Related Experiment Video
Updated: Apr 18, 2026
A Machine Learning Approach to Design an Efficient Selective Screening of Mild Cognitive Impairment
Published on: January 11, 2020
Montreal Cognitive Assessment for screening mild cognitive impairment: variations in test performance and scores by
Tze Pin Ng1, Lei Feng, Wee Shiong Lim
1Gerontology Research Programme, Department of Psychological Medicine, Yong Loo Lin School of Medicine, National University of Singapore, Singapore, Singapore.
Background:
The Montreal Cognitive Assessment (MoCA) was developed as a screening instrument for mild cognitive impairment (MCI). We evaluated the MoCA's test performance by educational groups among older Singaporean Chinese adults.
Method:
The MoCA and Mini-Mental State Examination (MMSE) were evaluated in two independent studies (clinic-based sample and community-based sample) of MCI and normal cognition (NC) controls, using receiver operating characteristic curve analyses: area under the curve (AUC), sensitivity (Sn), and specificity (Sp).
Results:
The MoCA modestly discriminated MCI from NC in both study samples (AUC = 0.63 and 0.65): Sn = 0.64 and Sp = 0.36 at a cut-off of 28/29 in the clinic-based sample, and Sn = 0.65 and Sp = 0.55 at a cut-off of 22/23 in the community-based sample. The MoCA's test performance was least satisfactory in the highest (>6 years) education group: AUC = 0.50 (p = 0.98), Sn = 0.54, and Sp = 0.51 at a cut-off of 27/28. Overall, the MoCA's test performance was not better than that of the MMSE. In multivariate analyses controlling for age and gender, MCI diagnosis was associated with a <1-point decrement in MoCA score (η(2) = 0.010), but lower (1-6 years) and no education was associated with a 3- to 5-point decrement (η(2) = 0.115 and η(2) = 0.162, respectively).
Conclusion:
The MoCA's ability to discriminate MCI from NC was modest in this Chinese population, because it was far more sensitive to the effect of education than MCI diagnosis.
