The utility of handwriting analysis and selected neuropsychological tests in the diagnosis of mild cognitive impairment in older patients

Neuropsychological assessment in mild cognitive impairment (MCI) increasingly includes executive functions evaluation to improve diagnostic accuracy. Handwriting analysis, though common in dementia studies, is less explored in MCI. This single-centre study aimed to compare neuropsychological tests and handwriting parameters, assessing their individual diagnostic value. The study included two groups: MCI (n = 46, female/male ratio 41/5, mean age 76.87 ± 5.08) and controls without cognitive impairment (n = 46, ratio 42/4, mean age 75.70 ± 5.97). The assessment included MoCA, MMSE, Comprehensive Trail Making Test (CTMT), verbal fluency test and handwriting analysis using Livescribe Echo Smartpen. Logistic Regression (LR), K-Nearest Neighbours (KNN) and Linear Discriminant Analysis (LDA) models were used to identify patients with MCI. Patients with MCI performed worse on neuropsychological tests, generating fewer words in verbal fluency (p < .01) and taking longer on CTMT (p < .01). Neuropsychological tests outperformed handwriting measures in MCI classification (AUC: CTMT = .81, semantic fluency = .76, phonemic fluency = .72). Among the handwriting measures, text height (AUC = .68) showed the best performance, while other kinematic features ranged from .63 to .64. After combining all neuropsychological tests, KNN achieved the best classification of MCI (AUC = .84, ACC = .82, MCC = .63), while handwriting-based models performed worse, with LR reaching the highest AUC (.64), ACC (.62) and MCC (.23). CTMT and verbal fluency tests are useful in diagnosing MCI, while handwriting measures showed limited classification value.