Glial and CSF Biomarkers for Cognitive Impairment Classification in 333 Adults

A dataset of 333 adults aged 60 years and older from the Knight Alzheimer’s Disease Research Center's longitudinal, open-access dataset. It includes cerebrospinal fluid biomarkers, glial biomarkers, and clinical features used to develop machine-learning models for classifying cognitive impairment. The data was published by Merve Türkegün Şengül on April 23, 2026.

Use Cases

• Training classification models to distinguish cognitive impairment based on biomarker levels.
• Evaluating the diagnostic value of glial biomarkers alongside core CSF biomarkers like tau and Aβ42.
• Developing parsimonious and interpretable machine-learning models for clinical stratification.
• Conducting feature importance analysis, such as SHAP, to identify influential biomarkers.

Strengths

• Includes data from 333 adults, providing a substantive sample size for analysis.
• Integrates multiple biomarker types (CSF and glial) and clinical features for multimodal modeling.
• Model performance metrics are reported, with an AUC of 0.959 and accuracy of 0.912 for the top model.
• Released under a CC-BY-4.0 license, permitting open reuse and modification.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Data is stored in a PDF file (426.6 KB), which may require extraction and preprocessing for analysis.

Provenance

Source

Knight Alzheimer’s Disease Research Center longitudinal, open-access dataset.

Collection Method

Data obtained from a research center's longitudinal study; machine-learning models were developed and evaluated.

Time Range

null

Freshness

Last updated 2026-04-23 04:20:27; freshness should be verified.

Geography

null