Spatial Navigation as a Digital Marker for Clinically Differentiating Cognitive Impairment Severity

Spatial Navigation as a Digital Marker for Clinically Differentiating Cognitive Impairment Severity | medRxiv /* */ /* */ <!-- <!-- /*! * yepnope1.5.4 * (c) WTFPL, GPLv2 */ (function(a,b,c){function d(a){return"[object Function]"==o.call(a)}function e(a){return"string"==typeof a}function f(){}function g(a){return!a||"loaded"==a||"complete"==a||"uninitialized"==a}function h(){var a=p.shift();q=1,a?a.t?m(function(){("c"==a.t?B.injectCss:B.injectJs)(a.s,0,a.a,a.x,a.e,1)},0):(a(),h()):q=0}function i(a,c,d,e,f,i,j){function k(b){if(!o&&g(l.readyState)&&(u.r=o=1,!q&&h(),l.onload=l.onreadystatechange=null,b)){"img"!=a&&m(function(){t.removeChild(l)},50);for(var d in y[c])y[c].hasOwnProperty(d)&&y[c][d].onload()}}var j=j||B.errorTimeout,l=b.createElement(a),o=0,r=0,u={t:d,s:c,e:f,a:i,x:j};1===y[c]&&(r=1,y[c]=[]),"object"==a?l.data=c:(l.src=c,l.type=a),l.width=l.height="0",l.onerror=l.onload=l.onreadystatechange=function(){k.call(this,r)},p.splice(e,0,u),"img"!=a&&(r||2===y[c]?(t.insertBefore(l,s?null:n),m(k,j)):y[c].push(l))}function j(a,b,c,d,f){return q=0,b=b||"j",e(a)?i("c"==b?v:u,a,b,this.i++,c,d,f):(p.splice(this.i++,0,a),1==p.length&&h()),this}function k(){var a=B;return a.loader={load:j,i:0},a}var l=b.documentElement,m=a.setTimeout,n=b.getElementsByTagName("script")[0],o={}.toString,p=[],q=0,r="MozAppearance"in l.style,s=r&&!!b.createRange().compareNode,t=s?l:n.parentNode,l=a.opera&&"[object Opera]"==o.call(a.opera),l=!!b.attachEvent&&!l,u=r?"object":l?"script":"img",v=l?"script":u,w=Array.isArray||function(a){return"[object Array]"==o.call(a)},x=[],y={},z={timeout:function(a,b){return b.length&&(a.timeout=b[0]),a}},A,B;B=function(a){function b(a){var a=a.split("!"),b=x.length,c=a.pop(),d=a.length,c={url:c,origUrl:c,prefixes:a},e,f,g;for(f=0;f<d;f++)g=a[f].split("="),(e=z[g.shift()])&&(c=e(c,g));for(f=0;f<b;f++)c=x[f](c);return c}function g(a,e,f,g,h){var i=b(a),j=i.autoCallback;i.url.split(".").pop().split("?").shift(),i.bypass||(e&&(e=d(e)?e:e[a]||e[g]||e[a.split("/").pop().split("?")[0]]),i.instead?i.instead(a,e,f,g,h):(y[i.url]?i.noexec=!0:y[i.url]=1,f.load(i.url,i.forceCSS||!i.forceJS&&"css"==i.url.split(".").pop().split("?").shift()?"c":c,i.noexec,i.attrs,i.timeout),(d(e)||d(j))&&f.load(function(){k(),e&&e(i.origUrl,h,g),j&&j(i.origUrl,h,g),y[i.url]=2})))}function h(a,b){function c(a,c){if(a){if(e(a))c||(j=function(){var a=[].slice.call(arguments);k.apply(this,a),l()}),g(a,j,b,0,h);else if(Object(a)===a)for(n in m=function(){var b=0,c;for(c in a)a.hasOwnProperty(c)&&b++;return b}(),a)a.hasOwnProperty(n)&&(!c&&!--m&&(d(j)?j=function(){var a=[].slice.call(arguments);k.apply(this,a),l()}:j[n]=function(a){return function(){var b=[].slice.call(arguments);a&&a.apply(this,b),l()}}(k[n])),g(a[n],j,b,n,h))}else!c&&l()}var h=!!a.test,i=a.load||a.both,j=a.callback||f,k=j,l=a.complete||f,m,n;c(h?a.yep:a.nope,!!i),i&&c(i)}var i,j,l=this.yepnope.loader;if(e(a))g(a,0,l,0);else if(w(a))for(i=0;i (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];var j=d.createElement(s);var dl=l!='dataLayer'?'&l='+l:'';j.src='//www.googletagmanager.com/gtm.js?id='+i+dl;j.type='text/javascript';j.async=true;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-P4HH5NV'); Skip to main content Home About Submit ALERTS / RSS Search for this keyword Advanced Search Spatial Navigation as a Digital Marker for Clinically Differentiating Cognitive Impairment Severity View ORCID Profile Giorgio Colombo , Karolina Minta , View ORCID Profile William R. Taylor , View ORCID Profile Jascha Grübel , Eddie Chong , Joyce R. Chong , Mark J.H. Lim , Paul Nichol G. Gonzales , Mitchell K. P. Lai , View ORCID Profile Christopher P. Chen , View ORCID Profile Victor R. Schinazi doi: https://doi.org/10.1101/2025.06.05.25329079 Giorgio Colombo 1 Future Health Technologies, Singapore-ETH Centre, Campus for Research Excellence And Technological Enterprise (CREATE) , Singapore 2 Chair of Cognitive Science , ETH Zurich, Zürich, 8092, Switzerland Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Giorgio Colombo For correspondence: gicolombo{at}ethz.ch Karolina Minta 1 Future Health Technologies, Singapore-ETH Centre, Campus for Research Excellence And Technological Enterprise (CREATE) , Singapore 3 Department of Pharmacology, Yong Loo Lin School of Medicine, National University of Singapore, Singapore Find this author on Google Scholar Find this author on PubMed Search for this author on this site William R. Taylor 1 Future Health Technologies, Singapore-ETH Centre, Campus for Research Excellence And Technological Enterprise (CREATE) , Singapore 4 Institute for Biomechanics, Department of Health Sciences and Technology , ETH Zurich, Zürich, Switzerland Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for William R. Taylor Jascha Grübel 5 Center for Sustainable Future Mobility , ETH Zurich, Zürich, 8092, Switzerland 6 Laboratory of Geo-information Science and Remote Sensing, Wageningen University , Wageningen, 6708 PB, The Netherlands Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Jascha Grübel Eddie Chong 3 Department of Pharmacology, Yong Loo Lin School of Medicine, National University of Singapore, Singapore 7 Memory, Aging and Cognition Centre, National University Health System , Singapore Find this author on Google Scholar Find this author on PubMed Search for this author on this site Joyce R. Chong 3 Department of Pharmacology, Yong Loo Lin School of Medicine, National University of Singapore, Singapore 7 Memory, Aging and Cognition Centre, National University Health System , Singapore Find this author on Google Scholar Find this author on PubMed Search for this author on this site Mark J.H. Lim 3 Department of Pharmacology, Yong Loo Lin School of Medicine, National University of Singapore, Singapore 7 Memory, Aging and Cognition Centre, National University Health System , Singapore 8 Healthy Longevity Translational Research Programme, Yong Loo Lin School of Medicine, National University of Singapore , Singapore, Singapore Find this author on Google Scholar Find this author on PubMed Search for this author on this site Paul Nichol G. Gonzales 7 Memory, Aging and Cognition Centre, National University Health System , Singapore Find this author on Google Scholar Find this author on PubMed Search for this author on this site Mitchell K. P. Lai 3 Department of Pharmacology, Yong Loo Lin School of Medicine, National University of Singapore, Singapore 7 Memory, Aging and Cognition Centre, National University Health System , Singapore 8 Healthy Longevity Translational Research Programme, Yong Loo Lin School of Medicine, National University of Singapore , Singapore, Singapore Find this author on Google Scholar Find this author on PubMed Search for this author on this site Christopher P. Chen 3 Department of Pharmacology, Yong Loo Lin School of Medicine, National University of Singapore, Singapore 7 Memory, Aging and Cognition Centre, National University Health System , Singapore 8 Healthy Longevity Translational Research Programme, Yong Loo Lin School of Medicine, National University of Singapore , Singapore, Singapore Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Christopher P. Chen Victor R. Schinazi 1 Future Health Technologies, Singapore-ETH Centre, Campus for Research Excellence And Technological Enterprise (CREATE) , Singapore 9 Department of Psychology, Bond University , Gold Coast, Queensland, Australia Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Victor R. Schinazi Abstract Full Text Info/History Metrics Supplementary material Data/Code Preview PDF Abstract Navigation impairments emerge early in Alzheimer’s disease, but assessments targeting these deficits remain underutilised or impractical for cognitive screening. The Spatial Performance Assessment for Cognitive Evaluation (SPACE) is a digital tool that evaluates spatial navigation deficits associated with cognitive impairment. In a memory clinic and community cohort (n = 300), SPACE reliably distinguished Clinical Dementia Rating levels, exceeding the accuracy of demographic models and matching or surpassing most traditional neuropsychological tests. Including SPACE significantly increased the AUC for distinguishing between no dementia from mild dementia (0.76 to 0.94), no dementia from moderate dementia (0.79 to 0.95), and questionable dementia from mild dementia (0.70 to 0.91), all with consistently high sensitivity and specificity. A short version of SPACE (< 11 minutes) reduced administration time by 40% while maintaining high diagnostic accuracy. These findings highlight the potential of digital navigation assessments to advance early detection, contributing to scalable and accessible healthcare. Introduction Dementia affects approximately 55 million people worldwide, and this number is projected to reach 150 million by 2050 1 , 2 . The global economic burden of dementia is estimated at $16.9 trillion, imposing substantial financial and social strain on individuals, families, and healthcare systems 3 . Alzheimer’s Disease (AD) is the leading cause of dementia and is characterised by progressive cognitive decline. AD typically unfolds across distinct stages, from Mild Cognitive Impairment (MCI) 4 , where subtle memory and navigation difficulties emerge, to moderate and severe stages characterised by profound impairments in multiple cognitive domains, including memory, spatial orientation, and daily functioning 2 , 5 – 9 . Identifying individuals with MCI provides an opportunity to delay disease progression through targeted interventions 8 – 11 . Episodic memory impairments have long been the hallmark of MCI and AD, forming the cornerstone of diagnostic criteria. Current neuropsychological assessments typically focus on the detection of memory deficits alongside impairments in other cognitive domains (e.g., attention, executive function). However, the limited sensitivity of some standard screening tools frequently hinders the detection of early or subtle cognitive changes, particularly in the preclinical stages of neurodegenerative disease 11 – 16 . As a result, many diagnoses are still made only after significant pathologic processes has already occurred, when interventions are limited. While blood biomarkers simplify early detection compared to Positron Emission Tomography (PET) and Cerebrospinal Fluid (CSF) methods, their reliance on clinic-based sampling and interpretation does not fully overcome the constraints of late-stage, in-clinic diagnosis. The early stages of AD are characterised by neurodegenerative changes in the hippocampus and the entorhinal cortex, including neuronal loss as well as the accumulation of amyloid-beta (Aβ) peptides and phosphorylated tau (p-tau) protein 17 – 21 . Critically, these subcortical regions also play an essential role in coding spatial information in the environment 22 and in assisting individuals to keep track of changes in position and orientation during navigation 23 . Indeed, an accumulating body of evidence from real-world 24 – 26 and Virtual Reality (VR) 27 – 32 studies has identified spatial navigation deficits as a promising marker of genetic risk for sporadic AD (APOE ε4-carriers) 33 – 35 and a reliable means of distinguishing individuals across the spectrum of cognitive impairment 16 , 36 . Here, experimental paradigms that require individuals to learn and recall the configuration of landmarks in an environment (i.e. cognitive mapping) 24 , 31 , 37 , maintain orientation and track self-motion (i.e. path integration) 27 , 28 , 32 , 38 , as well as infer landmark positions from different viewpoints (i.e. perspective taking) 39 , 40 have all exhibited sensitivity to early cognitive impairment. Among these tasks, path integration deficits have been consistently associated with hippocampal 24 , 31 , 37 and entorhinal cortex 28 , 34 atrophy, along with reduced activity in these brain areas 33 , 34 . Advances in digital technologies have enabled the development of scalable, gamified, spatial navigation assessments that offer engaging 41 , ecologically valid 42 , and accessible tools for early cognitive screening 43 – 45 . Such digital platforms offer remote deployment, reducing reliance on specialised clinical settings and costly visits, while enabling continuous monitoring of disease progression 8 , 9 , 46 , 47 . Despite strong theoretical and empirical support for spatial navigation as an early marker of cognitive decline, most existing tools have yet to be validated against a gold standard diagnosis. As a result, few, if any, digital navigation tools have transitioned into clinical practice, representing a missed opportunity for early detection and intervention. In this study, we investigate whether the Spatial Performance Assessment for Cognitive Evaluation (SPACE) 48 , deployed on a tablet (i.e. iPad), can discriminate between patients with various levels of cognitive impairment, as independently assessed by clinicians. SPACE assesses spatial navigation abilities through five tasks, including path integration, egocentric pointing, mapping, associative memory, and perspective taking. The different tasks in SPACE require participants to navigate a virtual environment, keep track of their position and orientation, reconstruct spatial layouts, recall object-location associations, and estimate spatial relationships from various perspectives. We administered SPACE to 300 individuals referred from memory clinics and the community and compared their performance with the Clinical Dementia Rating (CDR) scale and a battery of neuropsychological tests. Results Group differences in spatial navigation performance Table 1 summarises the demographic and clinical characteristics of the cohort, as well as their performance in neuropsychological tests and SPACE across CDR levels (CDR 0 = no dementia; CDR 0.5 = questionable dementia; CDR 1 = mild dementia; CDR 2 = moderate dementia; CDR 3 = severe dementia). Due to the very small number of participants (n = 3) with a CDR score of 3, the CDR 2 and CDR 3 groups were combined for analysis (see Supplementary Information B). We performed robust ANOVAs to assess differences in spatial navigation performance across CDR groups (CDR 0, CDR 0.5, CDR 1, and CDR 2+) for each task in SPACE ( Figure 1 and Supplementary Table 2). Given that only two participants with a CDR of 2+ completed the path integration, pointing, and mapping tasks, this group were excluded from the analysis for these tasks. Results revealed significant CDR group differences in training time ( F (3, 296) = 14.7, p < 0.001, ξ = 0.657), path integration distance error ( F (3, 296) = 7.0, p = 0.004, ξ = 0.587), and perspective taking error ( F (3, 296) = 11.0, p < 0.001, ξ = 0.51). No significant group differences were observed for egocentric pointing error ( F (3, 296) = 2.61, p = 0.101) or mapping accuracy ( F (3, 296) = 1.33, p = 0.291). Download figure Open in new tab Figure 1 | Between-group comparisons across Clinical Dementia Rating (CDR) levels for (a) Training time, (b) Path integration distance, and (c) Perspective taking error. Boxplots display group medians and interquartile ranges, with individual data points overlaid. Asterisks indicate levels of statistical significance (* p < 0.05; ** p < 0.01; *** p < 0.001). View this table: View inline View popup Table 1 | Sample characteristics. Demographic, clinical, neuropsychological, and task-performance characterisation of the sample by CDR status. Post-hoc comparisons revealed that participants with CDR 0 completed the training phase significantly faster than those with CDR 0.5 ( p = 0.022, ξ = 0.229), CDR 1 ( p < 0.001, ξ = 0.717), and CDR 2+ ( p < 0.001, ξ = 0.886). The CDR 0.5 group was also significantly faster than both the CDR 1 ( p = 0.001, ξ = 0.419) and CDR 2+ ( p < 0.001, ξ = 0.761) groups. Finally, participants with CDR 1 were faster than those with CDR 2+ ( p = 0.022, ξ = 0.483). For path integration distance error, participants with CDR 0 performed significantly better than those with CDR 1 ( p < 0.001, ξ = 0.666), but not better than those with CDR 0.5 ( p = 0.44). Participants with CDR 0.5 also outperformed those with CDR 1 ( p < 0.001, ξ = 0.714). For perspective taking error, participants with CDR 0 performed significantly better than those with CDR 1 ( p < 0.001, ξ = 0.496) and CDR 2+ ( p < 0.001, ξ = 0.661) but did not differ from those with CDR 0.5 ( p = 0.572). Participants with CDR 0.5 also outperformed those with CDR 1 ( p < 0.001, ξ = 0.518) and CDR 2+ ( p < 0.001, ξ = 0.613). No differences emerged between participants with CDR 1 and CDR 2+ ( p = 0.476). Diagnostic classification performance We conducted logistic regression analyses to evaluate the diagnostic utility of the SPACE measures (i.e. training, path integration, and perspective taking), which showed significant effects in the previous ANOVAs. These analyses aimed to classify participants across CDR stages by comparing between diagnostic groups. Here, we compared baseline models that included demographic variables (i.e. age, education, and gender) to models that additionally incorporated the significant SPACE metrics. We assessed improvements in model fit using likelihood-ratio tests (see Table 2 and Supplementary Table 3 for full model statistics) and evaluated classification accuracy using the Area Under the Curve (AUC), sensitivity, and specificity. View this table: View inline View popup Download powerpoint Table 2 | Model fit and classification performance. Comparison of model fit and predictive accuracy for distinguishing clinical groups, with and without the inclusion of SPACE measures. View this table: View inline View popup Download powerpoint Table 3 | Contribution of SPACE measures to model fit. Likelihood ratio tests comparing baseline demographic models to extended models including SPACE measures across clinical group contrasts. Incorporating SPACE measures improved model performance for all diagnostic contrasts except for CDR 0 vs. 0.5 ( Figure 2 ). Importantly, these measures improved classification not only for CDR 2+ but also for early stages, including CDR 0.5 and CDR 1 (Tables 2 & 3). Comparison between CDR 0 and CDR 0.5 produced the smallest improvements. Here, the demographic model yielded an AUC of 0.57, with sensitivity of 0.40 and specificity of 0.75. Adding SPACE raised the AUC slightly to 0.61 ( p DeLong = 0.256), sensitivity to 0.47 and specificity to 0.79, but this change was not statistically significant (Δχ² = 4.88, p = 0.181). Greater improvements were achieved for all the other five models. As expected, the addition of SPACE to contrasts involving the CDR 2+ group resulted in considerable gains in diagnostic performance. More importantly, however, for the comparison between CDR 0 and CDR 1, the demographic model achieved an AUC of 0.76, with sensitivity of 0.74 and specificity of 0.71. Critically, adding SPACE significantly improved model fit (Δχ² = 34.7, p < 0.001), increasing the AUC to 0.94 ( p DeLong < .001), sensitivity to 1.00, and specificity to 0.85. Of similar clinical importance was the differentiation between CDR 0.5 from CDR 1, where the demographic model produced an AUC of 0.70, with sensitivity of 0.74 and specificity of 0.59. Again, the addition of SPACE significantly improved the model (Δχ² = 32.9, p < 0.001), increasing the AUC to 0.91 ( p DeLong <.001), sensitivity to 0.95 and specificity to 0.73. Download figure Open in new tab Figure 2 | Receiver Operating Characteristic (ROC) curves comparing pairwise classification performance across contrasts between cognitive stages: Dem (sociodemographic variables, blue), SPACE (performance on training, path integration, and perspective taking tasks, red) and Total ( Dem + SPACE , orange). AUC values and 95% confidence intervals (CI) are reported for each model. Cutoff values reflect optimal thresholds based on the Youden Index. The diagonal dashed lines indicate chance-level performance. Cross-validation performance To further evaluate the classification performance of SPACE, we performed stratified k-fold cross-validation (k = 10) and Leave-One-Out Cross-Validation (LOOCV) for each of the contrasts ( Table 4 ). In both analyses, SPACE retained high diagnostic accuracy across most group comparisons. As expected, classification between CDR 0 and CDR 0.5 produced the weakest results. However, performance remained strong in all other clinically relevant contrasts. For distinguishing CDR 0 from CDR 1, k-fold cross-validation produced an AUC of 0.87 with sensitivity of 0.79 and specificity of 0.83. LOOCV yielded similar results with an AUC of 0.85, sensitivity of 0.79, and specificity of 0.82. For the comparison between CDR 0.5 and CDR 1, k-fold cross-validation produced an AUC of 0.80, with sensitivity of 0.74, and specificity of 0.75. LOOCV obtained an identical AUC of 0.80, sensitivity of 0.63, and specificity of 0.75. Importantly, in the comparison between CDR 0 and CDR 2+, k-fold cross-validation and LOOCV reached AUCs of 0.89 and 0.88 respectively, with sensitivity higher than 0.69 and specificity above 0.86 in both cases. View this table: View inline View popup Download powerpoint Table 4 | Cross-validation performance of SPACE-based models. Classification accuracy of SPACE models across clinical groups using K-fold and LOOCV, reporting AUC, sensitivity, and specificity. Neuropsychological tests To compare the diagnostic accuracy of SPACE relative to standard neuropsychological measures, ROC analyses were performed, and differences in AUCs were assessed using Bonferroni-corrected DeLong test (see Figure 3 and Supplementary Table 4). For CDR 0 vs 0.5, no neuropsychological test differed significantly from SPACE, and all showed modest accuracy (AUC ≈ 0.60–0.73). When comparing CDR 0 vs 1, MoCA outperformed SPACE ( p = 0.021), whereas the QDRS ( p = 0.010), Maze ( p = 0.010), and Dual Task ( p = 0.007) performed significantly worse. In differentiating CDR 0.5 vs 1, SPACE again achieved high discrimination accuracy (AUC = 0.91) and was significantly better than the QDRS (p = 0.003), TMT-A ( p = 0.025), Maze ( p = 0.008), and Dual Task ( p = 0.004). No significant differences were observed for CDR 0 vs 2+, 0.5 vs 2+, or 1 vs 2+, where all tasks, including SPACE, performed strongly (AUCs > 0.85). Overall, SPACE showed the same or better discrimination than most conventional tests, particularly for transitions from very mild to mild dementia, with AUCs above 0.90 in all but the earliest (CDR 0 vs 0.5) and most advanced (CDR 1 vs 2+) stage comparisons. Download figure Open in new tab Figure 3 | Forest plots showing the classification performance (AUC ± 95% CI) of the SPACE tasks and neuropsychological tests across pairwise CDR group comparisons. Each panel represents a distinct comparison. Circles denote models including demographic variables (age, gender, education), while squares denote models excluding demographic variables (test performance only). Orange circles represent SPACE, and dark blue markers represent traditional neuropsychological tests. Horizontal lines indicate 95% confidence intervals for the models, including demographics. While the primary goal of this paper was to assess whether SPACE can discriminate between various CDR scores, we also compared it relative to consensus diagnosis. For each comparison, we calculated AUC values along with sensitivity and specificity at the optimal cut-off. The p-values reported in the table indicate the statistical tests comparing the AUC of each assessment with that of SPACE, which served as the reference model. For these analyses, we merged the No Cognitive Impairment (NCI) and Subjective Cognitive Decline/Impairment (SCD/SCI) groups into a single NCI category, as SCD/SCI reflects concerns about cognitive changes but does not constitute a formal clinical diagnosis of cognitive impairment. In distinguishing between NCI from Dementia, SPACE demonstrated excellent discrimination accuracy (AUC = 0.94). In this comparison, MoCA performed significantly better than SPACE (AUC = 1.00, p = 0.006), whereas QDRS performed significantly worse (AUC = 0.82, p = 0.037). All other tests showed no significant difference from SPACE. For NCI versus MCI, MoCA again outperformed SPACE (AUC 0.86 vs AUC 0.72, p = 0.002), while the remaining assessments were statistically indistinguishable from it. Finally, for MCI versus Dementia, none of the cognitive measures differed significantly from SPACE (see Supplementary Tables 5 and 6). Short SPACE On average, participants took 281 seconds to complete the SPACE training, 465 seconds for path integration, and 386 seconds for perspective taking tasks. A repeated measures ANOVA revealed significant differences in time to completion for these tasks ( F (2, 498) = 75.28, p < 0.001, η² = 0.232). Participants completed the training phase significantly faster than both the path integration and perspective taking tasks. Additionally, participants were faster at completing the perspective taking task than the path integration task. All Bonferroni post-hoc comparisons were significant ( p < 0.001). To assess whether a more time-efficient and user-friendly version of SPACE could also effectively discriminate between CDR stages, we repeated the classification analyses using only the training phase and perspective taking task ( Table 5 ). In this short version of SPACE (sSPACE), which required only an average of 667 seconds, the CDR 0 vs. CDR 0.5 contrast yielded an AUC of 0.61 (sensitivity = 0.49, specificity = 0.72). Again, the clinically relevant CDR 0 vs. CDR 1 contrast resulted in an AUC of 0.91, (0.91, 0.81), while the CDR 0.5 vs. CDR 1 comparison yielded an AUC of 0.82 (0.84, 0.78). Since the models with the moderate group did not include the path integration task, contrasts for sSPACE were limited to the CDR 0, CDR 0.5, and CDR 1 groups ( Figure 4 ). Download figure Open in new tab Figure 4 | ROC curves for the short version of SPACE (sSPACE). ROC curves showing the classification performance of sSPACE (comprising only the training and perspective taking tasks) for three contrasts: CDR 0 vs. CDR 0.5, CDR 0 vs. CDR 1, and CDR 0.5 vs. CDR 1. View this table: View inline View popup Download powerpoint Table 5 | Classification performance of the short SPACE version (sSPACE). Predictive performance of sSPACE models using K-fold and LOOCV procedures across clinical group comparisons. Given the known role of path integration in discriminating impaired from not-impaired individuals, we compared AUCs for each CDR contrast in sSPACE with and without the path integration task. Here again, we excluded contrasts involving CDR 2+ because participants in these groups frequently aborted the path integration task due to its length and complexity, opting instead to proceed directly to the perspective taking task. The addition of the path integration task did not significantly improve discrimination accuracy for CDR 0 vs CDR 0.5 ( p = 0.777, p bonferroni = 1.000 ) and CDR 0 vs CDR 1 ( p = 0.107, p bonferroni = 0.322 ). We found that including the path integration task significantly improved the AUC for CDR 0.5 and CDR 1 ( p = 0.019), increasing it from 0.83 to 0.91. However, this comparison did not survive Bonferroni correction ( p Bonferroni = 0.056). Discussion We developed a tablet-based spatial navigation assessment (SPACE) and tested it in a clinical and community cohort of 300 individuals diagnosed with varying levels of cognitive impairment using the CDR scale. SPACE improved classification performance beyond demographic models across most group comparisons, including CDR 0 vs. CDR 1 and CDR 2+, as well as CDR 0.5 vs. CDR 1 and CDR 2+. Classification accuracy was consistently high across these comparisons, with AUCs ranging from 0.91 to 0.95, sensitivities from 0.88 to 1.00, and specificities from 0.73 to 0.88. SPACE maintained high accuracy even when distinguishing between adjacent early stages. We also evaluated a shorter version of SPACE (sSPACE), designed for rapid clinical screening and unsupervised deployment at home and in community settings. In under 11 minutes, sSPACE retained diagnostic accuracy comparable to the full version. Cross-validation analyses confirmed the reliability and robustness of these models, further highlighting their suitability for scalable cognitive screening. Altogether, these findings underscore the importance of incorporating spatial navigation tests into cognitive assessments and their sensitivity for detecting subtle differences in cognitive status. While demographic models alone achieved modest classification performance, the inclusion of SPACE consistently improved model accuracy, sensitivity, and specificity across all diagnostic contrasts, although gains were not significant when comparing CDR 0 vs. CDR 0.5 and CDR 1 vs. CDR 2+. As expected, SPACE increased the AUC for the model distinguishing CDR 0 vs. CDR 2+ (AUC increase from 0.79 to 0.95) and when distinguishing CDR 0.5 vs. CDR 2+ (AUC increase from 0.80 to 0.91). Across both of these contrasts, sensitivity and specificity remained high, ranging from 0.88 to 0.94 and 0.85 to 0.88, respectively. Notably, there was also a substantial improvement when differentiating CDR 0 vs. CDR 1 (AUC increase from 0.76 to 0.94) and CDR 0.5 vs. CDR 1 (AUC increase from 0.70 to 0.91). Here, we also observed substantial gains in both sensitivity and specificity. For the contrast between CDR 0 vs. CDR 1, sensitivity increased from 0.74 to 1.00 and specificity from 0.71 to 0.85. Similarly, for the contrast between CDR 0.5 vs. CDR. 1, sensitivity rose from 0.74 to 0.95 and specificity from 0.59 to 0.73. SPACE classification performance for CDR 0 vs. CDR 0.5 and CDR 1 vs. CDR 2+ was modest. This result is likely due to a small overlap between the clinical classification of these disease stages, rather than a limitation of the navigation tests themselves 49 . The CDR classification between CDR 0.0 vs. 0.5 is based on the detection of subtle cognitive deficits, primarily relying on memory performance during the interview. In fact, previous work has sometimes sidestepped this challenge by simply combining the classifications of questionable dementia and mild dementia 50 , 51 or by completely omitting one of the categories 52 . However, even when addressing marginal comparisons, SPACE consistently improved model performance compared to sociodemographic variables alone. These findings were validated using stratified 10-fold cross-validation and LOOCV, with both methods yielding similarly high AUCs, sensitivity, and specificity across most contrasts. These results are also consistent with those from the consensus diagnosis. Apart from the MoCA, which performed exceedingly well in this cohort, SPACE demonstrated clear discriminative ability comparable to that of other neuropsychological tests. This convergence indicates that SPACE captures clinically meaningful differences in cognitive status across independent diagnostic methods. Previous research using an unsupervised mobile app focusing on episodic memory demonstrated promising cross-validated performance in distinguishing mild cognitive impairment from unimpaired individuals in memory clinic settings, with an AUC of 0.83, increasing to 0.87 when based on multiple sessions 44 . Other researchers have shown that memory (AUC = 0.92) and orientation tasks (AUC = 0.91) can discriminate between AD and frontotemporal dementia 32 , and that cognitive mapping tasks, rather than episodic memory, can predict disease progression (AUC = 0.8) 31 . Building on this evidence, we show that a tablet-based spatial navigation test battery is able to achieve comparable or superior classification accuracies (AUCs = 0.94–0.95), confirmed through cross-validation, both for detecting CDR 1 and CDR 2+ impairments. Importantly, deficits in spatial navigation appear to be an excellent marker for discretising subtle clinical stages such as those from questionable (CDR 0.5) to mild (CDR 1) impairments. To address the challenges associated with typical, costly clinical assessments, it is critical that novel tools deliver efficient, scalable, and user-friendly digital solutions that can be deployed in community-based or other unsupervised screening contexts. By focusing on two of the most discriminatory tasks (i.e. training time and perspective taking), sSPACE considerably reduced the overall administration time (to < 11 min) of the full SPACE battery while maintaining strong diagnostic accuracy. Notably, our results revealed that sSPACE achieved AUCs of up to 0.91 in distinguishing participants with CDR 0 from those with CDR 1, and 0.82 for differentiating CDR 0.5 from CDR 1. These values compare favourably to recent digital drawing tests, such as PENSIEVE-AI, which reported an AUC of 0.93 for distinguishing cognitively normal individuals from those with MCI/dementia in a large community sample 53 . Critically, the PENSIVE-AI study grouped MCI and dementia cases together and did not report classification performance for adjacent distinctions, such as between cognitively normal participants and those with MCI only. Additionally, PENSIEVE-AI still requires some degree of supervision for more severely impaired cases and relies on a stylus-enabled tablet, which may limit its scalability in non-clinical settings. In contrast, sSPACE combines strong diagnostic accuracy with the potential for fully unsupervised deployment 54 , offering a practical and scalable solution for early cognitive screening. sSPACE can serve as an initial screening tool to flag individuals at risk, who may then undergo more detailed clinical assessment. However, because path integration showed greater sensitivity for discriminating between CDR 0.5 and CDR 1, administering the full SPACE battery remains advisable whenever feasible. The tasks in SPACE were designed to engage a distributed network of brain regions along a continuum from egocentric to allocentric processing. Training engages visuospatial processing and motor control from an egocentric perspective supported by the posteromedial parietal (e.g., precuneus, posterior cingulate) and occipital regions 55 , 56 . Path integration is a primarily egocentric task that relies on the updating of position and orientation based on vection (illusory sensation of self-motion), which relies on the medial entorhinal cortex. This region interacts with the hippocampus, retrosplenial and parahippocampal cortices, medial prefrontal regions, and cerebellum, in order to integrate optic-flow, vestibular, and “proprioceptive” self-motion cues to maintain spatial orientation 36 , 57 , 58 . However, path integration can also provide the self-motion framework from which allocentric representations, such as the relationships between landmarks, can emerge 59 – 61 . The pointing task requires participants to retrieve and transform this self-referenced information to estimate the direction of unseen landmarks, engaging both posterior parietal regions for egocentric updating (e.g., the waterfall is to my right) and retrosplenial regions for coordinate transformations (e.g., the waterfall is northeast of the rocket) 32 , 62 . The mapping task further demands integrating multiple viewpoints into a coherent representation of landmark locations, a process that relies on hippocampal and parahippocampal structures associated with allocentric spatial memory 63 , 64 . However, the mapping task can technically also be solved procedurally by reconstructing the individual paths to previously visited landmarks supported by the basal ganglia 65 , 66 . Finally, the perspective taking task involves mentally adopting a viewpoint other than one’s own and judging where landmarks are from that imagined perspective. This task draws on allocentric representations to reconstruct spatial relations and egocentric simulation to visualise the scene from the imagined perspective. This process reflects a dynamic integration of both reference frames, engaging hippocampal and retrosplenial regions that enable flexible spatial transformations 62 , 67 . Together, these tasks illustrate a somewhat hierarchical but not absolute shift from self-based updating to world-based representation that underpins spatial navigation 59 . Of all the spatial navigation tasks performed in SPACE, the training phase, path integration, and perspective taking were best able to discriminate levels of cognitive impairment. These navigation impairments likely reflect early pathology in regions critical for spatial orientation and memory 16 , 21 , 31 , deficits that are known to be more prominent in AD and MCI compared to normal aging 26 , 68 . A notable finding was the progressively longer times required to complete the SPACE training phase as the severity of cognitive impairment increased. This is likely due to deficits in both basic visuospatial integration 39 , 69 , 70 and fine motor skills for tracking objects during rotation and movement manoeuvres 56 . Similarly, a study using a smartphone-based app called Altoida showed that a tapping task added efficacy in distinguishing between MCI and early AD 45 . More recently, research has shown that reaching movements under visual interference can distinguish MCI from healthy aging 71 . Importantly, a novel cognitive-motor framework further implicates disrupted integration of motor and spatial systems as a driver of navigation errors in early-stage AD 70 . Thus, prolonged training time in cognitively impaired participants likely reflects not only spatial learning deficits but also deterioration of visuospatial and motor skills needed to interact with the control interface. We found that deficits in the path integration task were sensitive indicators of early signs of cognitive impairment. Path integration has been widely studied and has consistently demonstrated a strong ability to discriminate across stages of cognitive impairment 33 – 36 . In our cohort, individuals with mild dementia exhibited significantly greater path integration errors than those with no or questionable dementia, aligning with prior evidence 27 , 28 . Path integration deficits are associated with reduced grid-like representations 33 , 34 in the entorhinal cortex, as well as atrophy in both the entorhinal cortex 34 and hippocampus 27 , 28 . These deficits have been shown to emerge in individuals at genetic risk for AD, appearing well before symptom onset, even when performance on standard cognitive tests is accurate 33 – 35 . While these findings reinforce the interpretation of path integration tasks as sensitive markers of early AD, the inherent complexity of these tasks may limit their utility for widespread screening, particularly for older adults or unsupervised populations. The perspective taking task in SPACE emerged as a valuable marker, with individuals in the mild and moderate dementia groups exhibiting significantly greater errors compared to those with no or questionable dementia. This task assesses the ability to mentally represent and manipulate spatial relationships from different viewpoints, tapping into both visuospatial ability and executive function 64 , 72 . Indeed, research using immersive VR has found that early AD disrupts the ability to convert allocentric into egocentric information, highlighting compromised reference-frame transformations 73 . Despite its simplicity and suitability for unsupervised deployment, perspective taking remains understudied in both research and clinical screening. To our knowledge, only one study has specifically focused on perspective taking with amnestic MCI and AD participants, showing significant deficits in tasks requiring first-person viewpoint transformations 39 . In SPACE, we digitalised the Spatial Orientation Test 72 , into a scalable, sensitive perspective taking assessment. However, its suitability for routine cognitive screening across diverse clinical populations requires further validation in broader clinical settings. While the absence of significant group differences in the pointing and mapping tasks might initially appear inconsistent, these findings can be understood by considering their dependence on the path integration task. Both the pointing and mapping tasks rely on participants successfully encoding the position of the various landmarks when completing the various path integration trials. Here, participants who fail to accurately encode the locations of the landmarks are likely to perform worse on the path integration task and the following tasks (i.e. pointing and mapping) that depend on this knowledge. In contrast, performance on the perspective taking task does not depend on prior landmark encoding, as participants are provided with a map of the landmark locations. To clarify the relationship between these tasks, we conducted additional regression analyses examining whether clinical status and path integration performance jointly explained variation in pointing and mapping. These models showed that individuals who made larger errors during path integration also tended to perform worse on the pointing and mapping tasks, suggesting that difficulties in encoding landmark positions carried over into these later measures (Supplementary Information F). Comparing SPACE to standard neuropsychological tests allowed for direct benchmarking of its diagnostic validity. ROC analyses and DeLong tests confirmed that SPACE performed on par with, or better than, most conventional neuropsychological tests across clinical contrasts. Critically, for contrasts distinguishing CDR 0 vs 1 and 0.5 vs 1, SPACE showed comparable performance to the MoCA, TMT-B, and DCT, and outperformed the QDRS, TMT-A, Maze Test, and Dual Task. This pattern suggests that SPACE engages executive, attentional, working memory and processing speed components similar to those tapped by these tasks. The strong performance of the MoCA across all CDR stages reinforces its recognised sensitivity for detecting dementia. This strong performance was expected given the MoCA’s multi-domain coverage and the fact that part of the reconstructed CDR scores were derived from clinical notes and neuropsychological tests that include the MoCA 74 . In this context, SPACE’s ability to achieve AUCs close to those of the MoCA is particularly noteworthy. Our findings build on prior work that has demonstrated that spatial navigation ability, assessed through digital tools, can be a sensitive marker of cognitive impairment 32 , 41 , 44 , 45 , 53 , 75 – 77 . Unlike traditional assessments, digital tools can enhance engagement and adherence 11 , 13 , 78 , 79 , making them especially suitable for testing older adults in supervised and unsupervised settings. The pioneering citizen-study with Sea Hero Quest revealed the potential for scalable population-level screening, although its clinical relevance remains to be established, but see 35 . Other virtual paradigms, such as the Virtual Supermarket Test 32 and the Four Mountains Test 77 , have successfully discriminated between cognitively impaired and healthy individuals 32 and shown potential to predict conversion from MCI to dementia 80 . However, these tools have been mainly confined to controlled laboratory or research contexts. SPACE aims to unify these two strands by translating the strengths of population-scale and laboratory-based approaches into a clinically validated tool, benchmarked against standard neuropsychological assessments. Here, the shorter administration times and similar diagnostic accuracy in sSPACE make it ideal for future large-scale implementation in clinics and unsupervised settings. Despite the clear benefits of our digital assessment, several limitations should be acknowledged. First, the cross-sectional design limits conclusions about the ability of SPACE to track cognitive decline over time. Here, longitudinal studies are required to assess predictive validity. Second, although performance was robust under supervised conditions, SPACE has not yet been fully evaluated in fully unsupervised settings, where variability in digital literacy and user familiarity with touchscreen interfaces may affect task performance. To mitigate these technological barriers, we previously conducted a detailed usability study that evaluated task controls, interface aids, and the configuration of landmarks in the virtual environment across age groups 48 . This work has allowed optimisation of the current design and reduced some of these confounds. More recently, we have showed that sSPACE can reliably predict MoCA scores when deployed unsupervised 54 . Third, although the proportion of MCI cases in this study is consistent with prevalence estimates in an ageing population 81 , 82 and large-scale studies on digital assessment 44 , 53 , the sample was nonetheless skewed toward individuals with CDR 0–0.5, with relatively few participants having a CDR > 1. Finally, our findings were drawn from a single study cohort. As such, replication in other samples is necessary to support generalisability as well as differences in demographic characteristics. We acknowledge that SPACE is not intended to replace comprehensive clinical or neuropsychological assessments. While SPACE does not offer a definitive diagnosis, this work has shown its clear potential for serving as a rapid, scalable tool for widespread screening of cognitive impairment. To maximise clinical relevance, SPACE and sSPACE metrics could also be integrated with complementary data sources such as neuropsychological assessments, fluid and imaging biomarkers, genetic profiles, and other clinical data 83 . A promising future direction is the continuous monitoring of spatial navigation ability, supported by mobile sensor technologies and integrated health data 47 , 84 , 85 . This approach could enable early detection of subtle cognitive decline, support longitudinal tracking of neurological changes, and contribute to digital phenotyping for MCI. Incorporating these tools into clinical workflows and public health systems may ultimately strengthen early detection strategies and support timely, targeted interventions. In summary, this study provides clear evidence that spatial navigation tasks are a sensitive marker for detecting early cognitive impairment. Through immersive and engaging tasks that tap a distributed network of brain regions associated with spatial and navigation abilities, the tablet-based SPACE tool provides a domain-specific and cost-effective alternative to traditional clinical assessments. Our results demonstrate strong diagnostic accuracy across cognitive levels, particularly across the questionable and mild levels of impairment. The shortened version, sSPACE, maintains performance while reducing assessment time, supporting its suitability for widespread and unsupervised deployment. This approach can alleviate pressure on healthcare systems and extend reach to underserved populations. While additional validation in unsupervised environments is clearly needed, SPACE offers promising avenues for expanding access to cognitive screening and advancing early intervention strategies. Methods Participants The study included 300 participants (M age = 74 years; 41% male). Patients were recruited from memory clinics at the National University Hospital (NUH) and St. Luke’s Hospital in Singapore, while control participants were recruited from these sites (90%) and the community (10%). Participants recruited from memory clinics were part of ongoing cohorts (HARMONISATION 86 and SINGER 87 ) at the Memory, Aging & Cognition Centre (MACC). All recruited participants underwent an initial telephone screening to assess their eligibility for the study. Participants were eligible if they were over 50 years old and could walk 10 meters without using a walking aid. Participants were excluded if they manifested severe visual impairment or hearing loss, a history of seizure, epilepsy, or acute cardiac events. Signed informed consent was obtained from all participants or their legal representative prior to undergoing the experimental procedure. Ethical approval for this study was provided by the NHG Domain Specific Review Board (DSRB) in Singapore (2021/01160). Sociodemographic and Health Questionnaire Participants completed a sociodemographic and health questionnaire that included questions on their age, gender, education level, handedness, tablet experience, previous navigation training, vision impairments, chronic health conditions, history of head trauma and falls, depression, anxiety and stress, smoking status, alcohol consumption, sleep, and physical activity. Furthermore, participants were asked to self-evaluate their spatial navigation abilities using the Santa Barbara Sense of Direction (SBSOD) scale 72 . Global Clinical Dementia Rating Diagnosis The global Clinical Dementia Rating (CDR) score was determined relative to six cognitive and functional domains: memory, orientation, judgment and problem solving, community affairs, home and hobbies, and personal care. The CDR was calculated using the CDR-assignment algorithm 88 . The CDR for each participant was either obtained from existing study records conducted within one year of the SPACE visit or reconstructed from comprehensive clinical notes and neuropsychological data when these records were unavailable. This method has been shown to achieve good to excellent agreement (intraclass correlation coefficient = 0.81–0.92) with face-to-face assessments 89 . CDR scores were used to classify impairment severity: 0 = no dementia, 0.5 = questionable dementia, 1 = mild dementia, 2 = moderate dementia, and 3 = severe dementia ( Table 1 ). Due to the limited number of participants with a CDR score of 3 (n = 3), the moderate and severe groups were merged for analysis. Clinical Consensus Dementia Diagnosis In addition to CDR classification, patients received a diagnosis by clinical consensus based on MACC’s procedures, which consider performance on the Vascular Dementia Battery (VDB), structural MRI findings, and relevant biomarkers. The VDB evaluates six cognitive domains (i.e. attention, language, verbal and visual memory, visuoconstruction, and visuomotor speed) and classifies domain impairment as “borderline” (failure in one domain-specific test but <50% of tasks impaired) or “impaired” (≥50% of tasks failed). This battery has been validated in Singaporean stroke and dementia cohorts 90 . Control participants had Mini-Mental State Examination (MMSE) scores of ≥ 23 (secondary or tertiary education) or ≥ 21 (primary or no education), and no cognitive domain impairment on the VDB. The diagnosis classified participants into groups reflecting cognitive severity (NCI = No Cognitive Impairment, SCD/SCI = Subjective Cognitive Decline/Impairment, MCI = Mild Cognitive Impairment, and Dementia), following the recommendation of the National Institute on Aging and the Alzheimer’s Association (NIA-AA) for the diagnosis and evaluation of Alzheimer’s Disease 91 . Etiological classification followed the recommendations of the NIA–AA and was supported by plasma and neuroimaging biomarkers. Plasma p-tau217 levels were quantified via SIMOA immunoassay and used to stratify participants into low (≤ 0.388 pg/mL), intermediate (0.387–0.470 pg/mL), or high (≥ 0.471 pg/mL) risk for PET positivity for brain amyloid pathology, according to thresholds validated in Singaporean cohorts 92 , 93 . A subset of participants (n = 209) underwent multimodal MRI, with cerebrovascular disease (CeVD) defined as cortical infarcts and/or ≥ 2 lacunes and/or confluent white matter lesions in ≥ 2 regions (ARWMC ≥ 8) 94 . Accordingly, the cohort included individuals with AD Non-Vascular, AD Vascular, Non-AD Non-Vascular, and Non-AD Vascular pathologies – capturing the pathological diversity encountered in clinical settings. For participants without research scans, medical history (including CT scans when available) was used to determine CeVD status. Spatial Performance Assessment for Cognitive Evaluation (SPACE) SPACE is a serious game deployed on the iPad to assess spatial navigation abilities that may be indicative of cognitive impairment. In SPACE, participants assume the role of an astronaut tasked with exploring and learning the positions of different landmarks on a planet while completing a series of spatial navigation tasks ( Figure 5 ). A comprehensive description of SPACE is provided in a previous publication 48 . Download figure Open in new tab Figure 5 | Screenshots from SPACE tasks. Start (top images) and end (bottom images) screens from three core tasks within the SPACE assessment. (a) Training phase, rotation: Participants rotate their viewpoint by swiping left or right on the screen to find the robot. (b) Path integration: Starting from the rocket and following the robot to two landmarks, participants must navigate back to the starting location without assistance. (c) Perspective taking: Participants view a top-down map of the environment and are asked to identify the direction of a landmark by imagining themselves standing at a specified location while facing a given orientation (e.g., "imagine you are standing at the tree, facing the rocket point to the waterfall). SPACE begins with a training phase in which participants complete a series of manoeuvres designed to familiarise them with the controls while also assessing basic visuospatial skills. During this phase, participants learn how to rotate, move, and follow a robot from their starting point (rocket) to various destinations and back ( Figure 5a ). Following the training phase, participants complete five spatial tasks that target different aspects of navigation. The path integration task evaluates how well participants can track changes in position and orientation as they navigate from the rocket to three landmarks. In each trial, participants follow the robot along two legs of a triangle and are then asked to return to the rocket via the third leg. At each landmark, the robot scans an item that participants will later be asked to recall ( Figure 5b ). The pointing task measures how accurately participants can indicate the direction of other landmarks from a first-person perspective. Over a series of trials, participants are placed in front of a landmark and asked to indicate the direction of other landmarks in the environment. The mapping task assesses their ability to form a top-down mental map of the environment by reconstructing the spatial layout of landmarks. The associative memory task tests their recall of the specific items scanned at each landmark during the path integration task. Finally, the perspective taking task measures participants’ ability to calculate the spatial relation between the landmarks from various perspectives. Here, the player is asked to imagine standing at a landmark facing another landmark by looking at the map of the planet. Participants are then required to indicate the correct bearing towards a third landmark ( Figure 5c ). Performance in SPACE is quantified using the following measures. Training time is the duration (in seconds) required for participants to complete the entire training task. Path integration distance error refers to the mean distance between the player’s final position and the rocket’s original position across all path integration trials. Pointing error is the mean absolute angular deviation (in degrees) from a starting position to the target location. Mapping accuracy is determined using bidimensional regression 95 , 96 , a statistical method for assessing the degree of association (R 2 ) between the reference map and the map generated by the participants. The associative memory score is the percentage of correct pairings between scanned items and corresponding landmarks. Perspective taking error is the mean absolute angular deviation (in degrees) between the participant’s directional estimate and the target landmark across all trials. Study procedure Upon arriving at the study site, participants received an overview of the study procedures. Once written informed consent was obtained, participants completed a series of questionnaires to collect sociodemographic data, health information, and self-reported navigation ability. Participants then completed a battery of neuropsychological tests and SPACE. The neuropsychological tests included the MoCA, QDRS, Trail Making Test A and B (TMT-A/B), Maze Task, Animal Fluency, Digit Cancellation Test (DCT), and a Dual Task measure of divided attention. After completing SPACE, blood (n = 300) and saliva samples (n = 96) were collected for biomarker analyses. Participants also completed a short gait assessment using single and dual tasks. The entire testing procedure lasted approximately 3 hours. This study is part of a wider data collection effort to validate SPACE in community and clinical samples. Statistical analyses All statistical analyses and plots were performed using Jamovi (v.2.6.26.0) and RStudio (v.2024.12.0+467), with p-values < 0.05 considered statistically significant, and Bonferroni corrections were applied whenever appropriate. Descriptive statistics were computed for key demographic, clinical, neuropsychological, and task-performance variables across CDR levels. Given violations of the assumptions of normality of residuals and homogeneity of variance, we examined group differences in performance on the spatial navigation tasks of SPACE using robust one-way ANOVAs (modified one-step estimator; 5,000 bootstrap samples). To account for age and gender differences in navigation performance, we additionally conducted ANCOVAs for each SPACE task with these variables included as covariates. As the ANCOVAs supported the group differences reported with the robust ANOVAs, we retained the robust ANOVA as our primary approach. Robust ANOVAs provide unbiased estimates under violations of normality, homogeneity of variance, and unequal group sizes 97 , which was the case for most of our models. We report the results of the ANCOVA in the Supplementary Information G for transparency and comparison. Logistic regressions were then used to model the relationship between predictors and all combinations of binary outcomes from the CDR scale. Following the methodology from previous researchers 44 , we compared three model types: a demographic model with age, education and gender ( Dem ), a model including only the significant tasks in SPACE ( SPACE ), and a combined model including both demographic variables and the SPACE tasks ( Total ). The purpose of this comparison was to establish a baseline model using known demographic risk factors for dementia, against which the added discriminative value of SPACE performance could be evaluated. Model performance was compared using DeLong’s test to evaluate the incremental contribution of SPACE beyond demographic predictors. The optimal classification cutoff for each model was determined using the Youden index and assessed using the AUC, sensitivity, specificity, and accuracy. Model performance was further evaluated using deviance, AIC, and the Brier score. Given that only two participants with a CDR of 2+ completed the path integration, pointing, and mapping tasks, this group were excluded from the analysis for these tasks. To account for the results of the ANCOVA, we additionally evaluated a SPACE model that included the pointing task alongside demographic variables (Supplementary Table 9). We cross-validated our models using k-fold cross-validation and Leave-One-Out Cross-Validation (LOOCV). For both methods, we provide estimates of model performance, including AUC, sensitivity, and specificity, along with their corresponding confidence intervals. The k-fold cross-validation involved partitioning the dataset into k subsets (k = 10), training the model on k-1 subsets, and testing it on the remaining subset. LOOCV involves training the model on all but one data point and testing it on the remaining point, iterating this process for each data point in the dataset. To compare the performance of SPACE relative to traditional neuropsychological tests, we conducted ROC analyses for each CDR stage contrast. Differences in AUC values between SPACE and other neuropsychological tests were examined using the DeLong test with Bonferroni correction. The same analysis was conducted to assess whether SPACE can discriminate between categories from the consensus diagnosis (NCI vs. Dementia, NCI vs. MCI, and MCI vs. Dementia). For these analyses, we merged the No Cognitive Impairment (NCI) and Subjective Cognitive Decline/Impairment (SCD/SCI) groups into a single NCI category, as SCD/SCI reflects concerns about cognitive changes but does not constitute a formal clinical diagnosis of cognitive impairment. Finally, we evaluated whether a short version of SPACE, excluding the path integration task, could discriminate between different CDR levels. Here, a Repeated Measures ANOVA was employed to assess differences in the time taken to complete the three tasks (training, path integration, and perspective taking). Logistic regressions, ROC curve analysis, and cross-validation (stratified k-fold and LOOCV) were also performed on the short version of SPACE. Since the primary focus of this paper is to assess the performance of SPACE relative to the CDR scores, various other measures were excluded from the main analyses. A complete list of all measures collected and additional analyses is reported in the Supplementary Information. Data Availability Data and biomaterials can be shared with partners based on individual data and biomaterial transfer agreements. Author Contributions G.C., K.M., and V.R.S. conceived the study. G.C., J.G., and V.R.S. were responsible for software conceptualisation and development. E.C., J.R.C., and M.K.P.L. were responsible for participant recruitment and data collection. E.C. and C.P.C. provided clinical oversight and contributed to data collection coordination. G.C. and V.R.S. curated and analysed the data. G.C., K.M., and V.R.S. contributed to data interpretation. K.M. procured ethics approval. G.C. wrote the first draft of the manuscript and prepared the data visualisations. G.C., W.R.T., and V.R.S. contributed to critical revisions and substantive refinement of the manuscript. V.R.S. supervised the project. All authors reviewed and approved the final manuscript. 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. Acknowledgments This research is supported by the National Research Foundation Singapore (NRF) under its Campus for Research Excellence and Technological Enterprise (CREATE) programme. Footnotes The revised manuscript introduces substantial new material that strengthens the methodological transparency, analytical rigour, and interpretative depth of the study. Several updates clarify the study design, expand performance analyses of SPACE, and provide a clearer comparison with established neuropsychological tools. 1. New Comparative Analyses With Neuropsychological Tests A major addition is the set of ROC analyses comparing SPACE against standard neuropsychological measures (MoCA, TMT-A/B, Maze, DCT, fluency, dual-task performance). These analyses identify where SPACE performs similarly or exceeds existing tests across CDR contrasts, and include Bonferroni-corrected statistical comparisons. 2. Expanded Theoretical Rationale and Task-Level Interpretation The discussion now includes a more detailed neurocognitive account of how each SPACE task (training, path integration, pointing, mapping, perspective taking) maps onto brain systems involved in egocentric-allocentric processing. Additional literature is incorporated to frame the interpretation of deficits in different tasks and their relevance for early cognitive impairment. 3. Enhanced Clinical and Demographic Characterisation New descriptive statistics and tables provide comparative demographic, neuropsychological, biomarker, and clinical diagnostic information across CDR levels, including APOE genotype distributions and CeVD status. 4. Clarified Methods and Procedures Additional procedural detail has been added regarding recruitment, test-ordering, biomarker collection, gait assessments, and ethical approvals. Supplementary materials and exclusions (e.g., missing path integration data in CDR 2+) are now explicitly explained. References 1. ↵ GBD 2019 Dementia Forecasting Collaborators . Estimation of the global prevalence of dementia in 2019 and forecasted prevalence in 2050: an analysis for the Global Burden of Disease Study 2019 . Lancet Public Health 7 , e105 – e125 ( 2022 ). OpenUrl CrossRef PubMed 2. ↵ Livingston , G. et al. Dementia prevention, intervention, and care: 2024 report of the Lancet standing Commission . Lancet ( 2024 ) doi: 10.1016/s0140-6736(24)01296-0 . OpenUrl CrossRef PubMed 3. ↵ Nandi , A. et al. Global and regional projections of the economic burden of Alzheimer’s disease and related dementias from 2019 to 2050: A value of statistical life approach . EClinicalMedicine 51 , 101580 ( 2022 ). OpenUrl PubMed 4. ↵ Petersen , R. C . Normal aging, mild cognitive impairment and Alzheimer’s disease . The neurologist 4 , 326 – 344 ( 1995 ). OpenUrl 5. ↵ Alzheimer’s Association . 2024 Alzheimer’s disease facts and figures . Alzheimers. Dement . 20 , 3708 – 3821 ( 2024 ). OpenUrl CrossRef PubMed 6. Ngandu , T. et al. A 2 year multidomain intervention of diet, exercise, cognitive training, and vascular risk monitoring versus control to prevent cognitive decline in at-risk elderly people (FINGER): a randomised controlled trial . Lancet 385 , 2255 – 2263 ( 2015 ). OpenUrl CrossRef PubMed 7. Gustavsson , A. et al. Global estimates on the number of persons across the Alzheimer’s disease continuum . Alzheimers. Dement . 19 , 658 – 670 ( 2023 ). OpenUrl CrossRef PubMed 8. ↵ Niotis , K. , Saperia , C. , Saif , N. , Carlton , C. & Isaacson , R. S . Alzheimer’s disease risk reduction in clinical practice: a priority in the emerging field of preventive neurology . Nat. Ment. Health 2 , 25 – 40 ( 2024 ). OpenUrl 9. ↵ van der Flier , W. M. , de Vugt , M. E. , Smets , E. M. A. , Blom , M. & Teunissen , C. E . Towards a future where Alzheimer’s disease pathology is stopped before the onset of dementia. Nat . Aging 3 , 494 – 505 ( 2023 ). OpenUrl PubMed 10. Aisen , P. S. , Jimenez-Maggiora , G. A. , Rafii , M. S. , Walter , S. & Raman , R . Early-stage Alzheimer disease: getting trial-ready . Nat. Rev. Neurol . 18 , 389 – 399 ( 2022 ). OpenUrl CrossRef PubMed 11. ↵ Hampel , H. et al. Designing the next-generation clinical care pathway for Alzheimer’s disease. Nat . Aging 2 , 692 – 703 ( 2022 ). OpenUrl PubMed 12. Ilardi , C. R. et al. On the clinimetrics of the Montreal Cognitive Assessment: Cutoff analysis in patients with mild cognitive impairment due to Alzheimer’s disease . J. Alzheimers. Dis . 101 , 293 – 308 ( 2024 ). OpenUrl PubMed 13. ↵ Polk , S. E. et al. A scoping review of remote and unsupervised digital cognitive assessments in preclinical Alzheimer’s disease . NPJ Digit. Med . 8 , 266 ( 2025 ). OpenUrl PubMed 14. Watermeyer , T. & Calia , C . Neuropsychological assessment in preclinical and prodromal Alzheimer disease: a global perspective . J. Glob. Health 9 , 010317 ( 2019 ). OpenUrl PubMed 15. Rentz , D. M. et al. Promising developments in neuropsychological approaches for the detection of preclinical Alzheimer’s disease: a selective review . Alzheimers. Res. Ther . 5 , 58 ( 2013 ). OpenUrl PubMed 16. ↵ Coughlan , G. , Laczó , J. , Hort , J. , Minihane , A.-M. & Hornberger , M . Spatial navigation deficits - overlooked cognitive marker for preclinical Alzheimer disease? Nat. Rev. Neurol . 14 , 496 – 506 ( 2018 ). OpenUrl CrossRef PubMed 17. ↵ Jack , C. R. , Jr . et al. NIA-AA Research Framework: Toward a biological definition of Alzheimer’s disease . Alzheimers. Dement . 14 , 535 – 562 ( 2018 ). OpenUrl CrossRef PubMed 18. Brier , M. R. et al. Tau and Aβ imaging, CSF measures, and cognition in Alzheimer’s disease . Sci. Transl. Med . 8 , 338r a66-338ra66 ( 2016 ). OpenUrl CrossRef 19. de Calignon , A. et al. Propagation of tau pathology in a model of early Alzheimer’s disease . Neuron 73 , 685 – 697 ( 2012 ). OpenUrl CrossRef PubMed Web of Science 20. Braak , H. & Braak , E . Neuropathological stageing of Alzheimer-related changes . Acta Neuropathol . 82 , 239 – 259 ( 1991 ). OpenUrl CrossRef PubMed Web of Science 21. ↵ Storandt , M. , Mintun , M. A. , Head , D. & Morris , J. C . Cognitive decline and brain volume loss as signatures of cerebral amyloid-beta peptide deposition identified with Pittsburgh compound B: cognitive decline associated with Abeta deposition . Arch. Neurol . 66 , 1476 – 1481 ( 2009 ). OpenUrl CrossRef PubMed Web of Science 22. ↵ O’Keefe , J. & Nadel , L . The hippocampus as a cognitive map . Behav. Brain Sci . 2 , 487 – 494 ( 1979 ). OpenUrl CrossRef Web of Science 23. ↵ Hafting , T. , Fyhn , M. , Molden , S. , Moser , M.-B. & Moser , E. I . Microstructure of a spatial map in the entorhinal cortex . Nature 436 , 801 – 806 ( 2005 ). OpenUrl CrossRef PubMed Web of Science 24. ↵ deIpolyi , A. R. , Rankin , K. P. , Mucke , L. , Miller , B. L. & Gorno-Tempini , M. L. Spatial cognition and the human navigation network in AD and MCI . Neurology 69 , 986 – 997 ( 2007 ). OpenUrl CrossRef PubMed 25. Tangen , G. G. , Nilsson , M. H. , Stomrud , E. , Palmqvist , S. & Hansson , O . Spatial navigation and its association with biomarkers and future dementia in memory clinic patients without dementia . Neurology 99 , e2081 – e2091 ( 2022 ). OpenUrl PubMed 26. ↵ Hort , J. et al. Spatial navigation deficit in amnestic mild cognitive impairment . Proc. Natl. Acad. Sci. U. S. A . 104 , 4042 – 4047 ( 2007 ). OpenUrl Abstract / FREE Full Text 27. ↵ Castegnaro , A. , Ji , Z. , Rudzka , K. , Chan , D. & Burgess , N . Overestimation in angular path integration precedes Alzheimer’s dementia . Curr. Biol . ( 2023 ) doi: 10.1016/j.cub.2023.09.047 . OpenUrl CrossRef 28. ↵ Howett , D. et al. Differentiation of mild cognitive impairment using an entorhinal cortex-based test of virtual reality navigation . Brain 142 , 1751 – 1766 ( 2019 ). OpenUrl CrossRef PubMed 29. Mapstone , M. , Steffenella , T. M. & Duffy , C. J . A visuospatial variant of mild cognitive impairment: getting lost between aging and AD . Neurology 60 , 802 – 808 ( 2003 ). OpenUrl CrossRef PubMed 30. Cushman , L. A. , Stein , K. & Duffy , C. J . Detecting navigational deficits in cognitive aging and Alzheimer disease using virtual reality . Neurology 71 , 888 – 895 ( 2008 ). OpenUrl CrossRef PubMed 31. ↵ Levine , T. F. et al. Spatial navigation ability predicts progression of dementia symptomatology . Alzheimers. Dement . 16 , 491 – 500 ( 2020 ). OpenUrl CrossRef PubMed 32. ↵ Tu , S. et al. Lost in spatial translation - A novel tool to objectively assess spatial disorientation in Alzheimer’s disease and frontotemporal dementia . Cortex 67 , 83 – 94 ( 2015 ). OpenUrl CrossRef PubMed 33. ↵ Newton , C. et al. Entorhinal-based path integration selectively predicts midlife risk of Alzheimer’s disease . Alzheimers Dement 20 , 2779 – 2793 ( 2024 ). OpenUrl CrossRef PubMed 34. ↵ Bierbrauer , A. et al. Unmasking selective path integration deficits in Alzheimer’s disease risk carriers . Sci. Adv . 6 , eaba1394 ( 2020 ). 35. ↵ Coughlan , G. et al. Toward personalized cognitive diagnostics of at-genetic-risk Alzheimer’s disease . Proc. Natl. Acad. Sci. U. S. A . 116 , 9285 – 9292 ( 2019 ). OpenUrl Abstract / FREE Full Text 36. ↵ Segen , V. , Ying , J. , Morgan , E. , Brandon , M. & Wolbers , T . Path integration in normal aging and Alzheimer’s disease . Trends Cogn. Sci . 26 , 142 – 158 ( 2022 ). OpenUrl CrossRef PubMed 37. ↵ Minta , K. *, et al. SPACE: A novel digital tool for assessing hippocampal structural integrity in older adults . Research Square ( 2024 ) doi: 10.21203/rs.3.rs-5484040/v1 . OpenUrl CrossRef 38. ↵ Mokrisova , I. et al. Real-space path integration is impaired in Alzheimer’s disease and mild cognitive impairment . Behav. Brain Res . 307 , 150 – 158 ( 2016 ). OpenUrl CrossRef PubMed 39. ↵ Marková , H. , Laczó , J. , Andel , R. , Hort , J. & Vlček , K . Perspective taking abilities in amnestic mild cognitive impairment and Alzheimer’s disease . Behav. Brain Res . 281 , 229 – 238 ( 2015 ). OpenUrl PubMed 40. ↵ Laczó , M. et al. Spatial navigation and visuospatial strategies in typical and atypical aging . Brain Sci . 11 , 1421 ( 2021 ). OpenUrl PubMed 41. ↵ Coutrot , A. et al. Global determinants of navigation ability . Curr. Biol . 28 , 2861 – 2866 .e4 ( 2018 ). OpenUrl CrossRef PubMed 42. ↵ Goodroe , S. et al. Predicting real-world navigation performance from a virtual navigation task in older adults . PLoS One 20 , e0317026 ( 2025 ). OpenUrl PubMed 43. ↵ Thompson , L. I. et al. Remote and in-clinic digital cognitive screening tools outperform the MoCA to distinguish cerebral amyloid status among cognitively healthy older adults . Alzheimers Dement. (Amst .) 15 , ( 2023 ). 44. ↵ Berron , D. et al. A remote digital memory composite to detect cognitive impairment in memory clinic samples in unsupervised settings using mobile devices . NPJ Digit Med 7 , 79 ( 2024 ). OpenUrl PubMed 45. ↵ Meier , I. B. et al. Using a Digital Neuro Signature to measure longitudinal individual-level change in Alzheimer’s disease: the Altoida large cohort study . npj Digital Medicine 4 , ( 2021 ). 46. ↵ Öhman , F. , Hassenstab , J. , Berron , D. , Schöll , M. & Papp , K. V . Current advances in digital cognitive assessment for preclinical Alzheimer’s disease . Alzheimers Dement. (Amst .) 13 , e12217 ( 2021 ). OpenUrl 47. ↵ Butler , P. M. et al. Smartwatch- and smartphone-based remote assessment of brain health and detection of mild cognitive impairment . Nat. Med . 31 , 829 – 839 ( 2025 ). OpenUrl PubMed 48. ↵ Colombo , G. et al. Detecting cognitive impairment through an age-friendly serious game: The development and usability of the Spatial Performance Assessment for Cognitive Evaluation (SPACE) . Comput. Human Behav . 160 , 108349 ( 2024 ). OpenUrl 49. ↵ Wada-Isoe , K. et al. Global clinical dementia rating score of 0.5 may not be an accurate criterion to identify individuals with mild cognitive impairment . J. Alzheimers Dis. Rep . 3 , 233 – 239 ( 2019 ). OpenUrl PubMed 50. ↵ Kleiman , M. J. , Barenholtz , E. , Galvin , J. E. & Alzheimer’s Disease Neuroimaging Initiative. Screening for early-stage Alzheimer’s disease using optimized feature sets and machine learning . J. Alzheimers. Dis . 81 , 355 – 366 ( 2021 ). OpenUrl PubMed 51. ↵ Alim-Marvasti , A. et al. Validation of a rapid remote digital test for impaired cognition using clinical dementia rating and mini-mental state examination: An observational research study . Front. Digit. Health 4 , 1029810 ( 2022 ). OpenUrl PubMed 52. ↵ An , D. , Shin , J. S. , Bae , N. , Seo , S. W. & Na , D. L . Validity of the tablet-based digital cognitive test (SCST) in identifying different degrees of cognitive impairment . J. Korean Med. Sci . 39 , e247 ( 2024 ). OpenUrl PubMed 53. ↵ Liew , T. M. et al. PENSIEVE-AI a brief cognitive test to detect cognitive impairment across diverse literacy . Nat. Commun . 16 , 2847 ( 2025 ). OpenUrl PubMed 54. ↵ Tian , N. , Colombo , G. & Schinazi , V. From play to detection: Mini-SPACE as a serious game for unsupervised Cognitive Impairment screening . arXiv [cs.HC] ( 2025 ) doi: 10.48550/ARXIV.2511.12068 . OpenUrl CrossRef 55. ↵ Cavanna , A. E. & Trimble , M. R . The precuneus: a review of its functional anatomy and behavioural correlates . Brain 129 , 564 – 583 ( 2006 ). OpenUrl CrossRef PubMed Web of Science 56. ↵ Ilardi , C. R. , Chieffi , S. , Iachini , T. & Iavarone , A . Neuropsychology of posteromedial parietal cortex and conversion factors from Mild Cognitive Impairment to Alzheimer’s disease: systematic search and state-of-the-art review . Aging Clin. Exp. Res . 34 , 289 – 307 ( 2022 ). OpenUrl PubMed 57. ↵ Schinazi , V. R. & Thrash , T. Cognitive neuroscience of spatial and geographic thinking . in Handbook of Behavioral and Cognitive Geography ( Edward Elgar Publishing , 2018 ). 58. ↵ Epstein , R. A. , Patai , E. Z. , Julian , J. B. & Spiers , H. J . The cognitive map in humans: spatial navigation and beyond . Nat. Neurosci . 20 , 1504 – 1513 ( 2017 ). OpenUrl CrossRef PubMed 59. ↵ Ekstrom , A. D. , Arnold , A. E. G. F. & Iaria , G . A critical review of the allocentric spatial representation and its neural underpinnings: toward a network-based perspective . Front. Hum. Neurosci . 8 , 803 ( 2014 ). OpenUrl CrossRef PubMed 60. Arnold , A. E. , Burles , F. , Bray , S. , Levy , R. M. & Iaria , G . Differential neural network configuration during human path integration . Front Hum Neurosci 8 , 263 ( 2014 ). OpenUrl CrossRef PubMed 61. ↵ Chrastil , E. R. , Sherrill , K. R. , Aselcioglu , I. , Hasselmo , M. E. & Stern , C. E . Individual Differences in Human Path Integration Abilities Correlate with Gray Matter Volume in Retrosplenial Cortex, Hippocampus, and Medial Prefrontal Cortex . eNeuro 4 , ( 2017 ). 62. ↵ Schinazi , V. R. , Nardi , D. , Newcombe , N. S. , Shipley , T. F. & Epstein , R. A . Hippocampal size predicts rapid learning of a cognitive map in humans . Hippocampus 23 , 515 – 528 ( 2013 ). OpenUrl CrossRef PubMed Web of Science 63. ↵ Wolbers , T. & Buchel , C . Dissociable retrosplenial and hippocampal contributions to successful formation of survey representations . J Neurosci 25 , 3333 – 3340 ( 2005 ). OpenUrl Abstract / FREE Full Text 64. ↵ Ruggiero , G. , Ruotolo , F. , Iavarone , A. & Iachini , T . Allocentric coordinate spatial representations are impaired in aMCI and Alzheimer’s disease patients . Behav. Brain Res . 393 , 112793 ( 2020 ). OpenUrl PubMed 65. ↵ Iaria , G. , Petrides , M. , Dagher , A. , Pike , B. & Bohbot , V. D . Cognitive strategies dependent on the hippocampus and caudate nucleus in human navigation: variability and change with practice . J Neurosci 23 , 5945 – 5952 ( 2003 ). OpenUrl Abstract / FREE Full Text 66. ↵ Weisberg , S. M. , Schinazi , V. R. , Newcombe , N. S. , Shipley , T. F. & Epstein , R. A . Variations in cognitive maps: Understanding individual differences in navigation . J. Exp. Psychol. Learn. Mem. Cogn . 40 , 669 – 682 ( 2014 ). OpenUrl CrossRef PubMed 67. ↵ Lambrey , S. , Doeller , C. , Berthoz , A. & Burgess , N . Imagining being somewhere else: neural basis of changing perspective in space . Cereb. Cortex 22 , 166 – 174 ( 2012 ). OpenUrl CrossRef PubMed 68. ↵ Lester , A. W. , Moffat , S. D. , Wiener , J. M. , Barnes , C. A. & Wolbers , T . The Aging Navigational System . Neuron 95 , 1019 – 1035 ( 2017 ). OpenUrl CrossRef PubMed 69. ↵ Mitolo , M. et al. Visuospatial memory and neuroimaging correlates in mild cognitive impairment . J. Alzheimers. Dis . 35 , 75 – 90 ( 2013 ). OpenUrl PubMed 70. ↵ Hill , P. F. & Ekstrom , A. D . A cognitive-motor framework for spatial navigation in aging and early-stage Alzheimer’s disease . Cortex 185 , 133 – 150 ( 2025 ). OpenUrl PubMed 71. ↵ Ilardi , C. R. et al. Deficits in reaching movements under visual interference as a novel diagnostic marker for mild cognitive impairment . Sci. Rep . 15 , 1901 ( 2025 ). OpenUrl PubMed 72. ↵ Kozhevnikov , M. & Hegarty , M . A dissociation between object manipulation spatial ability and spatial orientation ability . Mem. Cognit . 29 , 745 – 756 (7 2001 ). OpenUrl CrossRef PubMed Web of Science 73. ↵ Morganti , F. , Stefanini , S. & Riva , G . From allo- to egocentric spatial ability in early Alzheimer’s disease: a study with virtual reality spatial tasks . Cogn. Neurosci . 4 , 171 – 180 ( 2013 ). OpenUrl CrossRef PubMed 74. ↵ Haji , B. , Zhang , Q. , & Monfared , A. A. T . Bridging the gap: A conversion framework for CDR-SB and MoCA scores in Alzheimer’s disease and related dementia . The Journal of Prevention of Alzheimer’s Disease ( 2025 ) doi: 10.1016/j.tjpad.2025.100226 . OpenUrl CrossRef 75. ↵ van der Ham , I. J. M. , Claessen , M. H. G. , Evers , A. W. M. & van der Kuil , M. N. A . Large-scale assessment of human navigation ability across the lifespan . Sci. Rep . 10 , 3299 ( 2020 ). OpenUrl PubMed 76. Rekers , S. & Finke , C . Translating spatial navigation evaluation from experimental to clinical settings: The virtual environments navigation assessment (VIENNA) . Behav. Res. Methods 56 , 2033 – 2048 ( 2024 ). OpenUrl PubMed 77. ↵ Chan , D. et al. The 4 Mountains Test: A Short Test of Spatial Memory with High Sensitivity for the Diagnosis of Pre-dementia Alzheimer’s Disease . J. Vis. Exp . ( 2016 ) doi: 10.3791/54454 . OpenUrl CrossRef PubMed 78. ↵ Hamari , J. , Koivisto , J. & Sarsa , H. Does gamification work?--a literature review of empirical studies on gamification . in 47th Hawaii international conference on system sciences 3025 – 3034 ( Ieee , 2014 ). 79. ↵ Mcalearney , A. S. High Touch and High Tech (HT2) proposal: transforming patient engagement throughout the continuum of care by engaging patients with portal technology at the bedside . JMIR Res. Protoc 5 , ( 2016 ). 80. ↵ Wood , R. A. , Moodley , K. K. , Lever , C. , Minati , L. & Chan , D . Allocentric spatial memory testing predicts conversion from mild cognitive impairment to dementia: An initial proof-of-concept study . Front. Neurol . 7 , 215 ( 2016 ). OpenUrl CrossRef PubMed 81. ↵ Bai , W. et al. Worldwide prevalence of mild cognitive impairment among community dwellers aged 50 years and older: a meta-analysis and systematic review of epidemiology studies . Age Ageing 51 , ( 2022 ). 82. ↵ Salari , N. et al. The global prevalence of mild cognitive impairment in geriatric population with emphasis on influential factors: a systematic review and meta-analysis . BMC Geriatr . 25 , 313 ( 2025 ). OpenUrl PubMed 83. ↵ Berron , D. et al. Remote and unsupervised digital memory assessments can reliably detect cognitive impairment in Alzheimer’s disease . Alzheimers. Dement . 20 , 4775 – 4791 ( 2024 ). OpenUrl PubMed 84. ↵ Marquardt , J. et al. Identifying older adults at risk for dementia based on smartphone data obtained during a wayfinding task in the real world. PLOS Digit . Health 3 , e0000613 ( 2024 ). OpenUrl 85. ↵ Ghosh , A. , Puthusseryppady , V. , Chan , D. , Mascolo , C. & Hornberger , M . Machine learning detects altered spatial navigation features in outdoor behaviour of Alzheimer’s disease patients . Sci. Rep . 12 , 3160 ( 2022 ). OpenUrl PubMed 86. ↵ Lim , M. J. H. et al. HARMONISATION – A multimodal prospective study of vascular cognitive impairment in multi-ethnic Asians: Cohort profile, progress, current contributions, and future impact . Journal of Alzheimer’s Disease 0 , ( 2025 ). 87. ↵ Xu , X. et al. The SINgapore GERiatric intervention study to reduce cognitive decline and physical frailty (SINGER): Study design and protocol . J. Prev. Alzheimers Dis . 9 , 40 – 48 ( 2022 ). OpenUrl PubMed 88. ↵ Morris , J. C . The Clinical Dementia Rating (CDR): current version and scoring rules: Current version and scoring rules . Neurology 43 , 2412 – 2414 ( 1993 ). OpenUrl CrossRef PubMed 89. ↵ Dauphinot , V. et al. Reliability of the assessment of the clinical dementia rating scale from the analysis of medical records in comparison with the reference method . Alzheimers. Res. Ther . 16 , 198 ( 2024 ). OpenUrl PubMed 90. ↵ Tham , W. et al. Progression of cognitive impairment after stroke: one year results from a longitudinal study of Singaporean stroke patients . J. Neurol. Sci . 203-204 , 49 – 52 ( 2002 ). OpenUrl CrossRef 91. ↵ Jack , C. R. , Jr . et al. Revised criteria for diagnosis and staging of Alzheimer’s disease: Alzheimer’s Association Workgroup . Alzheimers. Dement . 20 , 5143 – 5169 ( 2024 ). OpenUrl CrossRef PubMed 92. ↵ Chong , J. R. et al. Blood-based high sensitivity measurements of beta-amyloid and phosphorylated tau as biomarkers of Alzheimer’s disease: a focused review on recent advances . J. Neurol. Neurosurg. Psychiatry 92 , 1231 – 1241 ( 2021 ). OpenUrl Abstract / FREE Full Text 93. ↵ Chong , J. R. et al. Clinical utility of plasma p-tau217 in identifying abnormal brain amyloid burden in an Asian cohort with high prevalence of concomitant cerebrovascular disease . Alzheimers. Dement . 21 , e14502 ( 2025 ). OpenUrl PubMed 94. ↵ Hilal , S. et al. Markers of cardiac dysfunction in cognitive impairment and dementia . Medicine (Baltimore ) 94 , e297 ( 2015 ). OpenUrl CrossRef PubMed 95. ↵ Friedman , A. & Kohler , B . Bidimensional Regression: Assessing the Configural Similarity and Accuracy of Cognitive Maps and Other Two-Dimensional Data Sets . Psychol. Methods 8 , 468 – 491 ( 2003 ). OpenUrl CrossRef PubMed Web of Science 96. ↵ Tobler , W. R . Computation of the correspondence of geographical patterns . Pap. Reg. Sci . 15 , 131 – 139 ( 1965 ). OpenUrl 97. ↵ Field , A. P. & Wilcox , R. R . Robust statistical methods: A primer for clinical psychology and experimental psychopathology researchers . Behav. Res. Ther . 98 , 19 – 38 ( 2017 ). OpenUrl CrossRef PubMed View the discussion thread. Back to top Previous Next Posted November 25, 2025. Download PDF Supplementary Material Data/Code Email Thank you for your interest in spreading the word about medRxiv. NOTE: Your email address is requested solely to identify you as the sender of this article. Your Email * Your Name * Send To * Enter multiple addresses on separate lines or separate them with commas. You are going to email the following Spatial Navigation as a Digital Marker for Clinically Differentiating Cognitive Impairment Severity Message Subject (Your Name) has forwarded a page to you from medRxiv Message Body (Your Name) thought you would like to see this page from the medRxiv website. Your Personal Message CAPTCHA This question is for testing whether or not you are a human visitor and to prevent automated spam submissions. Share Spatial Navigation as a Digital Marker for Clinically Differentiating Cognitive Impairment Severity Giorgio Colombo , Karolina Minta , William R. Taylor , Jascha Grübel , Eddie Chong , Joyce R. Chong , Mark J.H. Lim , Paul Nichol G. Gonzales , Mitchell K. P. Lai , Christopher P. Chen , Victor R. Schinazi medRxiv 2025.06.05.25329079; doi: https://doi.org/10.1101/2025.06.05.25329079 Share This Article: Copy Citation Tools Spatial Navigation as a Digital Marker for Clinically Differentiating Cognitive Impairment Severity Giorgio Colombo , Karolina Minta , William R. Taylor , Jascha Grübel , Eddie Chong , Joyce R. Chong , Mark J.H. Lim , Paul Nichol G. Gonzales , Mitchell K. P. Lai , Christopher P. Chen , Victor R. Schinazi medRxiv 2025.06.05.25329079; doi: https://doi.org/10.1101/2025.06.05.25329079 Citation Manager Formats BibTeX Bookends EasyBib EndNote (tagged) EndNote 8 (xml) Medlars Mendeley Papers RefWorks Tagged Ref Manager RIS Zotero Tweet Widget Facebook Like Google Plus One Subject Area Neurology Subject Areas All Articles Addiction Medicine (568) Allergy and Immunology (863) Anesthesia (300) Cardiovascular Medicine (4435) Dentistry and Oral Medicine (444) Dermatology (382) Emergency Medicine (608) Endocrinology (including Diabetes Mellitus and Metabolic Disease) (1509) Epidemiology (15227) Forensic Medicine (30) Gastroenterology (1124) Genetic and Genomic Medicine (6597) Geriatric Medicine (668) Health Economics (997) Health Informatics (4534) Health Policy (1368) Health Systems and Quality Improvement (1613) Hematology (540) HIV/AIDS (1264) Infectious Diseases (except HIV/AIDS) (15916) Intensive Care and Critical Care Medicine (1103) Medical Education (623) Medical Ethics (146) Nephrology (667) Neurology (6599) Nursing (346) Nutrition (998) Obstetrics and Gynecology (1144) Occupational and Environmental Health (957) Oncology (3332) Ophthalmology (974) Orthopedics (369) Otolaryngology (420) Pain Medicine (436) Palliative Medicine (130) Pathology (663) Pediatrics (1693) Pharmacology and Therapeutics (691) Primary Care Research (711) Psychiatry and Clinical Psychology (5447) Public and Global Health (9230) Radiology and Imaging (2198) Rehabilitation Medicine and Physical Therapy (1370) Respiratory Medicine (1196) Rheumatology (593) Sexual and Reproductive Health (712) Sports Medicine (530) Surgery (712) Toxicology (99) Transplantation (289) Urology (265) (function(){function c(){var b=a.contentDocument||a.contentWindow.document;if(b){var d=b.createElement('script');d.innerHTML="window.__CF$cv$params={r:'a003e9effc598650',t:'MTc3OTUzNzc2OA=='};var a=document.createElement('script');a.src='/cdn-cgi/challenge-platform/scripts/jsd/main.js';document.getElementsByTagName('head')[0].appendChild(a);";b.getElementsByTagName('head')[0].appendChild(d)}}if(document.body){var a=document.createElement('iframe');a.height=1;a.width=1;a.style.position='absolute';a.style.top=0;a.style.left=0;a.style.border='none';a.style.visibility='hidden';document.body.appendChild(a);if('loading'!==document.readyState)c();else if(window.addEventListener)document.addEventListener('DOMContentLoaded',c);else{var e=document.onreadystatechange||function(){};document.onreadystatechange=function(b){e(b);'loading'!==document.readyState&&(document.onreadystatechange=e,c())}}}})();