Graphomotor performance as a quantitative marker of Alzheimer's disease: a tablet-based case-control ROC study

From the clinical and neurobiological perspective of cortico subcortical loop concepts, Alzheimer's disease (AD) is associated with progressive deterioration of cognitive, emotional, and motor functions, which can be assessed subclinically using digital motion tracking technology in an electromagnetic digitizing tablet technology. Graphomotor analysis provides a source of sensitive digital biomarkers with potential applications in the early diagnosis and monitoring of AD.

A case-control study was conducted including 72 patients with clinically diagnosed Alzheimer's disease (AD) and 44 cognitively healthy controls. Participants performed three graphomotor tasks (signature, free drawing, and the Trail Making Test A/B) using a digitizing tablet. Temporal, kinematic, and pressure-related parameters were recorded, including execution time, path length, mean and maximum velocity, pressure variability on the surface, and total harmonic distortion indices. Temporal measures reflected the duration and timing structure of task execution, kinematic measures described movement speed and variability, and pressure-related measures quantified the force applied by the pen and its stability during graphomotor performance. Statistical analysis was performed using the Mann-Whitney U test, and diagnostic accuracy was evaluated with receiver operating characteristic (ROC) analyses.

Patients with Alzheimer's disease exhibited significantly longer execution times, reduced mean and maximum velocities, and greater variability of velocity and pressure compared with healthy controls. In the signature task, kinematic measures showed the strongest discriminatory power, whereas in the drawing task, pressure variability emerged as the most sensitive parameter. In the Trail Making Test A, patients with AD performed significantly worse than controls across temporal, kinematic, and pressure domains, and the majority were unable to complete part B. ROC analyses demonstrated that execution time in TMT A (AUC 0.79; 95% CI 0.71-0.86), maximum velocity in TMT B (0.77; 0.68-0.85), and pressure variability (0.78; 0.69-0.85) achieved the highest diagnostic discriminatory value.

Digital analysis of biomechanical signals from graphomotor tasks indicates significantly reduced motor control and diminished visuospatial-executive functioning in individuals with Alzheimer's disease (AD) compared with healthy controls. Patients with AD demonstrated longer task execution times, decreased movement velocity, and greater variability of pressure relative to the control group. These findings suggest that tablet-derived graphomotor parameters may serve as candidate digital markers of motor and executive dysfunction in Alzheimer's disease. Further external validation and longitudinal studies are needed before their use as diagnostic biomarkers can be recommended.