Related Experiment Video
Updated: Jan 7, 2026

A Metadata Extraction Approach for Clinical Case Reports to Enable Advanced Understanding of Biomedical Concepts
Published on: September 20, 2018
Clinical Manifestations
Daniel J Schulman1, David J Libon2,3, Ali Jannati1
1Linus Health, Boston, MA, USA.
This study used a digital cognitive assessment tool to classify participants into cognitively unimpaired, dementia, and two mild cognitive impairment (MCI) subtypes. The 7-minute digital assessment shows promise for identifying MCI subtypes in Alzheimer's disease clinical trials.
Area of Science:
- Neuroscience
- Biomarkers
- Digital Health
Background:
- Alzheimer's disease (AD) clinical trials face high screen failure rates.
- Digital biomarkers, including the Digital Assessment of Cognition (DAC), are being developed to improve participant selection.
- The Apheleia-001 study utilized DAC and blood-based biomarkers to reduce screen failures in AD trials.
Purpose of the Study:
- To develop and validate a method for classifying participants into cognitively unimpaired (CU), mild cognitive impairment (MCI), and dementia cohorts using DAC metrics.
- To explore the utility of DAC in differentiating between subtypes of MCI.
- To augment limited labeled data with a larger unlabeled dataset using a semi-supervised approach.
Main Methods:
- Combined Apheleia-001 data (n=1189) with a memory clinic dataset (n=106) with established diagnoses.
- Applied modified finite mixture clustering models to DAC scores to identify 3-5 clusters (CU, dementia, 1-3 MCI).
- Selected a final 4-cluster model by expert consensus, validating with age, education, Clock and Recall (DCR), and Mini-Mental State Examination (MMSE) scores.
Main Results:
- A 4-cluster model was selected, identifying distinct CU, dementia, and two MCI subtypes (amnestic and mixed/dysexecutive).
- Significant differences in age, education, DCR, and MMSE scores were observed between clusters, aligning with hypothesized cognitive impairment levels.
- The amnestic MCI cluster showed significantly lower scores on DCR delayed free-recall compared to the mixed/dysexecutive MCI cluster.
Conclusions:
- The 7-minute DAC automated assessment can effectively differentiate between cognitively intact individuals, dementia patients, and two distinct MCI subtypes.
- This digital tool shows potential for improving diagnostic accuracy and participant stratification in clinical trials for Alzheimer's disease.
- Further validation is required, but the DAC demonstrates promise as a scalable and efficient cognitive assessment method.
Related Concept Videos
Chronic Kidney Disease II: Clinical Manifestations
Coronary Artery Disease III: Clinical Manifestations
Endocarditis II: Clinical Features of Infective Endocarditis
Heart Failure III: Clinical Manifestations
Gastroesophageal Reflux Disease II: Clinical Features and Management
Clinical Manifestations
GERD presents itself in a multitude of ways, with symptoms varying from person to person. The hallmark symptoms are...
Hypertension III: Clinical Manifestations and Diagnostic Studies

