Related Experiment Video
Updated: Jun 4, 2026
A Machine Learning Approach to Design an Efficient Selective Screening of Mild Cognitive Impairment
Published on: January 11, 2020
Improving precision in the quantification of cognition using the Montreal Cognitive Assessment and the Mini-Mental
Lisa Koski1, Haiqun Xie, Susanna Konsztowicz
1Division of Experimental Medicine, Department of Medicine, McGill University, Montreal, Quebec, Canada. lisa.koski@mcgill.ca
Background:
The Montreal Cognitive Assessment (MoCA) can be used to quantify cognitive ability in older persons undergoing screening for cognitive impairment. Although highly sensitive in detecting mild cognitive impairment, its measurement precision is weakest among persons with milder forms of impairment. We sought to overcome this limitation by integrating information from the Mini-Mental State Examination (MMSE) into the calculation of cognitive ability.
Methods:
Data from 185 geriatric outpatients screened for cognitive impairment with the MoCA and the MMSE were Rasch analyzed to evaluate the extent to which the MMSE items improved measurement precision in the upper ability ranges of the population.
Results:
Adding information from the MMSE resulted in a 13.8% (13.3-14.3%) reduction in measurement error, with significant improvements in all quartiles of patient ability. The addition of three-word repetition and recall, copy pentagons, repeat sentence, and write sentence improved measurement of cognition in the upper levels of ability.
Conclusions:
The algorithm presented here maximizes the yield of available clinical data while improving measurement of cognitive ability, which is particularly important for tracking changes over time in patients with milder levels of impairment.
