Explore the vast range of titles available.

Digital cognitive assessments have rapidly expanded in Alzheimer's disease (AD) research, offering a sensitive, scalable, and cost‐effective alternative to traditional neuropsychological tests. This systematic review examines the validation and utility of digital cognitive assessments in cognitively normal (CN) individuals and explores their potential classification within the US Food and Drug Administration's Biomarkers, Endpoints, and other Tools (FDA BEST) framework. Additionally, we provide recommendations to consider for their implementation. Following the Preferred Reporting Items for Systematic Reviews and Meta‐Analyses (PRISMA) guidelines, we searched PubMed for studies validating digital cognitive tools against paper‐based tests and standard AD biomarkers, including measures of amyloid beta and tau in fluid and neuroimaging biomarkers. Our findings suggest potential use as risk or monitoring biomarkers, though further longitudinal validation is needed. This review highlights the latest advancements in digital cognitive assessments, their role as novel AD biomarkers, and essential considerations for their effective use in AD. Highlights Digital cognitive assessments for preclinical Alzheimer's disease (AD) are associated with established biomarkers, including paper‐based neuropsychological tests and amyloid beta and tau measures in both fluid and neuroimaging techniques. These assessments have the potential to serve as novel AD biomarkers classified within the US Food and Drug Administration's Biomarkers, Endpoints, and other Tools framework and context of use, but long‐term studies spanning different disease stages are needed to fully establish their validity for some of the biomarker categories. Several biases may be present when conducting digital cognitive assessments; their optimal use should follow specific recommendations to minimize them.

The original counterclockwise study carried out in the late 1970s provided an extreme example of \"reminiscence therapy,\" reporting improvements in older adults' cognitive and physical functioning after they had lived for 5 days in a house set up as if decades earlier (the 1950s). We tested a virtual reality (VR) analog of this approach, enhanced by embodying participants in a virtual body that looked like themselves at the corresponding younger age. This study aimed to examine whether brief VR exposures combining (1) embodiment in a virtual body as one's younger self and (2) immersion in an iconic past event improve age-related subjective and performance outcomes compared with a current-self VR control condition. We carried out a between-groups study with 23 healthy older adults (aged 65-85 years; mean age 71.2, SD 4.03 years). Participants were randomly allocated to either a Young Self condition (n=11; mean 72.3, SD 4.17), where they were embodied in a virtual body that looked like themselves from the 1960s, or in a Current Self control condition (n=12; mean 70.1, SD 3.75), where participants were embodied in their current body. There were 5 sessions. In Session 1, participants completed a baseline assessment. There were then 2 VR exposures, approximately 1 week apart (Sessions 2-3), and follow-ups at 1 week (Session 4) and approximately 2 weeks (Session 5) after the final VR exposure. Outcomes included subjective age, awareness of age-related change, World Health Organization-Five Well-Being Index, Trail Making Test performance, and physical functioning (eg, grip strength). A hierarchical Bayesian analysis revealed that 1 week after the final VR exposure, those in the Young Self condition demonstrated lower subjective age than those in the Current Self condition (prob=.95). They had higher awareness of positive age-related change (prob=.89) and a higher score on the World Health Organization-Five Well-Being Index (prob=.84). Moreover, with respect to performance variables, they took less time to trace a trail (prob≥.99), made fewer mistakes in doing so (prob=.89), had greater right-hand (prob=.85) and left-hand (prob≥.99) grip strength. However, 2 weeks after their final VR exposure, these differences diminished apart from positive awareness of age-related change (prob=.82), trail-making mistakes (prob=.83), and left-hand grip strength (prob≥.99). Here, \"prob\" refers to posterior probability. The results demonstrate that even 2 short VR exposures, where people were embodied in their younger body and immersed in an iconic event from more than 50 years earlier, resulted in improvement in some age-related responses. This is encouraging for further research with more extensive VR experiences over a longer time period.

INTRODUCTION Early detection of Alzheimer's disease (AD) is critical for timely intervention as disease‐modifying treatments emerge. Speech‐based digital biomarkers offer scalable options for remotely capturing speech‐derived functional changes associated with early cognitive decline, but validation across real‐world populations remains limited. METHODS We evaluated the speech biomarker for cognition (SB‐C), an automated speech‐derived measure associated with cognitive status, in 736 participants across five European cohorts (Barcelonaβeta Brain Research Center's Alzheimer's at‐risk cohort, European Prevention of Alzheimer's Dementia Scotland, Dementia Study of Cognitive and Biomarker Dynamics, Longitudinal Cognitive Impairment and Dementia Study, and Biomarkers for Identifying Neurodegenerative Disorders Early and Reliably [BioFINDER‐Primary Care]). Participants completed verbal learning and semantic fluency tasks via automated phone or app‐based platforms. SB‐C performance was compared to Mini‐Mental State Examination, Clinical Dementia Rating, Preclinical Alzheimer Cognitive Composite 5, and cerebrospinal fluid amyloid beta and phosphorylated tau181 biomarker status. RESULTS SB‐C significantly differentiated cognitively unimpaired and impaired groups (P < 0.001), correlated with standard cognitive measures, and showed moderate‐to‐high area under the curve (0.56–0.82) for classifying biomarker positivity, with strongest results in BioFINDER‐Primary Care. DISCUSSION SB‐C is a scalable, remote speech‐derived marker associated with cognitive status and AD biomarker group differences. Highlights The speech biomarker for cognition (SB‐C) detects early Alzheimer's disease (AD)‐related cognitive impairment remotely. The SB‐C was validated across five European cohorts with diverse cognitive and biomarker profiles. SB‐C scores associated with cerebrospinal fluid amyloid beta and phosphorylated tau181 biomarker positivity. Significant differences were observed in cognitively unimpaired individuals with subjective cognitive decline by amyloid/tau status. SB‐C supports scalable screening for AD in decentralized clinical trial settings.

Remote cognitive assessments offer a scalable and efficient alternative to traditional paper-and-pencil tests for large-scale research and clinical studies. However, factors such as digital literacy, demographics, and socioeconomic status may influence participant engagement. This study examined the influence of these factors on response rate (test completion) in the web-based FLAME cognitive assessment among non-demented participants in BBRC observational cohorts. 609 participants (mean[SD] age:67[5.3]; 354 women; mean[SD] education:14.2[3.6]) were invited by e-mail to complete FLAME. Previously, they had been invited to complete an ad-hoc questionnaire assessing digital literacy, which included seven self-rated items on familiarity and comfort with ICTs use. Based on a median split, participants were classified as having high or low ICT literacy. Age, sex, years of education, household income, subjective socioeconomic status (SES) (low, medium, or high), and living area within Catalonia, (levels: highly populated, medium populated, and rural) were also collected. Univariate and multivariate logistic regression models were used to examine if digital literacy, demographic, and socioeconomic factors predicted test completion. Additionally, predictors of digital literacy were explored using demographic and socioeconomic variables. The FLAME test completion rate was 44.49%. High ICT literacy emerged as the strongest predictor of FLAME test completion (OR: 2.31; p =  0.002), independent of age, sex, education, income, and SES. In a multivariate model assessing demographic and socioeconomic as predictors of ICT literacy, younger age (OR: 0.87; p =  <0.0001) and higher education (OR: 1.18; p =  <0.0001) were significant predictors of higher ICT literacy, while female sex (OR: 0.57; p =  0.051) showed a strong trend toward lower ICT literacy, and income and SES were not significant. Although living area did not significantly influence ICT literacy, participants from rural areas had a higher FLAME completion rate (72.4%) compared to those from medium-density (50%) and high-density (54%) populated areas. Higher digital literacy was significantly associated with higher response rates in remote cognitive assessments, emphasizing the importance of assessing it beforehand. Additionally, demographic and socioeconomic factors may influence digital literacy and should be considered in these evaluations.

Background Remote cognitive assessments offer a scalable and efficient alternative to traditional paper‐and‐pencil tests for large‐scale research and clinical studies. However, factors such as digital literacy, demographics, and socioeconomic status may influence participant engagement. This study examined the influence of these factors on response rate (test completion) in the web‐based FLAME cognitive assessment among non‐demented participants in BBRC observational cohorts. Method 609 participants (mean[SD] age:67[5.3]; 354 women; mean[SD] education:14.2[3.6]) were invited by e‐mail to complete FLAME. Previously, they had been invited to complete an ad‐hoc questionnaire assessing digital literacy, which included seven self‐rated items on familiarity and comfort with ICTs use. Based on a median split, participants were classified as having high or low ICT literacy. Age, sex, years of education, household income, subjective socioeconomic status (SES) (low, medium, or high), and living area within Catalonia, (levels: highly populated, medium populated, and rural) were also collected. Univariate and multivariate logistic regression models were used to examine if digital literacy, demographic, and socioeconomic factors predicted test completion. Additionally, predictors of digital literacy were explored using demographic and socioeconomic variables. Result The FLAME test completion rate was 44.49%. High ICT literacy emerged as the strongest predictor of FLAME test completion (OR: 2.31; p = 0.002), independent of age, sex, education, income, and SES. In a multivariate model assessing demographic and socioeconomic as predictors of ICT literacy, younger age (OR: 0.87; p = <0.0001) and higher education (OR: 1.18; p = <0.0001) were significant predictors of higher ICT literacy, while female sex (OR: 0.57; p = 0.051) showed a strong trend toward lower ICT literacy, and income and SES were not significant. Although living area did not significantly influence ICT literacy, participants from rural areas had a higher FLAME completion rate (72.4%) compared to those from medium‐density (50%) and high‐density (54%) populated areas. Conclusion Higher digital literacy was significantly associated with higher response rates in remote cognitive assessments, emphasizing the importance of assessing it beforehand. Additionally, demographic and socioeconomic factors may influence digital literacy and should be considered in these evaluations.

The use of digital biomarkers to assess cognition in Alzheimer's disease (AD) offers scalable, efficient alternatives to paper-and-pencil tests. Validating these tools against clinical and biomarker-defined groups is critical for their adoption in research, clinical trials and clinical contexts. This study evaluates the performance of a remote, unsupervised cognitive assessment (FLAME-Factors of Longitudinal Attention, Memory and Executive Function) in distinguishing cognitive profiles across diagnostic categories and amyloid status in two cohorts from BBRC. Cognitively normal (CN) participants from ALFA+ cohort and subjective cognitive decline (SCD) or mild cognitive impairment (MCI) patients from Beta-AARC cohort were invited via email to FLAME remote and unsupervised assessment. 249 participants completed FLAME tasks, that include working memory (Self Ordered Search Score, Paired Associate Learning Score, Digit Span Score), episodic memory (Picture Recognition Accuracy), attention (Digit Vigilance Accuracy, Digit Vigilance False Alarms, Digit Vigilance Reaction Time Mean, Choice Reaction Time Accuracy) and executive function (Verbal Reasoning Accuracy). Analysis of covariance (ANCOVA) with post-hoc (Tukey) were used to examine differences by clinical and amyloid status. Logistic regression models were employed to evaluate if the digital tasks predicted MCI. All analyses were adjusted for age, sex, and education. MCI group showed reduced performance in paired associate learning score, attention variables, picture recognition accuracy and verbal reasoning compared to CN and SCD participants. Digit vigilance false alarms, picture recognition accuracy and verbal reasoning accuracy were able to significantly distinguish between CN and SCD groups (Figure 1). Several cognitive variables significantly predicted MCI, including paired associate learning score (OR=1.93,95%CI[1.09-3.51],p=0.02), digit vigilance accuracy (OR=1.17,95%CI[1.02-1.35],p=0.02) and false alarms (OR=1.4,95%CI[1.13-1.76],p=0.002), and accuracy from choice reaction time task (OR=1.23,95%CI[1.03-1.47],p=0.01), picture recognition (OR=1.39,95%CI[1.17-1.72],p<0.001), and verbal reasoning (OR=1.05,95%CI [1.01-1.11],p=0.03). Additionally, self ordered search score and picture recognition accuracy were significantly lower in amyloid-positive individuals (Figure 2). A remote unsupervised assessment reliably differentiates diagnostic and AD biomarker-defined groups and predicts MCI, underscoring its promise value for research and clinical contexts.

Background The use of digital biomarkers to assess cognition in Alzheimer’s disease (AD) offers scalable, efficient alternatives to paper‐and‐pencil tests. Validating these tools against clinical and biomarker‐defined groups is critical for their adoption in research, clinical trials and clinical contexts. This study evaluates the performance of a remote, unsupervised cognitive assessment (FLAME‐Factors of Longitudinal Attention, Memory and Executive Function) in distinguishing cognitive profiles across diagnostic categories and amyloid status in two cohorts from BBRC. Method Cognitively normal (CN) participants from ALFA+ cohort and subjective cognitive decline (SCD) or mild cognitive impairment (MCI) patients from Beta‐AARC cohort were invited via email to FLAME remote and unsupervised assessment. 249 participants completed FLAME tasks, that include working memory (Self Ordered Search Score, Paired Associate Learning Score, Digit Span Score), episodic memory (Picture Recognition Accuracy), attention (Digit Vigilance Accuracy, Digit Vigilance False Alarms, Digit Vigilance Reaction Time Mean, Choice Reaction Time Accuracy) and executive function (Verbal Reasoning Accuracy). Analysis of covariance (ANCOVA) with post‐hoc (Tukey) were used to examine differences by clinical and amyloid status. Logistic regression models were employed to evaluate if the digital tasks predicted MCI. All analyses were adjusted for age, sex, and education. Result MCI group showed reduced performance in paired associate learning score, attention variables, picture recognition accuracy and verbal reasoning compared to CN and SCD participants. Digit vigilance false alarms, picture recognition accuracy and verbal reasoning accuracy were able to significantly distinguish between CN and SCD groups (Figure 1). Several cognitive variables significantly predicted MCI, including paired associate learning score (OR=1.93,95%CI[1.09‐3.51],p=0.02), digit vigilance accuracy (OR=1.17,95%CI[1.02‐1.35],p=0.02) and false alarms (OR=1.4,95%CI[1.13‐1.76],p=0.002), and accuracy from choice reaction time task (OR=1.23,95%CI[1.03‐1.47],p=0.01), picture recognition (OR=1.39,95%CI[1.17‐1.72],p<0.001), and verbal reasoning (OR=1.05,95%CI [1.01‐1.11],p=0.03). Additionally, self ordered search score and picture recognition accuracy were significantly lower in amyloid‐positive individuals (Figure 2). Conclusion A remote unsupervised assessment reliably differentiates diagnostic and AD biomarker‐defined groups and predicts MCI, underscoring its promise value for research and clinical contexts.

We aimed to provide validation data on Mili-platform-a remote, automated speech-based cognitive assessment conducted via telephone-by comparing speech-derived cognitive scores with standard cognitive paper-and-pencil evaluations in a cohort of individuals with Subjective Cognitive Decline (SCD). Data were collected from 91 participants (mean[SD]age:65.8[6.7]; 59,34% women; mean[SD]education:15.5[3.8]) with SCD, enrolled in the B-AARC cohort in Spain, and participating in PROSPECT-AD study, a multi-cohort European longitudinal study aiming to develop algorithms to identify speech biomarkers for AD. The B-AARC cohort included paper-and-pencil cognitive assessments, from which composite scores for executive function, episodic memory, and language were derived, as well as the PACC for a global cognition score. Speech samples were automatically collected during cognitive tasks via the Mili platform (ki:elements) by phone, including a semantic verbal fluency test and a four-trial 15-word Auditory Verbal Learning Test. These tasks generated speech-based cognitive scores (SB-C), measuring three cognitive domains-memory, processing speed, and executive function-and a global cognition score. Pearson and Spearman rank correlations were used to evaluate associations between SB-C domain scores and paper-and-pencil cognitive domains. SB-C scores demonstrated moderate correlations with paper-and-pencil cognitive domains. The SB-C global cognition score significantly correlated with the PACC (r:0.57; p-value:<0.0001). The SB-C executive function score correlated with the paper-based executive function composite (r:0.56; p-value:<0.0001), as well as with the paper-based language composite (r:0.64; p-value:<0.0001). The SB-C memory score correlated with the paper-based memory domain, though weakly (r:0.22; p-value:0.03), as well as the SB-C processing speed score and the paper-based attention composite (r:0.26; p-value:0.01). These findings provided validation on the utility of remote, automated speech-based cognitive assessments in individuals at risk of AD compared to traditional assessments, offering a more scalable and accessible tool. Some domains displayed low correlations, suggesting that adding speech features to the composites provides additional information, potentially more sensitive and granular to actual changes, which may be not reflected in the paper-pencil tests. Further studies on predicting brain pathology and cognitive decline are warranted.

Background We aimed to provide validation data on Mili‐platform–a remote, automated speech‐based cognitive assessment conducted via telephone–by comparing speech‐derived cognitive scores with standard cognitive paper‐and‐pencil evaluations in a cohort of individuals with Subjective Cognitive Decline (SCD). Method Data were collected from 91 participants (mean[SD]age:65.8[6.7]; 59,34% women; mean[SD]education:15.5[3.8]) with SCD, enrolled in the B‐AARC cohort in Spain, and participating in PROSPECT‐AD study, a multi‐cohort European longitudinal study aiming to develop algorithms to identify speech biomarkers for AD. The B‐AARC cohort included paper‐and‐pencil cognitive assessments, from which composite scores for executive function, episodic memory, and language were derived, as well as the PACC for a global cognition score. Speech samples were automatically collected during cognitive tasks via the Mili platform (ki:elements) by phone, including a semantic verbal fluency test and a four‐trial 15‐word Auditory Verbal Learning Test. These tasks generated speech‐based cognitive scores (SB‐C), measuring three cognitive domains—memory, processing speed, and executive function—and a global cognition score. Pearson and Spearman rank correlations were used to evaluate associations between SB‐C domain scores and paper‐and‐pencil cognitive domains. Result SB‐C scores demonstrated moderate correlations with paper‐and‐pencil cognitive domains. The SB‐C global cognition score significantly correlated with the PACC (r:0.57; p‐value:<0.0001). The SB‐C executive function score correlated with the paper‐based executive function composite (r:0.56; p‐value:<0.0001), as well as with the paper‐based language composite (r:0.64; p‐value:<0.0001). The SB‐C memory score correlated with the paper‐based memory domain, though weakly (r:0.22; p‐value:0.03), as well as the SB‐C processing speed score and the paper‐based attention composite (r:0.26; p‐value:0.01). Conclusion These findings provided validation on the utility of remote, automated speech‐based cognitive assessments in individuals at risk of AD compared to traditional assessments, offering a more scalable and accessible tool. Some domains displayed low correlations, suggesting that adding speech features to the composites provides additional information, potentially more sensitive and granular to actual changes, which may be not reflected in the paper‐pencil tests. Further studies on predicting brain pathology and cognitive decline are warranted.