Prodromal Diagnosis of Lewy Body Diseases Based on the Assessment of Graphomotor and Handwriting Difficulties

TL;DR

This study develops a non-invasive, computerized method to identify prodromal Lewy body diseases by analyzing handwriting and graphomotor tasks, achieving 74% accuracy in distinguishing affected individuals from healthy controls.

Contribution

The paper introduces a novel quantitative approach combining multiple handwriting and graphomotor features to detect early signs of Lewy body diseases.

Findings

Achieved 74% accuracy in identifying prodromal LBDs.

Demonstrated the potential of objective, non-invasive diagnosis methods.

Validated the effectiveness of combined task analysis for early detection.

Abstract

To this date, studies focusing on the prodromal diagnosis of Lewy body diseases (LBDs) based on quantitative analysis of graphomotor and handwriting difficulties are missing. In this work, we enrolled 18 subjects diagnosed with possible or probable mild cognitive impairment with Lewy bodies (MCI-LB), 7 subjects having more than 50% probability of developing Parkinson's disease (PD), 21 subjects with both possible/probable MCI-LB and probability of PD > 50%, and 37 age- and gender-matched healthy controls (HC). Each participant performed three tasks: Archimedean spiral drawing (to quantify graphomotor difficulties), sentence writing task (to quantify handwriting difficulties), and pentagon copying test (to quantify cognitive decline). Next, we parameterized the acquired data by various temporal, kinematic, dynamic, spatial, and task-specific features. And finally, we trained classification models for each task separately as well as a model for their combination to estimate the predictive power of the features for the identification of LBDs. Using this approach we were able to identify prodromal LBDs with 74% accuracy and showed the promising potential of computerized objective and non-invasive diagnosis of LBDs based on the assessment of graphomotor and handwriting difficulties.