Face recognition deficits occur in diseases such as prosopagnosia, autism, Alzheimer's disease, and dementias. The objective was to evaluate whether degrading the architecture of artificial intelligence (AI) face recognition algorithms can model deficits in diseases. Two established face recognition models, convolutional-classification neural network (C-CNN) and Siamese network (SN), were trained on the FEI faces dataset (~14 images/person for 200 persons). The trained networks were perturbed by reducing weights (weakening) and node count (lesioning) to emulate brain tissue dysfunction and lesions, respectively. Accuracy assessments were used as surrogates for face recognition deficits. The findings were compared to clinical outcomes from the Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset. Face recognition accuracy decreased gradually for weakening factors less than 0.55 for C-CNN, and 0.85 for SN. Rapid accuracy loss occurred at higher values. C-CNN accuracy was similarly affected by weakening any convolutional layer whereas SN accuracy was more sensitive to weakening of the first convolutional layer. SN accuracy declined gradually with a rapid drop when nearly all nodes were lesioned. C-CNN accuracy declined rapidly when as few as 10% of nodes were lesioned. CNN and SN were more sensitive to lesioning of the first convolutional layer. Overall, SN was more robust than C-CNN, and the findings from SN experiments were concordant with ADNI results. As predicted from modeling, brain network failure quotient was related to key clinical outcome measures for cognition and functioning. Perturbation of AI networks is a promising method for modeling disease progression effects on complex cognitive outcomes.

Repurposing Artificial Intelligence Tools for Disease Modeling: Case Study of Face Recognition Deficits in Neurodegenerative Diseases.

Objective: The Stricker Learning Span (SLS) is a computer-adaptive digital word list memory test specifically designed for remote assessment and self-administration on a web-based multi-device platform (Mayo Test Drive). We aimed to establish criterion validity of the SLS by comparing its ability to differentiate biomarker-defined groups to the person-administered Auditory Verbal Learning Test (AVLT). Participants and Methods: Participants (N=353; mean age=71, SD=11; 93% cognitively unimpaired [CU]) completed the AVLT during an in-person visit, the SLS remotely (within 3 months) and had brain amyloid and tau PET scans available (within 3 years). Overlapping groups were formed for 1) those on the Alzheimer disease (AD) continuum (amyloid PET positive, A+, n=125) or not (A-, n=228), and those with biological AD (amyloid and tau PET positive, A+T+, n=55) vs no evidence of AD pathology (A-T-, n=195). Analyses were repeated among CU participants only. Results: The SLS and AVLT showed similar ability to differentiate biomarker-defined groups when comparing AUROCs (p>.05). In logistic regression models, SLS contributed significantly to predicting biomarker group beyond age, education and sex, including when limited to CU participants. Medium (A- vs A+) to large (A-T- vs A+T+) unadjusted effect sizes were observed for both SLS and AVLT. Learning and delay variables were similar in terms of ability to separate biomarker groups. Conclusions: Remotely administered SLS performed similarly to in-person-administered AVLT in its ability to separate biomarker-defined groups, providing evidence of criterion validity. Results suggest the SLS may be sensitive to detecting subtle objective cognitive decline in preclinical AD.

/pdf/stricker-learning-span-criterion-validity-a-remote-self-2q6h27qt.pdf

Stricker Learning Span criterion validity: a remote self-administered multi-device compatible digital word list memory measure shows similar ability to differentiate amyloid and tau PET-defined biomarker groups as in-person Auditory Verbal Learning Test

Abstract Background Easily accessible and self-administered cognitive assessments that can aid early detection for Alzheimer’s disease (AD) dementia risk are critical for timely intervention. Objectives/Design This cross-sectional study investigated continuous associations between Mayo Test Drive (MTD) – a remote, self-administered, multi-device compatible, web-based cognitive assessment – and AD-related imaging biomarkers. Participants/Setting 684 adults from the Mayo Clinic Study of Aging and Mayo Clinic Alzheimer’s Disease Research Center participated (age=70.4±11.2, 49.7% female). Participants were predominantly cognitively unimpaired (CU; 94.0%). Measurements Participants completed (1) brain amyloid and tau PET scans and MRI scans for hippocampal volume (HV) and white matter hyperintensities (WMH); (2) MTD remotely, consisting of the Stricker Learning Span and Symbols Test which combine into an MTD composite; and (3) in-person neuropsychological assessment including measures to obtain Mayo Preclinical Alzheimer’s disease Cognitive Composite (Mayo-PACC) and Global-z. Multiple regressions adjusted for age, sex, and education queried associations between imaging biomarkers and scores from remote and in-person cognitive measures. Results Lower performances on MTD were associated with greater amyloid, entorhinal tau, and global tau PET burden, lower HV, and higher WMH. Mayo-PACC and Global-z were associated with all imaging biomarkers except global tau PET burden. MCI/Dementia participants showed lower performance on all MTD measures compared to CU with large effect sizes (Hedge’s g’s=1.65-2.02), with similar findings for CU versus MCI only (Hedge’s g’s=1.46-1.83). Conclusions MTD is associated with continuous measures of AD-related imaging biomarkers, demonstrating ability to detect subtle cognitive change using a brief, remote assessment in predominantly CU individuals and criterion validity for MTD. 

Continuous Associations Between Remote Self-Administered Cognitive Measures and Imaging Biomarkers of Alzheimers Disease

Abstract Objective Few normative data for unsupervised, remotely-administered computerized cognitive measures are available. We examined variables to include in normative models for Mayo Test Drive (a multi-device remote cognitive assessment platform) measures, developed normative data, and validated the norms. Method 1240 Cognitively Unimpaired (CU) adults ages 32-100-years (96% white) from the Mayo Clinic Study of Aging and Mayo Alzheimer’s Disease Research Center with Clinical Dementia Rating® of 0 were included. We converted raw scores to normalized scaled scores and derived regression-based normative data adjusting for age, age 2 , sex and education (base model); alternative norms are also provided (age+age 2 +sex; age+age 2 ). We assessed additional terms using an a priori cut-off of 1% variance improvement above the base model. We examined low test performance rates (&lt;-1 standard deviation) in independent validation samples (n=167 CU, n=64 mild cognitive impairment (MCI), n=14 dementia). Rates were significantly different when 95% confidence intervals (CI) did not include the expected 14.7% base rate. Results No model terms met the a priori cut-off beyond the base model, including device type, response input source (e.g., mouse, etc.) or session interference. Norms showed expected low performance rates in CU and greater rates of low performance in MCI and dementia in independent validation samples. Conclusion Typical normative models appear appropriate for remote self-administered MTD measures and are sensitive to cognitive impairment. Device type and response input source did not explain enough variance for inclusion in normative models but are important for individual-level interpretation. Future work will increase inclusion of individuals from under-represented groups. 

Mayo Normative Studies: regression-based normative data for remote self-administration of the Stricker Learning Span, Symbols Test and Mayo Test Drive Screening Battery Composite and validation in individuals with Mild Cognitive Impairment and dementia

Abstract Objective: The psychometric rigor of unsupervised, smartphone-based assessments and factors that impact remote protocol engagement is critical to evaluate prior to the use of such methods in clinical contexts. We evaluated the validity of a high-frequency, smartphone-based cognitive assessment protocol, including examining convergence and divergence with standard cognitive tests, and investigating factors that may impact adherence and performance (i.e., time of day and anticipated receipt of feedback vs. no feedback). Methods: Cognitively unimpaired participants (N = 120, M age = 68.8, 68.3% female, 87% White, M education = 16.5 years) completed 8 consecutive days of the Mobile Monitoring of Cognitive Change (M2C2), a mobile app-based testing platform, with brief morning, afternoon, and evening sessions. Tasks included measures of working memory, processing speed, and episodic memory. Traditional neuropsychological assessments included measures from the Preclinical Alzheimer’s Cognitive Composite battery. Results: Findings showed overall high compliance (89.3%) across M2C2 sessions. Average compliance by time of day ranged from 90.2% for morning sessions, to 77.9% for afternoon sessions, and 84.4% for evening sessions. There was evidence of faster reaction time and among participants who expected to receive performance feedback. We observed excellent convergent and divergent validity in our comparison of M2C2 tasks and traditional neuropsychological assessments. Conclusions: This study supports the validity and reliability of self-administered, high-frequency cognitive assessment via smartphones in older adults. Insights into factors affecting adherence, performance, and protocol implementation are discussed. 

Psychometric and adherence considerations for high-frequency, smartphone-based cognitive screening protocols in older adults

OBJECTIVE
Growing evidence supports the importance of learning as a central deficit in preclinical/prodromal Alzheimer's disease. The aims of this study were to conduct a series of neural network simulations to develop a functional understanding of a distributed, nonmodular memory system that can learn efficiently without interference. This understanding is applied to the development of a novel digital memory test.


METHOD
Simulations using traditional feed forward neural network architectures to learn simple logic problems are presented. The simulations demonstrate three limitations: (a) inefficiency, (b) an inability to learn problems consistently, and (c) catastrophic interference when given multiple problems. A new mirrored cascaded architecture is introduced to address these limitations, with support provided by a series of simulations.


RESULTS
The mirrored cascaded architecture demonstrates efficient and consistent learning relative to feed forward networks but also suffers from catastrophic interference. Addition of context values to add the capability of distinguishing features as part of learning eliminates the problem of interference in the mirrored cascaded, but not the feed forward, architectures.


CONCLUSIONS
A mirrored cascaded architecture addresses the limitations of traditional feed forward neural networks, provides support for a distributed memory system, and emphasizes the importance of context to avoid interference. These process models contributed to the design of a digital computer-adaptive word list learning test that places maximum stress on the capability to distinguish specific episodes of learning. Process simulations provide a useful method of testing models of brain function and contribute to new approaches to neuropsychological assessment. (PsycInfo Database Record (c) 2022 APA, all rights reserved).

Neural network process simulations support a distributed memory system and aid design of a novel computer adaptive digital memory test for preclinical and prodromal Alzheimer's disease.

Handle everyday research tasks with reliable, citation-backed results

Your personal Research Agent to handle research tasks with citation-backed results

Popular Tasks used by Researchers

How can I help with your research?

Meet SciSpace

Get more enhanced response by uploading the PDFs you want me to reference.

No relevant PDFs in your library

SciSpace is the AI research assistant for academics. Run systematic literature reviews on 280M+ papers, and write papers with cited sources. Trusted by 1M+ students, PhDs & researchers.

SciSpace | AI for Research

Analyze PDFs

Code & Manuscripts

Funding & Grants

Literature & Patents

Medical & Clinical Data

Systematic Review

Visualize & Present

Web & Data

Build a Google Scholar-like website for your research.

Build a website

Create charts and images for your research

Create a Chart

Write a paper for submission to a journal

Draft a manuscript

Patent Search

Design eye-catching scientific posters in minutes.

Scientific Poster Generation

Systematic Literature Review

One task is running at the moment. Your messages will be shown right after.

Drag and drop or click here to browse

Loved by <highlight>1 million+</highlight> researchers

Extract a list of specific topics and their sources from unstructured text

Topics

Compare and analyze relevant papers that matches with your search

Papers

Get insights from PDFs and bookmarked papers from your library

My library

Recent searches

Try searching for:

Catch AI-generated content in scholarly and non-scholarly content

{ai} Detector

Ai Writer

Get PDF Summaries, highlighted text explanations 

Chat with PDF

Effortlessly create in-text citations and bibliographies in APA and 2,500 other formats

Citation generator

Get explanations, summaries, and answers on academic papers

Ease up your research workflow with {scispace}'s cohort of exciting AI tools

Elevate your academic writing skills and convey your ideas the way you want

Paraphraser

Explore our range of reading and writing tools

Your file is being prepared and should be ready in a few minutes. If it's a large file, it might take a bit longer. You can close this window, and we'll email you the file when it's done.

You have reached a maximum limit of <strong>{limit}</strong> columns in the table. Remove at least <strong>1</strong> column to add or create another one.

Hugo Botha

Author Tools

Chat about Author

Papers

Neural network process simulations support a distributed memory system and aid design of a novel computer adaptive digital memory test for preclinical and prodromal Alzheimer's disease.