Related Experiment Video
Updated: Apr 12, 2026
A Machine Learning Approach to Design an Efficient Selective Screening of Mild Cognitive Impairment
Published on: January 11, 2020
An Abbreviated Montreal Cognitive Assessment (MoCA) for Dementia Screening
Daniel K Horton1, Linda S Hynan, Laura H Lacritz
1a Department of Psychiatry , The University of Texas Southwestern Medical Center , Dallas , TX , USA.
Objective:
The Montreal Cognitive Assessment (MoCA) is a cognitive screening instrument growing in popularity, but few studies have conducted psychometric item analyses or attempted to develop abbreviated forms. We sought to derive and validate a short-form MoCA (SF-MoCA) and compare its classification accuracy to the standard MoCA and Mini-Mental State Examination (MMSE) in mild cognitive impairment (MCI), Alzheimer disease (AD), and normal aging.
Methods:
408 subjects (MCI n = 169, AD n = 87, and normal n = 152) were randomly divided into derivation and validation samples. Item analysis in the derivation sample identified most sensitive MoCA items. Receiver Operating Characteristic (ROC) analyses were used to develop cut-off scores and evaluate the classification accuracy of the SF-MoCA, standard MoCA, and MMSE. Net Reclassification Improvement (NRI) analyses and comparison of ROC curves were used to compare classification accuracy of the three measures.
Results:
Serial subtraction (Cramer's V = .408), delayed recall (Cramer's V = .702), and orientation items (Cramer's V = .832) were included in the SF-MoCA based on largest effect sizes in item analyses. Results revealed 72.6% classification accuracy of the SF-MoCA, compared with 71.9% for the standard MoCA and 67.4% for the MMSE. Results of NRI analyses and ROC curve comparisons revealed that classification accuracy of the SF-MoCA was comparable to the standard version and generally superior to the MMSE.
Conclusions:
Findings suggest the SF-MoCA could be an effective brief tool in detecting cognitive impairment.
