Differential Diagnosis of Alzheimer's Disease and Mild Cognitive Impairment Using the Structural and Diffusion MRI Features of the Brain

Differential Diagnosis of Alzheimer's Disease and Mild Cognitive Impairment Using the Structural and Diffusion MRI Features of the Brain | 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 Differential Diagnosis of Alzheimer's Disease and Mild Cognitive Impairment Using the Structural and Diffusion MRI Features of the Brain Seyed Amir Zamanpour, Bahare Bigham, Mohamad-Hoseyn Sigari, Hoda Zare This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-573090/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Introduction: Accurate, fast, and reliable diagnosis of Alzheimer's Disease (AD) from Mild Cognitive Impairment (MCI) is crucial for prescribing proper treatment and prevention of disease progression. At first glance, structural and diffusion MRI images, are affected by neurodegenerative proceedings in AD and MCI. In this study, we are looking for the most effective features to detect and differentiate between healthy normal control (NC), AD, and MCI groups by non-invasive Magnetic Resonance Imaging (MRI) method and propose the automatic multi-class classification using the structural and diffusion MRI Features of the brain. Methods: The structural and diffusion MRI data were downloaded from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) database on three groups including AD, MCI, and NC subjects. Four famous classification models of machine learning were used to discover the best classification as a diagnostic tool for separation of the NC, AD and MCI groups. Results: Taken together, our results from this study lead to classify three groups for differentiation between the NC group and patients with MCI and AD, with average accuracy factor 89.9% for Support Vector Machine (SVM) and 91.9% for Artificial Neural Network (ANN) using selected features. Conclusions: Top 9 regions repetitive of WM based on four types of features are the caudate nucleus, corpus callosum, hippocampus, para hippocampus, temporal gyrus, putamen nucleus, cingulate gyrus, the region of 36 and 3 Brodmann. Therefore, these regions could be considered for identifying, monitoring, and future drug trials that could target this brain region to AD and MCI Management. Nuclear Medicine & Medical Imaging Alzheimer’s Disease Mild Cognitive Impairments Magnetic Resonance Imaging Machine Learning Diffusion Weighted imaging Structural MRI imaging Figures Figure 1 1. Introduction Alzheimer’s disease (AD) is the most common kind of dementia which is considered a progressive neurodegenerative disorder and it affects millions of people in the world. Disability to remember is one of the first signs of AD. Although, memory problems can be formed by Mild cognitive impairment (MCI). The AD consists of four stages: very early signs and symptoms, Mid AD, Moderate and, Severe AD. The first stage of AD is similar to MCI, also, with the available diagnostic tools people often are diagnosed in the mid AD stage. About eight of every ten elderly people with MCI go on to expand AD within 7 years; However, exactly unclear why some people with MCI do not advance to AD ( 1 ). Neuroimaging technologies, such as magnetic resonance imaging (MRI) or positron emission tomography (PET) are useful for AD diagnosis by detecting structural or functional markers at the early stage.( 1 ) MRI scans use magnetic fields and radio waves for the detection of the hydrogen atoms in tissues; However, PET uses radionuclides nuclear material for monitoring functional brain ( 2 ). Thus, the MRI scan is a more Non-invasive method than the PET scan. In the past, studies based on structural magnetic resonance imaging (sMRI), significant results have been achieved to distinguish between healthy normal control (NC) from AD groups with the help of machine learning. Nevertheless, conventional MRI is not able to highlight the construction of white matter (WM) regions ( 3 ). Extracted Parameters of segmented areas in SMRI include morphometry features such as volumetric and thickness measurements, or curvature of regions, and features based on pixel intensity of regions ( 4 ). Diffusion tensor imaging (DTI) is an MRI scan protocol that detects the random Brownian motion of water molecules in the body. This protocol identifies neural pathways by using data reconstruction models. In particular, DTI is an excellent tool for the identification of microstructural white matter degeneration in Alzheimer’s disease ( 5 , 6 ). DTI data reconstruction can be divided into two categories: the free-model and model-based methods. One of the models of reconstruction is based on diffusion tensor imaging (DTI) ( 7 , 8 ).Features derived from DTI inclusive: fractional anisotropy (FA), mean diffusivity (MD), axial diffusivity (AD) and, radial diffusivity (RD) and these mentioned features are obtained from diffusion matrix eigenvectors and eigenvalues. On the other, one of the free model reconstruction methods such as Generalized Q sampling Imaging (GQI) and Q-space diffeomorphic reconstruction (QSDR) is no assumption on the distribution. The QSDR method reconstructs data in the MNI space ( 9 ). Extractable features from QSDR methods include quantitative anisotropy (QA) related to tract compactness, the isotropic value (ISO) related to free water diffusion and, restricted diffusion imaging (RDI) ( 10 ). In the DTI reconstruction model, the fractional anisotropy (FA) parameter was measured for each voxel of the whole brain volume and the FA map was performed for each subject. FA generally is considered as an index of structural brain connection because it is sensitive to the axonal structure. Important factors affecting FA are myelin of axons. Subsequently, an analytical relation is provided in the reconstruction based on the GQI model. The calculation is not a numerical estimation but it is a direct calculation with using simple matrix multiplication. So the GQI is more accurate in terms of its mathematical form due to numerical estimation. Also, GQI can achieve with a linear and nonlinear transformation to reconstruction data in the T1W or MNI space. QSDR provides connectometry analysis and template construction. QA is more reliable than FA simply because it is less affected by the inflammation and edema that is due to the existence of the independent metric called ISO, which is calculated and extracted in this model. The QA factor is more sensitive to Pathological changes in the neurons. DSI Studio uses a voxel with water-free diffusion to calibrating the QA factor for ensuring better repeatability and compatibility ( 10 – 12 ). Recently, several studies have demonstrated that the features obtained from diffusion and structural measures contribute useful information for more accurate classification ( 13 – 16 ). In this study, the parameters extracted from brain images were used as features for classification. After extracting the features, the feature selection step is performed. Feature selection is a way of pre-processing features before entering the machine learning phase. It is a process of picking out a proper subset of all features so that the feature space is optimally decreasing matching to a specified evaluation criterion ( 17 ). The importance of feature selection is in eliminating idle and unused features, enhancing performance in machine learning models, improving learning efficiency such as predictive accuracy and raising the comprehensibility of learned conclusion ( 18 , 19 ). With the advancement of technology in recent years which consists of exponential growth in computing power, big data processing technologies, accessibility to large clinical data sets using electronic health records, and machine learning in medicine can lead to more precise diagnostic algorithms and individualize patient treatment ( 20 ). Machine learning is mathematical focuses on the predictive performance and generalization of models around cross-validation and iterative improvement of the algorithm ( 18 , 21 ). In this study, we used four classified models for diagnosis which included: Decision Tree (DT), k-Nearest Neighbors (KNN), Support Vector Machines (SVM) and, Artificial Neural Networks (ANN) and subsequently performed classification into three groups AD/MCI/NC simultaneously. Furthermore, we investigated which type of extracted features that were measured with MRI by using machine learning models can use as a tool for the diagnosis of normal aging from MCI and AD. 2. Materials And Methods 2.1. Data acquisition and preprocessing Participants in this study were selected from the Alzheimer’s disease neuroimaging initiative (ADNI) database ( http://www.loni.ucla.edu/ADNI/ ). The images that were used for analyzing included: diffusion-weighted, T1 weighted, T2 weighted and, FLAIR scans of 72 participants. Subjects including 24 NC (Male (M)/Female (F) = 11/13, mean age = 75.32 ± 8.31 year, Mini-Mental State Examination )MMSE( = 29.04 ± 1.23), 24 AD patients (M/F = 16/8, mean age = 76.47 ± 8.23 year, MMSE = 20.17 ± 4.97) and, 24 MCI (M/F = 12/12, mean age = 76.04 ± 8.65 year, MMSE = 26.70 ± 2.07). Two MCI levels (early or late) are generally defined according to the Wechsler Memory Scale Logical Memory II(22). In this paper, we considered early and late groups that each of them contains 12 subjects to create MCI. NC subjects do not have a history of neurological or psychiatric disorders; subjects with AD who have signatures of the NINCDS-ADRDA (National Institute of Neurological and Communicative Diseases and Stroke/Alzheimer's Disease and Related Disorders Association) criteria which means AD's disorder. Also, MCI subjects have symptoms a subjective memory concern; However, they were without any significant disability in other cognitive domains (they follow everyday activities with any signs of dementia). All scans that used in this study, were acquired by a 3T MRI Scanner (GE Medical Systems scanner). We collected structural MRI from all participants: T1- SPGE: TR=6.9 ms, TE=2.8 ms, slice thickness=1.2 mm; T2: TR=650 ms, TE=20 ms, slice thickness=4 mm; and FLAIR: TR=11000ms, TE=147.9 ms, slice thickness=5 mm; also, we collected diffusion MRI from all participants: matrix size = 256 × 256 × 46; voxel size = 1.36 × 1.36 × 2.7 mm; the sequence contained 5 images acquired without diffusion weighting and with diffusion weighting along 41 non-collinear directions (b = 1000 s/mm²). 2.2. Structural MRI processing We performed the processing of morphometrically and intensity level assessment by using the FreeSurfer image analysis pipeline (23) (version 6 on ubuntu 16.04) based on T1 and T2-FLAIR weighted images. The results of this process were divided into two parts that consisted of subcortical and cortical. In the first place, we performed the subcortical processing steps which included: volume, intensity pixel value, range of intensity, maximum, and minimum intensity pixel value in each region of atlas available in Freesurfer (Desikan-Killiany atlas). In the second place, the measurements related to cortical were done which included: thickness, surface area, gray matter volume and the curve in each region of Desikan-Killiany atlas from the right and left brain hemispheres. Part of the FreeSurfer pipeline conforms to the MRI scans to 1 × 1 × 1 mm 3 resolution and corrects for bias field using the non-parametric non-uniform intensity normalization (N3) algorithm. Briefly, this processing includes removal of non-brain tissue, automated Talairach transformation,surface deformation, tessellation of the grey matter-white matter boundary, topology correction, and segmentation (24, 25). Finally, after doing all the processing in this section, 496 features were extracted from every subject and saved in CSV format. 2.3 Diffusion MRI processing The diffusion MRI data were preprocessed using DSI-Studio software (developed by Fang-Cheng Yeh from the Advanced Biomedical MRI Lab, National Taiwan University Hospital, Taiwan, Supported by Fiber Tractography Lab, University of Pittsburgh, and made available at http://dsistudio . labsolver.org/Download/). For skull stripping and filtering the background region, we used the masks provided by DSI-Studio. Before DTI parameter measurement, head motion and eddy current effects were corrected using the DSI-Studio toolbox. We used two different reconstruction methods include model base (DTI) and free model (QSDR); with two different attitudes to process the diffusion images. Four general feature categories were extracted from the images; the first type of features was diffusion-based parameters in each region of the structural freesurferseg and Brodmann atlases, the second type of features was tract-based parameters in each region of the same atlases and the third type of features was structural connectome with two atlases that were mentioned above. Finally, the last type of feature was the graph network. This process was performed for DTI model according to the following recommended parameters: FA index was used to determine the fiber tracking threshold and Otsu’s method was used to set the anisotropy (FA) threshold to stop the fiber tracking, the number of seed points = 1,000,000, max angle = 60◦, step size = half of the spatial resolution, length constraint = 30–300 mm, and no spatial smoothing. Also, DSI studio software was implemented to calculate the connections and graphs of the "connectivity matrix and graph Network " tool with two "End" and "Pass" views. In total, 92763 features were extracted from 72 participants. The features extracted of structural and diffusion processed MRI images in our work are given in the chart below. (Shown in Figure 1) 2.4. Feature selection and Classification Feature selection means recognizing the most relevant features for pattern recognition and noise reduction that usually used in studies where the number of data is extremely high while few studied subjects are available The Fast Correlation-Based Filter (FCBF) method was better than the other models among all the feature selection models because less than quadratic time complexity (17). FCBF selected 66 features for the next step. After the feature selection step, we performed four classification models including the DT, KNN, SVM and ANN with 5, 10, 15, 20, 30, 40, 50-fold cross-validation were used to evaluate the performance of the classifier. To evaluating the ANN model, training, validation and test phase was repeated for 30 times and finally the mean and standard deviation were reported. For this purpose, data set was divided into three parts: 60% for training, 20% for validation, and 20% for test. The MATLAB software R2012 (The Math Works, USA) was used to design and implement classification models. 3. Result The number of extracted features using structural and diffusion processed MRI images for each subject were 92763. Features extracted from the DTI model included region-based and tract-based features of structural freesurferseg and Brodmann atlases. Features extracted from QSDR model include Region-based, Tract- based, connectivity and Network features from structural freesurferseg and Brodmann atlases; Eventually, Features extracted from the structural MRI images from structural freesurferseg atlas include Morphometric and Voxel intensity features (Figure 1). Table 1 shows the number of selected features with the FCBF model. According to this table, related features, Connectivity and Network have the most weight of selected features. However, it should be noted that the features selected for the three groups AD/MCI/NC classification. So, removing any of them has a significant effect on the classification performance. In total, 66 features were selected. In table 2, the performance of each classification model (DT, KNN, and SVM) is provided to classify three groups of AD / MCI / NC. These classification models were evaluated by cross validation fold of 5, 10, 15, 20, 30, 40, and 50. We have considered the accuracy factor to compare the performance of all classification models. In table 3 the performance of ANN is presented to classify three groups of AD / MCI / NC. This process is repeated 30 times and the mean and standard deviation are calculated. We run the SVM, DT, KNN and, ANN models for the diagnosis of AD/MCI/NC using MATLAB tools. These results showed good performance using selected features such as morphometry, network, connective, and tract base for the classifying AD/MCI/NC. Two highest accuracy for the classification of three groups (NC, MCI, and AD individuals) were obtained 89.9% and 90.7% for the SVM and ANN model respectively by 66 selected features. 4. Discussion In this study, we evaluated the structural and diffusion MRI brain images of the Alzheimer's patients from the ADNI dataset with machine learning techniques to classify the three groups AD, MCI and NC. For high-throughput imaging analysis, we used a well-established automated pipeline for morphometry (Freesurfer) and a pipeline to estimate rigorously individualized white matter structure (DSI_Studio). the study, which has used only structural image T1W ADNI databases, it was performed by Lauge Sorensen et al. In 2016, to simultaneously classify the three groups AD, MCI and NC by using the extracted features of the Freesurfer and Hippocampal texture area. The accuracy obtained for the three-group classification is 62.7%. The extracted areas are classified as Hippocampal, Parietal lobs, Cingulate cortex and ventricles ( 24 ). By comparing between our study and previous study: The study that only considers the diffusion image ADNI database for the classification of AD from NC is Luis R. Peraza et al., 2019, with an accuracy of 89.07% using 401 Connectivity feature extraction. For automated classification, only 89 features were selected between the two groups of AD and NC. In this study the regions involved in the temporal lobe, subcortical-posterior, occipital brain were introduced ( 26 ). Also, the study that used both structural and diffusion imaging to diagnose AD, MCI, and NC, it was performed by Yun Wang et al in 2019. The Structural images that were taken from the ADNI database, were processed by Freesurfer software and the diffusion images were processed by MRtrix3 software. The current results are also in broad agreement with a recent SVM study that used connectivity measures and morphometry. Accuracy for AD-NC, NC-MCI, and AD-MCI was 96%, 70%, and 75% respectively. For this classification, the selected areas are not mentioned ( 25 ). The current study results show that it is possible to classify the three groups NC, MCI and AD subjects with a high degree of accuracy using an automated procedure that combines DTI, QSDR, and morphometry sMRI. Our results from NC and MCI and AD classification which achieved an average accuracy of 89.9% and 90.7% for the SVM and ANN model respectively. Due to the high standard deviation in the ANN model, the SVM model is reported as a result of this study. The results of the KNN and DT classification performance were not considered due to the low accuracy. Many brain imaging studies have shown that some brain regions may play a key role in the diagnosis of MCI and AD that the present study can confirm these results.( 27 – 30 ) The extracted significant features could be used to find the most vulnerable regions of the brain to distinguish AD, MCI, and healthy elderly peoples. In the current study, the features of brain regions such as the caudate nucleus, corpus callosum, hippocampus, Para hippocampus, temporal gyrus, putamen nucleus, cingulate gyrus, the region of 36 and 3 Brodmann are the most important features for the separation of these three groups, especially the temporal lobe. Moreover, the Amygdala is the only brain region of which both left and right hemispheres are exposed to changes in imaging values between the three groups. According to similar studies, the areas are considered vulnerable and this data can provide a clinician with additional information. Top 9 regions repetitive of WM based on four types of features are the caudate nucleus, corpus callosum, hippocampus, Para hippocampus, temporal gyrus, putamen nucleus, cingulate gyrus, and the regions of 36 and 3 Brodmann. These areas have the highest frequency in various measurements such as connectivity and network of this region with another area, the volume of tracts on these regions, and the thickness of regions. Generally, the features extracted from the Atlas of Structure were much more efficient than the Broadman Atlas. 5. Conclusion In this study we evaluated different three-group classification approaches using combine the neuroimaging modalities MRI and DTI. Combining anatomical and diffusion features, provided information indicating changes in the brain of patients with AD and MCI. Considering the results, three-group classification compared to two-group classifications used in above studies have significantly more important. The highest frequency in this study could be considered for identifying, monitoring, and future drug trials that could target these brain regions to AD and MCI Management. Eventually, the MRI can be combined with machine learning algorithms to detect damage in the early stages of Alzheimer’s disease. Abbreviations AD : Alzheimer’s disease MCI : Mild cognitive impairment DTI : Diffusion tensor imaging DT : Decision Tree KNN : k-nearest neighbors SVM : support vector machines ANN : Artificial Neural Networks Declarations Ethics approval and consent to participate: The study was reviewed and approved by the Ethical Committee of Mashhad University of Medical Sciences. (Ethical number: IR.MUMS.MEDICAL.REC.1397.410) Consent for publication : Not applicable. Availability of data and material : The data used to support the findings of this study are available from the corresponding author upon request. Competing interests : The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Funding : This study was financially supported by a grant from the Vice Chancellor of Research of Mashhad University of Medical Sciences; grant number 970508, But was not involved in the decision to publish. Authors' contributions : Dr. Hoda Zare Seyed Amir Zamanpour and Bahare Bigham conceived of the presented idea and developed the theory and performed the computations. Dr. Hoda Zare and verified the analytical methods and supervised the findings of this work. Dr. Hoda Zare , Seyed Amir Zamanpour and Mohamad-Hoseyn Sigari discussed the results and commented on the manuscript. All authors contributed to and have approved the final manuscript. Each named author has substantially contributed to conducting the underlying research and drafting this manuscript. Additionally, to the best of our knowledge, the named authors have no conflict of interest, financial or otherwise. Acknowledgements: Data used in the preparation of this article were obtained from the Alzheimer's Disease Neuroimaging Initiative (ADNI) database ( adni.loni.usc.edu ). The ADNI was launched in 2003 as a public-private partnership, led by Principal Investigator Michael W. Weiner, MD. The primary goal of ADNI has been to test whether serial magnetic resonance imaging (MRI), positron emission tomography (PET), other biological markers, and clinical and neuropsychological assessment can be combined to measure the progression of mild cognitive impairment (MCI) and early Alzheimer's disease (AD). For up-to-date information, see www.adni-info.org . References Sutton AL. Alzheimer Disease Sourcebook: Basic Consumer Health Information about Alzheimer Disease and Other Forms of Dementia, Including Mild Cognitive Impairment, Corticobasal Degeneration, Dementia with Lewy Bodies, Frontotemporal Dementia, Huntington Disease, Parkinson Disease, and Vascular Dementia Along with Information about Recent Research on the Diagnosis and Prevention of Alzheimer Disease and Genetic Testing, Tips for Maintaining Cognitive Functioning, Strategies for Long-term Planning, Advice for Caregivers, a Glossary of Related Terms, and Directories of Resources for Additional Help and Information: Omnigraphics; 2011. Meyer P, Rijntjes M, Hellwig S, Klöppel S, Weiller C. Functional neuroimaging: functional magnetic resonance imaging, positron emission tomography, and single-photon emission computed tomography2016. Maggipinto T, Bellotti R, Amoroso N, Diacono D, Donvito G, Lella E, et al. DTI measurements for Alzheimer’s classification. Physics in Medicine & Biology. 2017;62(6):2361. Li M, Qin Y, Gao F, Zhu W, He X. Discriminative analysis of multivariate features from structural MRI and diffusion tensor images. Magnetic resonance imaging. 2014;32(8):1043-51. Wen Q, Mustafi SM, Li J, Risacher SL, Tallman E, Brown SA, et al. White matter alterations in early-stage Alzheimer's disease: A tract-specific study. Alzheimer's & Dementia: Diagnosis, Assessment & Disease Monitoring. 2019;11:576-87. Bigham B, Zamanpour SA, Zemorshidi F, Boroumand F, Zare H, Initiative AsDN. Identification of Superficial White Matter Abnormalities in Alzheimer’s Disease and Mild Cognitive Impairment Using Diffusion Tensor Imaging. Journal of Alzheimer's Disease Reports. (Preprint):1-11. Mori S. Introduction to diffusion tensor imaging: Elsevier; 2007. Alexander AL, Lee JE, Lazar M, Field AS. Diffusion tensor imaging of the brain. Neurotherapeutics : the journal of the American Society for Experimental NeuroTherapeutics. 2007;4(3):316-29. Yeh F-C, Wedeen VJ, Tseng W-YI. Generalized ${q} $-sampling imaging. IEEE transactions on medical imaging. 2010;29(9):1626-35. Zhang H, Wang Y, Lu T, Qiu B, Tang Y, Ou S, et al. Differences between generalized q-sampling imaging and diffusion tensor imaging in the preoperative visualization of the nerve fiber tracts within peritumoral edema in brain. Neurosurgery. 2013;73(6):1044-53. Yeh F-C, Tseng W-YI. NTU-90: a high angular resolution brain atlas constructed by q-space diffeomorphic reconstruction. NeuroImage. 2011;58(1):91-9. Yeh F-C. Diffusion MRI Reconstruction in DSI Studio n.d. [Available from: http://dsi-studio.labsolver.org/Manual/Reconstruction . Schouten TM, Koini M, de Vos F, Seiler S, van der Grond J, Lechner A, et al. Combining anatomical, diffusion, and resting state functional magnetic resonance imaging for individual classification of mild and moderate Alzheimer's disease. NeuroImage Clinical. 2016;11:46-51. de Vos F, Schouten TM, Hafkemeijer A, Dopper EG, van Swieten JC, de Rooij M, et al. Combining multiple anatomical MRI measures improves Alzheimer's disease classification. Human brain mapping. 2016;37(5):1920-9. Dai Z, Yan C, Wang Z, Wang J, Xia M, Li K, et al. Discriminative analysis of early Alzheimer's disease using multi-modal imaging and multi-level characterization with multi-classifier (M3). NeuroImage. 2012;59(3):2187-95. Schouten TM, Koini M, Vos F, Seiler S, Rooij M, Lechner A, et al. Individual classification of Alzheimer's disease with diffusion magnetic resonance imaging. NeuroImage. 2017;152:476-81. Yu L, Liu H, editors. Feature selection for high-dimensional data: A fast correlation-based filter solution. Proceedings of the 20th international conference on machine learning (ICML-03); 2003. Shalev-Shwartz S, Ben-David S. Understanding machine learning: From theory to algorithms: Cambridge university press; 2014. Guyon I, Elisseeff A. An introduction to variable and feature selection. Journal of machine learning research. 2003;3(Mar):1157-82. Handelman G, Kok H, Chandra R, Razavi A, Lee M, Asadi H. eD octor: machine learning and the future of medicine. Journal of internal medicine. 2018;284(6):603-19. Cleophas TJ, Zwinderman AH. Machine Learning in Medicine-a Complete Overview: Springer; 2015. Wechsler D. Wechsler memory scale-revised. Psychological Corporation. 1987. Fischl B. FreeSurfer. NeuroImage. 2012;62(2):774-81. Sørensen L, Igel C, Pai A, Balas I, Anker C, Lillholm M, et al. Differential diagnosis of mild cognitive impairment and Alzheimer's disease using structural MRI cortical thickness, hippocampal shape, hippocampal texture, and volumetry. NeuroImage: Clinical. 2017;13:470-82. Wang Y, Xu C, Park J-H, Lee S, Stern Y, Yoo S, et al. Diagnosis and prognosis of Alzheimer's disease using brain morphometry and white matter connectomes. NeuroImage: Clinical. 2019;23:101859. Peraza LR, Díaz-Parra A, Kennion O, Moratal D, Taylor J-P, Kaiser M, et al. Structural connectivity centrality changes mark the path toward Alzheimer's disease. Alzheimer's & Dementia: Diagnosis, Assessment & Disease Monitoring. 2019;11:98-107. Fellgiebel A, Yakushev I. Diffusion tensor imaging of the hippocampus in MCI and early Alzheimer's disease. Journal of Alzheimer's disease : JAD. 2011;26 Suppl 3:257-62. Huang J, Friedland R, Auchus A. Diffusion tensor imaging of normal-appearing white matter in mild cognitive impairment and early Alzheimer disease: preliminary evidence of axonal degeneration in the temporal lobe. American Journal of Neuroradiology. 2007;28(10):1943-8. Petersen RC, Smith GE, Waring SC, Ivnik RJ, Tangalos EG, Kokmen E. Mild cognitive impairment: clinical characterization and outcome. Archives of neurology. 1999;56(3):303-8. Zhang Y, Schuff N, Jahng GH, Bayne W, Mori S, Schad L, et al. Diffusion tensor imaging of cingulum fibers in mild cognitive impairment and Alzheimer disease. Neurology. 2007;68(1):13-9. Tables Table 1: The number of selected features with the FCBF feature selection model Type of features Number of selected features Percentage of total selected features DTI Model Region Base 4 6% Tract Base 9 13.6% QSDR Model Region base 3 4.6% Tract base 1 1.6% Connectivity 30 45.5% Network 11 16.7% Structural MRI Cortical Region 5 7.5% Sub-Cortical Region 3 4.5% Table 2: the performance of DT, KNN, and SVM with cross validation fold in term of the accuracy factor. cross validation fold 5 10 15 20 30 40 50 average Sd DT CN 54.2 54.2 50 58.3 50 58.4 45.8 52.9 4.6 MCI 62.5 62.5 50 41.7 62.5 66.6 62.5 58.3 8.9 AD 70.8 70.8 75 70.8 66.6 70.8 75 71.4 2.8 all 62.5 62.5 58.3 56.9 59.7 65.3 61.1 60.9 2.8 KNN CN 83.3 95.8 95.8 95.8 95.8 95.8 95.8 94.0 4.7 MCI 79.2 45.8 54.2 51.2 54.2 66.6 58.3 58.5 11.1 AD 83.3 91.7 95.8 91.7 91.6 91.7 91.7 91.0 3.7 all 81.9 77.8 81.9 80.6 80.6 84.7 81.9 81.3 2.0 SVM CN 95.8 95.8 91.6 95.8 95.8 95.8 91.7 94.6 2.0 MCI 83.3 87.5 87.5 87.5 83.3 83.3 87.5 85.7 2.2 AD 87.5 91.6 91.6 87.5 91.7 87.5 87.5 89.2 2.2 all 88.9 91.7 90.3 90.3 90.3 88.9 88.9 89.9 1.0 Table 3: the performance of ANN in term of the accuracy factor and the mean and standard deviation ANN index CN MCI AD All 1 100 83.3 91.7 91.7 17 83.3 87.5 100 90.3 2 95.8 83.3 91.7 90.3 18 95.8 87.5 91.7 91.7 3 100 83.3 91.7 91.7 19 83.3 87.5 91.7 87.5 4 91.7 83.3 100 91.7 20 91.7 79.2 91.7 87.5 5 87.5 83.3 91.7 87.5 21 100 87.5 95.8 94.4 6 95.8 87.5 95.8 93.1 22 100 87.5 95.8 94.4 7 91.7 83.3 83.3 86.1 23 91.7 95.8 91.7 93.1 8 95.8 95.8 95.8 95.8 24 91.7 91.7 95.8 93.1 9 91.7 83.3 95.8 90.3 25 95.8 91.7 100 95.8 10 100 70.8 100 90.3 26 91.7 87.5 95.8 91.7 11 95.8 91.7 95.8 94.4 27 95.8 54.2 91.7 80.6 12 91.7 83.3 91.7 88.9 28 100 66.7 100 88.9 13 91.7 87.5 95.8 91.7 29 83.3 87.5 75 81.9 14 95.8 83.3 100 93.1 30 100 70.8 100 90.3 15 100 79.2 95.8 91.7 mean 94.3 83.7 94.0 90.7 16 100 87.5 87.5 91.7 Sd 5.1 8.6 5.4 3.5 Additional Declarations No competing interests reported. Cite Share Download PDF Status: Posted 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-573090","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":31813219,"identity":"8011e001-0a27-48a2-9671-6886ca89357a","order_by":0,"name":"Seyed Amir Zamanpour","email":"","orcid":"","institution":"Mashhad University of Medical Sciences","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Seyed","middleName":"Amir","lastName":"Zamanpour","suffix":""},{"id":31813220,"identity":"c2dddf04-0383-441c-a6d3-b5994e24a8cd","order_by":1,"name":"Bahare Bigham","email":"","orcid":"","institution":"Mashhad University of Medical Sciences","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Bahare","middleName":"","lastName":"Bigham","suffix":""},{"id":31813221,"identity":"4ebd9389-3d57-4ce3-a3d8-17f2954a7112","order_by":2,"name":"Mohamad-Hoseyn Sigari","email":"","orcid":"","institution":"Quchan University of Advanced Technology","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Mohamad-Hoseyn","middleName":"","lastName":"Sigari","suffix":""},{"id":31813222,"identity":"f3c04982-c221-4fc4-820c-88b0925491fb","order_by":3,"name":"Hoda Zare","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAsklEQVRIiWNgGAWjYJACZoYKOFuCWC1nSNbC2EaKo+Tbzz58XDhvW+La/gOMH34wWOQT1MLYk25sPHPb7cRtNxKYJXsYJCwbCDqKIY1NmheshYFBGugXA4K2sPE/Y//NOweo5fwB5t9EaeGRSGNj5m0AajmQwEacLRISz5ileY7dNt52I7HNsseACC3y/WmMn3lqbstuO3/48I0fFXWEtSABxgYGBpI0jIJRMApGwSjACQB+ATS825CmmQAAAABJRU5ErkJggg==","orcid":"","institution":"Mashhad University of Medical Sciences","correspondingAuthor":true,"submittingAuthor":false,"prefix":"","firstName":"Hoda","middleName":"","lastName":"Zare","suffix":""}],"badges":[],"createdAt":"2021-05-29 18:29:05","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-573090/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-573090/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":10179477,"identity":"b0736e74-2fe6-40ff-bea4-ce39c380e77a","added_by":"auto","created_at":"2021-06-09 21:50:05","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":350354,"visible":true,"origin":"","legend":"the steps of processing and extracted features","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-573090/v1/7ee6793b4d46e31debafaa55.png"},{"id":13697978,"identity":"b233fd25-8b1e-4026-8c8a-e6bc6d59485f","added_by":"auto","created_at":"2021-09-17 13:12:18","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":505882,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-573090/v1/af6e9e67-b4ff-410f-86d5-61d9b4e2cd5b.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"\u003cp\u003eDifferential Diagnosis of Alzheimer's Disease and Mild Cognitive Impairment Using the Structural and Diffusion MRI Features of the Brain\u003c/p\u003e","fulltext":[{"header":"1. Introduction","content":" \u003cp\u003eAlzheimer\u0026rsquo;s disease (AD) is the most common kind of dementia which is considered a progressive neurodegenerative disorder and it affects millions of people in the world. Disability to remember is one of the first signs of AD. Although, memory problems can be formed by Mild cognitive impairment (MCI). The AD consists of four stages: very early signs and symptoms, Mid AD, Moderate and, Severe AD. The first stage of AD is similar to MCI, also, with the available diagnostic tools people often are diagnosed in the mid AD stage. About eight of every ten elderly people with MCI go on to expand AD within 7 years; However, exactly unclear why some people with MCI do not advance to AD (\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eNeuroimaging technologies, such as magnetic resonance imaging (MRI) or positron emission tomography (PET) are useful for AD diagnosis by detecting structural or functional markers at the early stage.(\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e) MRI scans use magnetic fields and radio waves for the detection of the hydrogen atoms in tissues; However, PET uses radionuclides nuclear material for monitoring functional brain (\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThus, the MRI scan is a more Non-invasive method than the PET scan. In the past, studies based on structural magnetic resonance imaging (sMRI), significant results have been achieved to distinguish between healthy normal control (NC) from AD groups with the help of machine learning. Nevertheless, conventional MRI is not able to highlight the construction of white matter (WM) regions (\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e). Extracted Parameters of segmented areas in SMRI include morphometry features such as volumetric and thickness measurements, or curvature of regions, and features based on pixel intensity of regions (\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eDiffusion tensor imaging (DTI) is an MRI scan protocol that detects the random Brownian motion of water molecules in the body. This protocol identifies neural pathways by using data reconstruction models. In particular, DTI is an excellent tool for the identification of microstructural white matter degeneration in Alzheimer\u0026rsquo;s disease (\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eDTI data reconstruction can be divided into two categories: the free-model and model-based methods.\u003c/p\u003e \u003cp\u003eOne of the models of reconstruction is based on diffusion tensor imaging (DTI) (\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e, \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e).Features derived from DTI inclusive: fractional anisotropy (FA), mean diffusivity (MD), axial diffusivity (AD) and, radial diffusivity (RD) and these mentioned features are obtained from diffusion matrix eigenvectors and eigenvalues. On the other, one of the free model reconstruction methods such as Generalized Q sampling Imaging (GQI) and Q-space diffeomorphic reconstruction (QSDR) is no assumption on the distribution. The QSDR method reconstructs data in the MNI space (\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e). Extractable features from QSDR methods include quantitative anisotropy (QA) related to tract compactness, the isotropic value (ISO) related to free water diffusion and, restricted diffusion imaging (RDI) (\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eIn the DTI reconstruction model, the fractional anisotropy (FA) parameter was measured for each voxel of the whole brain volume and the FA map was performed for each subject. FA generally is considered as an index of structural brain connection because it is sensitive to the axonal structure. Important factors affecting FA are myelin of axons. Subsequently, an analytical relation is provided in the reconstruction based on the GQI model. The calculation is not a numerical estimation but it is a direct calculation with using simple matrix multiplication. So the GQI is more accurate in terms of its mathematical form due to numerical estimation. Also, GQI can achieve with a linear and nonlinear transformation to reconstruction data in the T1W or MNI space. QSDR provides connectometry analysis and template construction. QA is more reliable than FA simply because it is less affected by the inflammation and edema that is due to the existence of the independent metric called ISO, which is calculated and extracted in this model. The QA factor is more sensitive to Pathological changes in the neurons. DSI Studio uses a voxel with water-free diffusion to calibrating the QA factor for ensuring better repeatability and compatibility (\u003cspan additionalcitationids=\"CR11\" citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eRecently, several studies have demonstrated that the features obtained from diffusion and structural measures contribute useful information for more accurate classification (\u003cspan additionalcitationids=\"CR14 CR15\" citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eIn this study, the parameters extracted from brain images were used as features for classification.\u003c/p\u003e \u003cp\u003eAfter extracting the features, the feature selection step is performed. Feature selection is a way of pre-processing features before entering the machine learning phase. It is a process of picking out a proper subset of all features so that the feature space is optimally decreasing matching to a specified evaluation criterion (\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e). The importance of feature selection is in eliminating idle and unused features, enhancing performance in machine learning models, improving learning efficiency such as predictive accuracy and raising the comprehensibility of learned conclusion (\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e, \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eWith the advancement of technology in recent years which consists of exponential growth in computing power, big data processing technologies, accessibility to large clinical data sets using electronic health records, and machine learning in medicine can lead to more precise diagnostic algorithms and individualize patient treatment (\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e). Machine learning is mathematical focuses on the predictive performance and generalization of models around cross-validation and iterative improvement of the algorithm (\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e, \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e). In this study, we used four classified models for diagnosis which included: Decision Tree (DT), k-Nearest Neighbors (KNN), Support Vector Machines (SVM) and, Artificial Neural Networks (ANN) and subsequently performed classification into three groups AD/MCI/NC simultaneously.\u003c/p\u003e \u003cp\u003eFurthermore, we investigated which type of extracted features that were measured with MRI by using machine learning models can use as a tool for the diagnosis of normal aging from MCI and AD.\u003c/p\u003e "},{"header":"2. Materials And Methods","content":"\u003cp\u003e\u003cstrong\u003e2.1. Data acquisition and preprocessing\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eParticipants in this study were selected from the Alzheimer\u0026rsquo;s disease neuroimaging initiative (ADNI) database (\u003ca href=\"http://www.loni.ucla.edu/ADNI/\"\u003ehttp://www.loni.ucla.edu/ADNI/\u003c/a\u003e). The images that were used for analyzing included: diffusion-weighted, T1 weighted, T2 weighted and, FLAIR scans of 72 participants. Subjects including 24 NC (Male (M)/Female (F) = 11/13, mean age = 75.32 \u0026plusmn; 8.31 year, Mini-Mental State\u0026nbsp;Examination )MMSE( = 29.04 \u0026plusmn; 1.23), 24 AD patients (M/F = 16/8, mean age = 76.47 \u0026plusmn; 8.23 year, MMSE = 20.17 \u0026plusmn; 4.97) and, 24 MCI (M/F = 12/12, mean age = 76.04 \u0026plusmn; 8.65 year, MMSE = 26.70 \u0026plusmn; 2.07). Two MCI levels (early or late) are generally defined according to the Wechsler Memory Scale Logical Memory II(22). In this paper, we considered early and late groups that each of them contains 12 subjects to create MCI. NC subjects do not have a history of neurological or psychiatric disorders; subjects with AD who have signatures of the NINCDS-ADRDA (National Institute of Neurological and Communicative Diseases and Stroke/Alzheimer's Disease and Related Disorders Association) criteria which means AD's disorder. Also, MCI subjects have symptoms a subjective memory concern; However, they were without any significant disability in other cognitive domains (they follow everyday activities with any signs of dementia). All scans that used in this study, were acquired by a 3T MRI Scanner (GE Medical Systems scanner). We collected structural MRI from all participants: T1- SPGE: TR=6.9 ms, TE=2.8 ms, slice thickness=1.2 mm; T2: TR=650 ms, TE=20 ms, slice thickness=4 mm; and FLAIR: TR=11000ms, TE=147.9 ms, slice thickness=5 mm; also, we collected diffusion MRI from all participants: matrix size = 256 \u0026times; 256 \u0026times; 46; voxel size = 1.36 \u0026times; 1.36 \u0026times; 2.7 mm; the sequence contained 5 images acquired without diffusion weighting and with diffusion weighting along 41 non-collinear directions (b = 1000 s/mm\u0026sup2;).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.2. Structural MRI processing\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe performed the processing of morphometrically and intensity level assessment by using the FreeSurfer image analysis pipeline (23) (version 6 on ubuntu 16.04) based on T1 and T2-FLAIR weighted images. The results of this process were divided into two parts that consisted of subcortical and cortical. In the first place, we performed the subcortical processing steps which included: volume, intensity pixel value, range of intensity, maximum, and minimum intensity pixel value in each region of atlas available in Freesurfer (Desikan-Killiany atlas). In the second place, the measurements related to cortical were done which included: thickness, surface area, gray matter volume and the curve in each region of Desikan-Killiany atlas from the right and left brain hemispheres. Part of the FreeSurfer pipeline conforms to the MRI scans to 1 \u0026times; 1 \u0026times; 1 mm\u003csup\u003e3\u003c/sup\u003e resolution and corrects for bias field using the non-parametric non-uniform intensity normalization (N3) algorithm.\u003c/p\u003e\n\u003cp\u003eBriefly, this processing includes removal of non-brain tissue, automated Talairach transformation,surface deformation, tessellation of the grey matter-white matter boundary, topology correction, and segmentation (24, 25).\u003c/p\u003e\n\u003cp\u003eFinally, after doing all the processing in this section, 496 features were extracted from every subject and saved in CSV format.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.3 Diffusion MRI processing\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe diffusion MRI data were preprocessed using DSI-Studio software (developed by Fang-Cheng Yeh from the Advanced Biomedical MRI Lab, National Taiwan University Hospital, Taiwan, Supported by Fiber Tractography Lab, University of Pittsburgh, and made available at \u003ca href=\"http://dsistudio\"\u003ehttp://dsistudio\u003c/a\u003e. labsolver.org/Download/). For skull stripping and filtering the background region, we used the masks provided by DSI-Studio. Before DTI parameter measurement, head motion and eddy current effects were corrected using the DSI-Studio toolbox. We used two different reconstruction methods include model base (DTI) and free model (QSDR); with two different attitudes to process the diffusion images. Four general feature categories were extracted from the images; the first type of features was diffusion-based parameters in each region of the structural freesurferseg and Brodmann atlases, the second type of features was tract-based parameters in each region of the same atlases and the third type of features was structural connectome with two atlases that were mentioned above. Finally, the last type of feature was the graph network.\u003c/p\u003e\n\u003cp\u003eThis process was performed for DTI model according to the following recommended parameters:\u003c/p\u003e\n\u003cp\u003eFA index was used to determine the fiber tracking threshold and Otsu\u0026rsquo;s method was used to set the anisotropy (FA) threshold to stop the fiber tracking, the number of seed points = 1,000,000, max angle = 60◦, step size = half of the spatial resolution, length constraint = 30\u0026ndash;300 mm, and no spatial smoothing. Also, DSI studio software was implemented to calculate the connections and graphs of the \"connectivity matrix and graph Network \" tool with two \"End\" and \"Pass\" views.\u003c/p\u003e\n\u003cp\u003eIn total, 92763 features were extracted from 72 participants. The features extracted of structural and diffusion processed MRI images in our work are given in the chart below. (Shown in Figure 1)\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.4. Feature selection and Classification \u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;Feature selection means recognizing the most relevant features for pattern recognition and noise reduction that usually used in studies where the number of data is extremely high while few studied subjects are available The Fast Correlation-Based Filter (FCBF) method was better than the other models among all the feature selection models because less than quadratic time complexity (17). FCBF selected 66 features for the next step.\u003c/p\u003e\n\u003cp\u003eAfter the feature selection step, we performed four classification models including the DT, KNN, SVM and ANN with 5, 10, 15, 20, 30, 40, 50-fold cross-validation were used to evaluate the performance of the classifier. To evaluating the ANN model, training, validation and test phase was repeated for 30 times and finally the mean and standard deviation were reported. For this purpose, data set was divided into three parts: 60% for training, 20% for validation, and 20% for test. The MATLAB software R2012 (The Math Works, USA) was used to design and implement classification models.\u003c/p\u003e"},{"header":"3. Result","content":"\u003cp\u003eThe number of extracted features using structural and diffusion processed MRI images for each subject were 92763.\u003c/p\u003e\n\u003cp\u003eFeatures extracted from the DTI model included region-based and tract-based features of structural freesurferseg and Brodmann atlases. Features extracted from QSDR model include Region-based, Tract- based, connectivity and Network features from structural freesurferseg and Brodmann atlases; Eventually, Features extracted from the structural MRI images from structural freesurferseg atlas include Morphometric and Voxel intensity features (Figure 1).\u003c/p\u003e\n\u003cp\u003eTable 1 shows the number of selected features with the FCBF model. According to this table, related features, Connectivity and Network have the most weight of selected features. However, it should be noted that the features selected for the three groups AD/MCI/NC classification. So, removing any of them has a significant effect on the classification performance. In total, 66 features were selected.\u003c/p\u003e\n\u003cp\u003eIn table 2, the performance of each classification model (DT, KNN, and SVM) is provided to classify three groups of AD / MCI / NC. These classification models were evaluated by cross validation fold of 5, 10, 15, 20, 30, 40, and 50. We have considered the accuracy factor to compare the performance of all classification models.\u003c/p\u003e\n\u003cp\u003eIn table 3 the performance of ANN is presented to classify three groups of AD / MCI / NC. This process is repeated 30 times and the mean and standard deviation are calculated.\u003c/p\u003e\n\u003cp\u003eWe run the SVM, DT, KNN and, ANN models for the diagnosis of AD/MCI/NC using MATLAB tools. These results showed good performance using selected features such as morphometry, network, connective, and tract base for the classifying AD/MCI/NC. Two highest accuracy for the classification of three groups (NC, MCI, and AD individuals) were obtained 89.9% and 90.7% for the SVM and ANN model respectively by 66 selected features.\u003c/p\u003e"},{"header":"4. Discussion","content":" \u003cp\u003eIn this study, we evaluated the structural and diffusion MRI brain images of the Alzheimer's patients from the ADNI dataset with machine learning techniques to classify the three groups AD, MCI and NC. For high-throughput imaging analysis, we used a well-established automated pipeline for morphometry (Freesurfer) and a pipeline to estimate rigorously individualized white matter structure (DSI_Studio).\u003c/p\u003e \u003cp\u003ethe study, which has used only structural image T1W ADNI databases, it was performed by Lauge Sorensen et al. In 2016, to simultaneously classify the three groups AD, MCI and NC by using the extracted features of the Freesurfer and Hippocampal texture area. The accuracy obtained for the three-group classification is 62.7%. The extracted areas are classified as Hippocampal, Parietal lobs, Cingulate cortex and ventricles (\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eBy comparing between our study and previous study: The study that only considers the diffusion image ADNI database for the classification of AD from NC is Luis R. Peraza et al., 2019, with an accuracy of 89.07% using 401 Connectivity feature extraction. For automated classification, only 89 features were selected between the two groups of AD and NC. In this study the regions involved in the temporal lobe, subcortical-posterior, occipital brain were introduced (\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eAlso, the study that used both structural and diffusion imaging to diagnose AD, MCI, and NC, it was performed by Yun Wang et al in 2019. The Structural images that were taken from the ADNI database, were processed by Freesurfer software and the diffusion images were processed by MRtrix3 software. The current results are also in broad agreement with a recent SVM study that used connectivity measures and morphometry. Accuracy for AD-NC, NC-MCI, and AD-MCI was 96%, 70%, and 75% respectively. For this classification, the selected areas are not mentioned (\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThe current study results show that it is possible to classify the three groups NC, MCI and AD subjects with a high degree of accuracy using an automated procedure that combines DTI, QSDR, and morphometry sMRI. Our results from NC and MCI and AD classification which achieved an average accuracy of 89.9% and 90.7% for the SVM and ANN model respectively. Due to the high standard deviation in the ANN model, the SVM model is reported as a result of this study. The results of the KNN and DT classification performance were not considered due to the low accuracy.\u003c/p\u003e \u003cp\u003eMany brain imaging studies have shown that some brain regions may play a key role in the diagnosis of MCI and AD that the present study can confirm these results.(\u003cspan additionalcitationids=\"CR28 CR29\" citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e)\u003c/p\u003e \u003cp\u003eThe extracted significant features could be used to find the most vulnerable regions of the brain to distinguish AD, MCI, and healthy elderly peoples. In the current study, the features of brain regions such as the caudate nucleus, corpus callosum, hippocampus, Para hippocampus, temporal gyrus, putamen nucleus, cingulate gyrus, the region of 36 and 3 Brodmann are the most important features for the separation of these three groups, especially the temporal lobe. Moreover, the Amygdala is the only brain region of which both left and right hemispheres are exposed to changes in imaging values between the three groups. According to similar studies, the areas are considered vulnerable and this data can provide a clinician with additional information.\u003c/p\u003e \u003cp\u003eTop 9 regions repetitive of WM based on four types of features are the caudate nucleus, corpus callosum, hippocampus, Para hippocampus, temporal gyrus, putamen nucleus, cingulate gyrus, and the regions of 36 and 3 Brodmann. These areas have the highest frequency in various measurements such as connectivity and network of this region with another area, the volume of tracts on these regions, and the thickness of regions. Generally, the features extracted from the Atlas of Structure were much more efficient than the Broadman Atlas.\u003c/p\u003e "},{"header":"5. Conclusion","content":" \u003cp\u003eIn this study we evaluated different three-group classification approaches using combine the neuroimaging modalities MRI and DTI. Combining anatomical and diffusion features, provided information indicating changes in the brain of patients with AD and MCI. Considering the results, three-group classification compared to two-group classifications used in above studies have significantly more important. The highest frequency in this study could be considered for identifying, monitoring, and future drug trials that could target these brain regions to AD and MCI Management. Eventually, the MRI can be combined with machine learning algorithms to detect damage in the early stages of Alzheimer\u0026rsquo;s disease.\u003c/p\u003e "},{"header":"Abbreviations","content":"\u003cp\u003eAD : Alzheimer\u0026rsquo;s disease\u003c/p\u003e\n\u003cp\u003eMCI : Mild cognitive impairment\u003c/p\u003e\n\u003cp\u003eDTI : Diffusion tensor imaging\u003c/p\u003e\n\u003cp\u003eDT : Decision Tree\u003c/p\u003e\n\u003cp\u003eKNN : k-nearest neighbors\u003c/p\u003e\n\u003cp\u003eSVM : support vector machines\u003c/p\u003e\n\u003cp\u003eANN : Artificial Neural Networks\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe study was reviewed and approved by the Ethical Committee of Mashhad University of Medical Sciences. (Ethical number: IR.MUMS.MEDICAL.REC.1397.410)\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u0026nbsp;:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and material :\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe data used to support the findings of this study are available from the corresponding author upon request.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests :\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding :\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study was financially supported by a grant from the Vice Chancellor of Research of Mashhad University of Medical Sciences; grant number 970508, But was not involved in the decision to publish.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthors' contributions :\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eDr. Hoda Zare Seyed Amir Zamanpour and Bahare Bigham conceived of the presented idea and developed the theory and performed the computations.\u003c/p\u003e\n\u003cp\u003eDr. Hoda Zare and verified the analytical methods and supervised the findings of this work.\u003c/p\u003e\n\u003cp\u003eDr. Hoda Zare , Seyed Amir Zamanpour and Mohamad-Hoseyn Sigari discussed the results and commented on the manuscript. All authors contributed to and have approved the final manuscript.\u003c/p\u003e\n\u003cp\u003eEach named author has substantially contributed to conducting the underlying research and drafting this manuscript. Additionally, to the best of our knowledge, the named authors have no conflict of interest, financial or otherwise.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eData used in the preparation of this article were obtained from the Alzheimer's Disease Neuroimaging Initiative (ADNI) database (\u003ca href=\"https://ida.loni.usc.edu/collaboration/access/adni.loni.usc.edu\"\u003eadni.loni.usc.edu\u003c/a\u003e). The ADNI was launched in 2003 as a public-private partnership, led by Principal Investigator Michael W. Weiner, MD. The primary goal of ADNI has been to test whether serial magnetic resonance imaging (MRI), positron emission tomography (PET), other biological markers, and clinical and neuropsychological assessment can be combined to measure the progression of mild cognitive impairment (MCI) and early Alzheimer's disease (AD). For up-to-date information, see\u0026nbsp;\u003ca href=\"https://ida.loni.usc.edu/collaboration/access/www.adni-info.org\"\u003ewww.adni-info.org\u003c/a\u003e.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eSutton AL. Alzheimer Disease Sourcebook: Basic Consumer Health Information about Alzheimer Disease and Other Forms of Dementia, Including Mild Cognitive Impairment, Corticobasal Degeneration, Dementia with Lewy Bodies, Frontotemporal Dementia, Huntington Disease, Parkinson Disease, and Vascular Dementia Along with Information about Recent Research on the Diagnosis and Prevention of Alzheimer Disease and Genetic Testing, Tips for Maintaining Cognitive Functioning, Strategies for Long-term Planning, Advice for Caregivers, a Glossary of Related Terms, and Directories of Resources for Additional Help and Information: Omnigraphics; 2011.\u003c/li\u003e\n\u003cli\u003eMeyer P, Rijntjes M, Hellwig S, Kl\u0026ouml;ppel S, Weiller C. Functional neuroimaging: functional magnetic resonance imaging, positron emission tomography, and single-photon emission computed tomography2016.\u003c/li\u003e\n\u003cli\u003eMaggipinto T, Bellotti R, Amoroso N, Diacono D, Donvito G, Lella E, et al. DTI measurements for Alzheimer\u0026rsquo;s classification. Physics in Medicine \u0026amp; Biology. 2017;62(6):2361.\u003c/li\u003e\n\u003cli\u003eLi M, Qin Y, Gao F, Zhu W, He X. Discriminative analysis of multivariate features from structural MRI and diffusion tensor images. Magnetic resonance imaging. 2014;32(8):1043-51.\u003c/li\u003e\n\u003cli\u003eWen Q, Mustafi SM, Li J, Risacher SL, Tallman E, Brown SA, et al. White matter alterations in early-stage Alzheimer's disease: A tract-specific study. Alzheimer's \u0026amp; Dementia: Diagnosis, Assessment \u0026amp; Disease Monitoring. 2019;11:576-87.\u003c/li\u003e\n\u003cli\u003eBigham B, Zamanpour SA, Zemorshidi F, Boroumand F, Zare H, Initiative AsDN. Identification of Superficial White Matter Abnormalities in Alzheimer\u0026rsquo;s Disease and Mild Cognitive Impairment Using Diffusion Tensor Imaging. Journal of Alzheimer's Disease Reports. (Preprint):1-11.\u003c/li\u003e\n\u003cli\u003eMori S. Introduction to diffusion tensor imaging: Elsevier; 2007.\u003c/li\u003e\n\u003cli\u003eAlexander AL, Lee JE, Lazar M, Field AS. Diffusion tensor imaging of the brain. Neurotherapeutics : the journal of the American Society for Experimental NeuroTherapeutics. 2007;4(3):316-29.\u003c/li\u003e\n\u003cli\u003eYeh F-C, Wedeen VJ, Tseng W-YI. Generalized ${q} $-sampling imaging. IEEE transactions on medical imaging. 2010;29(9):1626-35.\u003c/li\u003e\n\u003cli\u003eZhang H, Wang Y, Lu T, Qiu B, Tang Y, Ou S, et al. Differences between generalized q-sampling imaging and diffusion tensor imaging in the preoperative visualization of the nerve fiber tracts within peritumoral edema in brain. Neurosurgery. 2013;73(6):1044-53.\u003c/li\u003e\n\u003cli\u003eYeh F-C, Tseng W-YI. NTU-90: a high angular resolution brain atlas constructed by q-space diffeomorphic reconstruction. NeuroImage. 2011;58(1):91-9.\u003c/li\u003e\n\u003cli\u003eYeh F-C. Diffusion MRI Reconstruction in DSI Studio n.d. [Available from: \u003ca href=\"http://dsi-studio.labsolver.org/Manual/Reconstruction\"\u003ehttp://dsi-studio.labsolver.org/Manual/Reconstruction\u003c/a\u003e.\u003c/li\u003e\n\u003cli\u003eSchouten TM, Koini M, de Vos F, Seiler S, van der Grond J, Lechner A, et al. Combining anatomical, diffusion, and resting state functional magnetic resonance imaging for individual classification of mild and moderate Alzheimer's disease. NeuroImage Clinical. 2016;11:46-51.\u003c/li\u003e\n\u003cli\u003ede Vos F, Schouten TM, Hafkemeijer A, Dopper EG, van Swieten JC, de Rooij M, et al. Combining multiple anatomical MRI measures improves Alzheimer's disease classification. Human brain mapping. 2016;37(5):1920-9.\u003c/li\u003e\n\u003cli\u003eDai Z, Yan C, Wang Z, Wang J, Xia M, Li K, et al. Discriminative analysis of early Alzheimer's disease using multi-modal imaging and multi-level characterization with multi-classifier (M3). NeuroImage. 2012;59(3):2187-95.\u003c/li\u003e\n\u003cli\u003eSchouten TM, Koini M, Vos F, Seiler S, Rooij M, Lechner A, et al. Individual classification of Alzheimer's disease with diffusion magnetic resonance imaging. NeuroImage. 2017;152:476-81.\u003c/li\u003e\n\u003cli\u003eYu L, Liu H, editors. Feature selection for high-dimensional data: A fast correlation-based filter solution. Proceedings of the 20th international conference on machine learning (ICML-03); 2003.\u003c/li\u003e\n\u003cli\u003eShalev-Shwartz S, Ben-David S. Understanding machine learning: From theory to algorithms: Cambridge university press; 2014.\u003c/li\u003e\n\u003cli\u003eGuyon I, Elisseeff A. An introduction to variable and feature selection. Journal of machine learning research. 2003;3(Mar):1157-82.\u003c/li\u003e\n\u003cli\u003eHandelman G, Kok H, Chandra R, Razavi A, Lee M, Asadi H. eD octor: machine learning and the future of medicine. Journal of internal medicine. 2018;284(6):603-19.\u003c/li\u003e\n\u003cli\u003eCleophas TJ, Zwinderman AH. Machine Learning in Medicine-a Complete Overview: Springer; 2015.\u003c/li\u003e\n\u003cli\u003eWechsler D. Wechsler memory scale-revised. Psychological Corporation. 1987.\u003c/li\u003e\n\u003cli\u003eFischl B. FreeSurfer. NeuroImage. 2012;62(2):774-81.\u003c/li\u003e\n\u003cli\u003eS\u0026oslash;rensen L, Igel C, Pai A, Balas I, Anker C, Lillholm M, et al. Differential diagnosis of mild cognitive impairment and Alzheimer's disease using structural MRI cortical thickness, hippocampal shape, hippocampal texture, and volumetry. NeuroImage: Clinical. 2017;13:470-82.\u003c/li\u003e\n\u003cli\u003eWang Y, Xu C, Park J-H, Lee S, Stern Y, Yoo S, et al. Diagnosis and prognosis of Alzheimer's disease using brain morphometry and white matter connectomes. NeuroImage: Clinical. 2019;23:101859.\u003c/li\u003e\n\u003cli\u003ePeraza LR, D\u0026iacute;az-Parra A, Kennion O, Moratal D, Taylor J-P, Kaiser M, et al. Structural connectivity centrality changes mark the path toward Alzheimer's disease. Alzheimer's \u0026amp; Dementia: Diagnosis, Assessment \u0026amp; Disease Monitoring. 2019;11:98-107.\u003c/li\u003e\n\u003cli\u003eFellgiebel A, Yakushev I. Diffusion tensor imaging of the hippocampus in MCI and early Alzheimer's disease. Journal of Alzheimer's disease : JAD. 2011;26 Suppl 3:257-62.\u003c/li\u003e\n\u003cli\u003eHuang J, Friedland R, Auchus A. Diffusion tensor imaging of normal-appearing white matter in mild cognitive impairment and early Alzheimer disease: preliminary evidence of axonal degeneration in the temporal lobe. American Journal of Neuroradiology. 2007;28(10):1943-8.\u003c/li\u003e\n\u003cli\u003ePetersen RC, Smith GE, Waring SC, Ivnik RJ, Tangalos EG, Kokmen E. Mild cognitive impairment: clinical characterization and outcome. Archives of neurology. 1999;56(3):303-8.\u003c/li\u003e\n\u003cli\u003eZhang Y, Schuff N, Jahng GH, Bayne W, Mori S, Schad L, et al. Diffusion tensor imaging of cingulum fibers in mild cognitive impairment and Alzheimer disease. Neurology. 2007;68(1):13-9.\u003c/li\u003e\n\u003c/ol\u003e"},{"header":"Tables","content":"\u003cp\u003eTable 1:\u0026nbsp; The number of selected features with the FCBF feature selection model\u003c/p\u003e\n\u003ctable border=\"1\"\u003e\n\u003ctbody\u003e\n\u003ctr\u003e\n\u003ctd width=\"153\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"164\"\u003e\n\u003cp\u003eType of features\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"153\"\u003e\n\u003cp\u003eNumber of selected features\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"153\"\u003e\n\u003cp\u003ePercentage of total selected features\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd rowspan=\"2\" width=\"153\"\u003e\n\u003cp\u003eDTI Model\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"164\"\u003e\n\u003cp\u003eRegion Base\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"153\"\u003e\n\u003cp\u003e4\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"153\"\u003e\n\u003cp\u003e6%\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"164\"\u003e\n\u003cp\u003eTract Base\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"153\"\u003e\n\u003cp\u003e9\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"153\"\u003e\n\u003cp\u003e13.6%\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd rowspan=\"4\" width=\"153\"\u003e\n\u003cp\u003eQSDR Model\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"164\"\u003e\n\u003cp\u003eRegion base\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"153\"\u003e\n\u003cp\u003e3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"153\"\u003e\n\u003cp\u003e4.6%\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"164\"\u003e\n\u003cp\u003eTract base\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"153\"\u003e\n\u003cp\u003e1\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"153\"\u003e\n\u003cp\u003e1.6%\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"164\"\u003e\n\u003cp\u003eConnectivity\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"153\"\u003e\n\u003cp\u003e30\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"153\"\u003e\n\u003cp\u003e45.5%\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"164\"\u003e\n\u003cp\u003eNetwork\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"153\"\u003e\n\u003cp\u003e11\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"153\"\u003e\n\u003cp\u003e16.7%\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd rowspan=\"2\" width=\"153\"\u003e\n\u003cp\u003eStructural MRI\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"164\"\u003e\n\u003cp\u003eCortical Region\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"153\"\u003e\n\u003cp\u003e5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"153\"\u003e\n\u003cp\u003e7.5%\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"164\"\u003e\n\u003cp\u003eSub-Cortical Region\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"153\"\u003e\n\u003cp\u003e3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"153\"\u003e\n\u003cp\u003e4.5%\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTable 2: the performance of DT, KNN, and SVM with cross validation fold in term of the accuracy factor.\u003c/p\u003e\n\u003ctable border=\"1\" width=\"0\"\u003e\n\u003ctbody\u003e\n\u003ctr\u003e\n\u003ctd width=\"89\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd width=\"89\"\u003e\n\u003cp\u003ecross validation fold\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"52\"\u003e\n\u003cp\u003e5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e10\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e15\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e20\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e30\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e40\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e50\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003eaverage\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003eSd\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd rowspan=\"4\" width=\"89\"\u003e\n\u003cp\u003eDT\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"89\"\u003e\n\u003cp\u003eCN\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"52\"\u003e\n\u003cp\u003e54.2\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e54.2\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e50\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e58.3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e50\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e58.4\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e45.8\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e52.9\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e4.6\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"89\"\u003e\n\u003cp\u003eMCI\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"52\"\u003e\n\u003cp\u003e62.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e62.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e50\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e41.7\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e62.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e66.6\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e62.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e58.3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e8.9\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"89\"\u003e\n\u003cp\u003eAD\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"52\"\u003e\n\u003cp\u003e70.8\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e70.8\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e75\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e70.8\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e66.6\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e70.8\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e75\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e71.4\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e2.8\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"89\"\u003e\n\u003cp\u003eall\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"52\"\u003e\n\u003cp\u003e\u003cstrong\u003e62.5\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e\u003cstrong\u003e62.5\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e\u003cstrong\u003e58.3\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e\u003cstrong\u003e56.9\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e\u003cstrong\u003e59.7\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e\u003cstrong\u003e65.3\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e\u003cstrong\u003e61.1\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e\u003cstrong\u003e60.9\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e\u003cstrong\u003e2.8\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd rowspan=\"4\" width=\"89\"\u003e\n\u003cp\u003eKNN\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"89\"\u003e\n\u003cp\u003eCN\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"52\"\u003e\n\u003cp\u003e83.3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e95.8\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e95.8\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e95.8\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e95.8\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e95.8\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e95.8\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e94.0\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e4.7\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"89\"\u003e\n\u003cp\u003eMCI\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"52\"\u003e\n\u003cp\u003e79.2\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e45.8\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e54.2\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e51.2\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e54.2\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e66.6\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e58.3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e58.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e11.1\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"89\"\u003e\n\u003cp\u003eAD\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"52\"\u003e\n\u003cp\u003e83.3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e91.7\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e95.8\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e91.7\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e91.6\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e91.7\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e91.7\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e91.0\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e3.7\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"89\"\u003e\n\u003cp\u003eall\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"52\"\u003e\n\u003cp\u003e\u003cstrong\u003e81.9\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e\u003cstrong\u003e77.8\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e\u003cstrong\u003e81.9\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e\u003cstrong\u003e80.6\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e\u003cstrong\u003e80.6\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e\u003cstrong\u003e84.7\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e\u003cstrong\u003e81.9\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e\u003cstrong\u003e81.3\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e\u003cstrong\u003e2.0\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd rowspan=\"4\" width=\"89\"\u003e\n\u003cp\u003eSVM\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"89\"\u003e\n\u003cp\u003eCN\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"52\"\u003e\n\u003cp\u003e95.8\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e95.8\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e91.6\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e95.8\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e95.8\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e95.8\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e91.7\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e94.6\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e2.0\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"89\"\u003e\n\u003cp\u003eMCI\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"52\"\u003e\n\u003cp\u003e83.3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e87.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e87.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e87.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e83.3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e83.3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e87.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e85.7\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e2.2\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"89\"\u003e\n\u003cp\u003eAD\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"52\"\u003e\n\u003cp\u003e87.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e91.6\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e91.6\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e87.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e91.7\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e87.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e87.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e89.2\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e2.2\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"89\"\u003e\n\u003cp\u003eall\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"52\"\u003e\n\u003cp\u003e\u003cstrong\u003e88.9\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e\u003cstrong\u003e91.7\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e\u003cstrong\u003e90.3\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e\u003cstrong\u003e90.3\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e\u003cstrong\u003e90.3\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e\u003cstrong\u003e88.9\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e\u003cstrong\u003e88.9\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e\u003cstrong\u003e89.9\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"50\"\u003e\n\u003cp\u003e\u003cstrong\u003e1.0\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTable 3: the performance of ANN in term of the accuracy factor and the mean and standard deviation\u003c/p\u003e\n\u003ctable border=\"1\" width=\"0\"\u003e\n\u003ctbody\u003e\n\u003ctr\u003e\n\u003ctd width=\"64\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd colspan=\"4\" width=\"256\"\u003e\n\u003cp\u003eANN\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003eindex\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003eCN\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003eMCI\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003eAD\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003eAll\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e1\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e100\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e83.3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e91.7\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e91.7\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e17\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e83.3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e87.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e100\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e90.3\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e2\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e95.8\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e83.3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e91.7\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e90.3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e18\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e95.8\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e87.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e91.7\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e91.7\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e100\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e83.3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e91.7\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e91.7\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e19\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e83.3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e87.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e91.7\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e87.5\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e4\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e91.7\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e83.3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e100\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e91.7\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e20\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e91.7\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e79.2\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e91.7\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e87.5\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e87.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e83.3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e91.7\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e87.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e21\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e100\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e87.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e95.8\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e94.4\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e6\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e95.8\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e87.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e95.8\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e93.1\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e22\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e100\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e87.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e95.8\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e94.4\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e7\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e91.7\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e83.3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e83.3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e86.1\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e23\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e91.7\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e95.8\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e91.7\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e93.1\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e8\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e95.8\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e95.8\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e95.8\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e95.8\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e24\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e91.7\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e91.7\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e95.8\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e93.1\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e9\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e91.7\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e83.3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e95.8\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e90.3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e25\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e95.8\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e91.7\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e100\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e95.8\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e10\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e100\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e70.8\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e100\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e90.3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e26\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e91.7\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e87.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e95.8\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e91.7\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e11\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e95.8\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e91.7\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e95.8\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e94.4\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e27\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e95.8\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e54.2\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e91.7\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e80.6\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e12\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e91.7\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e83.3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e91.7\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e88.9\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e28\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e100\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e66.7\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e100\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e88.9\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e13\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e91.7\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e87.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e95.8\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e91.7\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e29\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e83.3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e87.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e75\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e81.9\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e14\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e95.8\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e83.3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e100\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e93.1\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e30\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e100\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e70.8\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e100\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e90.3\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e15\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e100\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e79.2\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e95.8\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e91.7\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003emean\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e\u003cstrong\u003e94.3\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e\u003cstrong\u003e83.7\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e\u003cstrong\u003e94.0\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e\u003cstrong\u003e90.7\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e16\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e100\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e87.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e87.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e91.7\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003eSd\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e\u003cstrong\u003e5.1\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e\u003cstrong\u003e8.6\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e\u003cstrong\u003e5.4\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"64\"\u003e\n\u003cp\u003e\u003cstrong\u003e3.5\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/tbody\u003e\n\u003c/table\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"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, Mild Cognitive Impairments, Magnetic Resonance Imaging, Machine Learning, Diffusion Weighted imaging, Structural MRI imaging","lastPublishedDoi":"10.21203/rs.3.rs-573090/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-573090/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003eIntroduction: \u003c/strong\u003eAccurate, fast, and reliable diagnosis of Alzheimer's Disease (AD) from Mild Cognitive Impairment (MCI) is crucial for prescribing proper treatment and prevention of disease progression. \u0026nbsp;At first glance, structural and diffusion MRI images, are affected by neurodegenerative proceedings in AD and MCI. In this study, we are looking for the most effective features to detect and differentiate between healthy normal control (NC), AD, and MCI groups by non-invasive Magnetic Resonance Imaging (MRI) method and propose the automatic multi-class classification using the structural and diffusion MRI Features of the brain. \u003c/p\u003e\u003cp\u003e\u003cstrong\u003eMethods:\u003c/strong\u003e The structural and diffusion MRI data were downloaded from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) database on three groups including AD, MCI, and NC subjects. Four famous classification models of machine learning were used to discover the best classification as a diagnostic tool for separation of the NC, AD and MCI groups. \u003c/p\u003e\u003cp\u003e\u003cstrong\u003eResults:\u003c/strong\u003e Taken together, our results from this study lead to classify three groups for differentiation between the NC group and patients with MCI and AD, with average accuracy factor 89.9% for Support Vector Machine (SVM) and 91.9% for Artificial Neural Network (ANN) using selected features. \u003c/p\u003e\u003cp\u003e\u003cstrong\u003eConclusions:\u003c/strong\u003e Top 9 regions repetitive of WM based on four types of features are the caudate nucleus, corpus callosum, hippocampus, para hippocampus, temporal gyrus, putamen nucleus, cingulate gyrus, the region of 36 and 3 Brodmann. Therefore, these regions could be considered for identifying, monitoring, and future drug trials that could target this brain region to AD and MCI Management.\u003c/p\u003e","manuscriptTitle":"Differential Diagnosis of Alzheimer's Disease and Mild Cognitive Impairment Using the Structural and Diffusion MRI Features of the Brain","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2021-06-09 21:45:26","doi":"10.21203/rs.3.rs-573090/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":"4b96d1e5-2550-4af1-9f96-582ea39bf83d","owner":[],"postedDate":"June 9th, 2021","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":4899953,"name":"Nuclear Medicine \u0026 Medical Imaging"}],"tags":[],"updatedAt":"2021-07-21T10:59:05+00:00","versionOfRecord":[],"versionCreatedAt":"2021-06-09 21:45:26","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-573090","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-573090","identity":"rs-573090","version":["v1"]},"buildId":"WrCJVZZCHTDjtuVLN7oU0","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}