDG PET Imaging in Neurodegenerative Dementing Disorders: Insights into Subtype Classification, Emerging Disease Categories, and Mixed Dementia with Copathologies. J Nucl Med 63(Supplement 1):2S\u0026ndash;12S. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.2967/jnumed.121.263194\u003c/span\u003e\u003cspan address=\"10.2967/jnumed.121.263194\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShiino A, Watanabe T, Maeda K, Kotani E, Akiguchi I, Matsuda M (2006) Four subgroups of Alzheimer\u0026rsquo;s disease based on patterns of atrophy using VBM and a unique pattern for early onset disease. NeuroImage 33(1):17\u0026ndash;26. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.neuroimage.2006.06.010\u003c/span\u003e\u003cspan address=\"10.1016/j.neuroimage.2006.06.010\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEwers M, Brendel M, Rizk-Jackson A et al (2013) Reduced FDG-PET brain metabolism and executive function predict clinical progression in elderly healthy subjects. NeuroImage Clin 4:45\u0026ndash;52. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.nicl.2013.10.018\u003c/span\u003e\u003cspan address=\"10.1016/j.nicl.2013.10.018\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMosconi L (2013) Glucose metabolism in normal aging and Alzheimer\u0026rsquo;s disease: methodological and physiological considerations for PET studies. Clin Transl Imaging 1(4):217\u0026ndash;233. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s40336-013-0026-y\u003c/span\u003e\u003cspan address=\"10.1007/s40336-013-0026-y\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBenvenutto A, Giusiano B, Koric L et al (2018) Imaging Biomarkers of Neurodegeneration in Alzheimer\u0026rsquo;s Disease: Distinct Contributions of Cortical MRI Atrophy and FDG-PET Hypometabolism. J Alzheimers Dis 65(4):1147\u0026ndash;1157. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3233/JAD-180292\u003c/span\u003e\u003cspan address=\"10.3233/JAD-180292\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePerini G, Rodriguez-Vieitez E, Kadir A, Sala A, Savitcheva I, Nordberg A (2021) Clinical impact of 18F-FDG-PET among memory clinic patients with uncertain diagnosis. Eur J Nucl Med Mol Imaging 48(2):612\u0026ndash;622. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s00259-020-04969-7\u003c/span\u003e\u003cspan address=\"10.1007/s00259-020-04969-7\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKapasi A, DeCarli C, Schneider JA (2017) Impact of multiple pathologies on the threshold for clinically overt dementia. Acta Neuropathol (Berl) 134(2):171\u0026ndash;186. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s00401-017-1717-7\u003c/span\u003e\u003cspan address=\"10.1007/s00401-017-1717-7\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMohanty R, Ferreira D, Frerich S et al (2022) Neuropathologic Features of Antemortem Atrophy-Based Subtypes of Alzheimer Disease. Neurology 99(4):e323\u0026ndash;e333. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1212/WNL.0000000000200573\u003c/span\u003e\u003cspan address=\"10.1212/WNL.0000000000200573\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRabinovici GD, Seeley WW, Kim EJ et al (2007) Distinct MRI atrophy patterns in autopsy-proven Alzheimer\u0026rsquo;s disease and frontotemporal lobar degeneration. Am J Alzheimers Dis Other Demen 22(6):474\u0026ndash;488. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1177/1533317507308779\u003c/span\u003e\u003cspan address=\"10.1177/1533317507308779\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSilva-Rodr\u0026iacute;guez J, Labrador-Espinosa MA, Moscoso A et al (2023) Characteristics of amnestic patients with hypometabolism patterns suggestive of Lewy body pathology. Brain J Neurol Published online June 7:awad194. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1093/brain/awad194\u003c/span\u003e\u003cspan address=\"10.1093/brain/awad194\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGrothe MJ, Moscoso A, Silva-Rodr\u0026iacute;guez J et al (2023) Differential diagnosis of amnestic dementia patients based on an FDG-PET signature of autopsy-confirmed LATE-NC. Alzheimers Dement 19(4):1234\u0026ndash;1244. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/alz.12763\u003c/span\u003e\u003cspan address=\"10.1002/alz.12763\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHerholz K, Nordberg A, Salmon E et al (1999) Impairment of Neocortical Metabolism Predicts Progression in Alzheimer\u0026rsquo;s Disease. Dement Geriatr Cogn Disord 10(6):494\u0026ndash;504. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1159/000017196\u003c/span\u003e\u003cspan address=\"10.1159/000017196\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePoulakis K, Ferreira D, Pereira JB, Smedby \u0026Ouml;, Vemuri P, Westman E (2020) Fully bayesian longitudinal unsupervised learning for the assessment and visualization of AD heterogeneity and progression. Aging 12(13):12622\u0026ndash;12647. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.18632/aging.103623\u003c/span\u003e\u003cspan address=\"10.18632/aging.103623\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVogel JW, Young AL, Oxtoby NP et al (2021) Four distinct trajectories of tau deposition identified in Alzheimer\u0026rsquo;s disease. Nat Med 27(5):871\u0026ndash;881. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41591-021-01309-6\u003c/span\u003e\u003cspan address=\"10.1038/s41591-021-01309-6\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCollij LE, Salvad\u0026oacute; G, Wottschel V et al (2022) Spatial-Temporal Patterns of β-Amyloid Accumulation: A Subtype and Stage Inference Model Analysis. Neurology 98(17):e1692\u0026ndash;e1703. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1212/WNL.0000000000200148\u003c/span\u003e\u003cspan address=\"10.1212/WNL.0000000000200148\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":false,"highlight":"","institution":"Karolinska Institute","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Alzheimer's disease, data-driven models, subtypes, hypometabolism, atrophy, PET, MRI","lastPublishedDoi":"10.21203/rs.3.rs-4454593/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-4454593/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e[\u003csup\u003e18\u003c/sup\u003eF] fluorodeoxyglucose (FDG)-PET and MRI are key imaging markers for neurodegeneration in Alzheimer's disease.\u0026nbsp; It is well-established that parieto-temporal hypometabolism on FDG-PET is closely associated with medial temporal atrophy on MRI in Alzheimer's disease. Substantial biological heterogeneity, expressed as distinct subtypes of hypometabolism or atrophy patterns, has been previously described in Alzheimer's disease using data-driven and hypothesis-driven methods. However, the link between these two imaging modalities has not yet been explored in the context of Alzheimer's disease subtypes. To investigate this link, the current study utilised FDG-PET and MRI scans from 180 amyloid-beta positive Alzheimer's disease dementia patients and 176 amyloid-beta negative cognitively normal controls from the Alzheimer’s Disease Neuroimaging Initiative (ADNI). Random forest hierarchical clustering, a data-driven model for identifying subtypes, was implemented in the two modalities: one with standard uptake value ratios and the other with grey matter volumes. Five subtypes hypometabolism- and atrophy-based subtypes were identified, exhibiting both cortical-predominant and limbic-predominant patterns although with differing percentages and clinical presentations. Three cortical-predominant hypometabolism subtypes found were: Cortical Predominant (32%), Cortical Predominant+ (11%), Cortical Predominant posterior (9%); and two limbic-predominant hypometabolism subtypes: Limbic Predominant (36%) and Limbic Predominant (13%). In addition, minimal and diffuse neurodegeneration subtypes were observed from the MRI data. The five atrophy subtypes were found: Cortical Predominant (19%), Limbic Predominant (27%), Diffuse (28%), Diffuse+ (6%) and Minimal (19%). \u0026nbsp;Inter-modality comparisons showed that all FDG-PET subtypes displayed medial temporal atrophy, whereas the distinct MRI subtypes showed topographically similar hypometabolism. Further, allocations of FDG-PET and MRI subtypes were not consistent when compared at an individual-level. Additional analysis comparing the data-driven clustering model with prior hypothesis-driven methods showed only partial agreement between these subtyping methods. FDG-PET subtypes had greater differences between limbic-predominant and cortical-predominant patterns and MRI subtypes had greater differences in severity of atrophy. In conclusion, this study highlighted that Alzheimer's disease subtypes identified using both FDG-PET and MRI capture distinct pathways showing cortical versus limbic predominance of neurodegeneration. However, the subtypes do not share a bidirectional relationship between modalities and are thus not interchangeable.\u003c/p\u003e","manuscriptTitle":"­­­Divergent Neurodegenerative Patterns: Comparison of FDG-PET- and MRI-based Alzheimer’s Disease Subtypes","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-05-31 20:30:55","doi":"10.21203/rs.3.rs-4454593/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"6c5ccb48-53b7-428d-884a-52248b357962","owner":[],"postedDate":"May 31st, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[{"id":32611441,"name":"Nuclear Medicine \u0026 Medical Imaging"},{"id":32611442,"name":"Neurobiology of Disease"}],"tags":[],"updatedAt":"2024-11-26T15:12:18+00:00","versionOfRecord":{"articleIdentity":"rs-4454593","link":"https://doi.org/10.1093/braincomms/fcae426","journal":{"identity":"brain-communications","isVorOnly":true,"title":"Brain Communications"},"publishedOn":"2024-11-23 00:00:00","publishedOnDateReadable":"November 23rd, 2024"},"versionCreatedAt":"2024-05-31 20:30:55","video":"","vorDoi":"10.1093/braincomms/fcae426","vorDoiUrl":"https://doi.org/10.1093/braincomms/fcae426","workflowStages":[]},"version":"v1","identity":"rs-4454593","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-4454593","identity":"rs-4454593","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}