A cognitively demanding dictation task could help differentiate older adults with cognitive impairment from cognitively healthy peers, whereas simpler handwriting exercises could show limited ability to distinguish the groups, according to a recent study.
The cross-sectional study included 58 institutionalized older adults residing in Portugal, 38 of whom had cognitive impairment and 20 of whom were cognitively healthy. Researchers classified cognitive status using education-adjusted Mini-Mental State Examination scores, Clock Drawing Test scores, and medical records confirming cognitive impairment.
The participants completed two pen-control tasks and four handwriting-speed tasks involving sentence copying and dictation on a digitizing tablet. The researchers analyzed nine handwriting kinematic variables, including vertical size, horizontal size, slant, start time, duration, relative pen-down duration, normalized y-jerk, average pen pressure, and number of strokes.
The third handwriting-speed task, a dictation exercise, yielded the strongest results.. Duration and number of strokes were significantly associated with cognitive impairment, while vertical size and start time showed marginal associations.
The dot task model was not statistically significant. While the line task model reached statistical significance overall, no individual kinematic predictor independently contributed to classification following correction for multiple comparisons.
The researchers also evaluated whether handwriting kinematic measures predicted handwriting performance. Among cognitively healthy participants, associations between process variables and handwriting speed were limited. Just horizontal size in the third dictation task predicted handwriting speed.
Among the participants with cognitive impairment, stronger associations emerged between handwriting kinematics and handwriting performance. In the second dictation task, start time and number of strokes significantly predicted handwriting speed. In the third dictation task, vertical size, start time, and duration significantly predicted handwriting performance, while normalized y-jerk approached statistical significance.
“Dictation tasks, which require phonological decoding, orthographic retrieval, serial maintenance, and motor programming, impose substantial demands on the phonological loop and central executive, as described in working memory models,” wrote lead study author João Galrinho, of the Department of Sports and Health at the School of Health and Human Development at the Universidade de Évora and Comprehensive Health Research Centre Polo di Evora in Portugal, and colleagues. “[D]iagnostic differentiation emerged only under tasks imposing higher cognitive-motor demands,” they added.
The researchers used binary logistic regression and multiple linear regression models with Bonferroni correction for multiple comparisons. They imputed missing data below 5% using median values prior to their analysis. Post-hoc power analyses suggested adequate sensitivity in detecting medium-to-large effects, although subgroup analyses had lower sensitivity for small effects because of sample size limitations.
There were several limitations noted. The researchers did not systematically analyze medication use; however, the participants with neurologic or psychiatric disorders known to affect motor control were excluded. Because all participants were recruited from residential care facilities, the findings may not apply to older adults living independently. The cross-sectional design also limited conclusions about temporal progression or prediction of future cognitive decline.
The researchers concluded: “[D]igital handwriting assessment represents a low-cost, noninvasive, and ecologically valid method with strong potential for early screening and monitoring of cognitive decline.”
The study received partial funding from Fundação Ciência e Tecnologia, IP national support through CHRC. The study authors reported no commercial or financial conflicts of interest. The article also stated that generative artificial intelligence tools were used to improve clarity and readability without altering the original scientific content.
Source: Frontiers in Human Neuroscience
