Abstract
Dementia impairs the ability to perform everyday activities. Reduced motor capacity and executive functions as well as loss of memory function and forms of apraxia and action disorganization syndrome can be reasons for such impairments. In this study, an analysis of the hand trajectories during the sequential movements in an adapted version of the trail making task, the reciprocal trail making task (RTMT), was used to predict performance in activities of daily living (ADL) of patients suffering from mild cognitive impairment and dementia. 1 patient with dementia of the Alzheimer's type and 15 healthy, age-matched adults were tested in the standardized ADL of tea making and document filing. The characteristics of the kinematic performance in the RTMT were assessed, and models of multiple linear regression were computed to predict the durations of the ADL. Patients showed increased trial durations (TDs) in the ADL (Cohen's d: tea making 1.64, document filing 1.25). Parameters and explained variability differed across patients and control as well as between different activities. The models for the patient sample were stronger and particularly high for the document filing task for which kinematics explained 71% of the variance ([Formula: see text]: tea making 0.62, document filing 0.71; both tasks combined patients 0.55, controls 0.25). The most relevant factors for the models were the TD and a parameter characterizing movement fluency and variability ("movement harmonicity") in the RTMT. The models of multiple linear regression suggested that the patients' activity of daily living performance was limited by cognitive demands, namely, identifying the varying targets during sequencing and the healthy controls' performance by their motor capacity. Such models could be used to estimate the severity of ADL impairments in patients.