Translating Fine-Motor Spiral Drawing Metrics into Accessible Digital Interfaces for Parkinson’s Disease

Abstract

Parkinson’s Disease (PD) is a neurodegenerative disorder characterized by motor impairments, including resting tremors, sluggish movement, and reduced motor control. Consequently, individuals with PD frequently encounter significant barriers when interacting with standard digital user interfaces (UI) and user experiences (UX). While existing research uses wearable technology for early PD diagnosis, comparatively few studies focus on optimizing digital accessibility for these tools for affected users. To address this gap, this study analyzed kinetic and kinematic motor features from the UCI Parkinson’s Disease Spiral Drawings dataset, which captures spatial coordinates (X, Y, Z), pen pressure, grip angle, and drawing time from 62 PD patients and 15 control subjects. Data preprocessing and time-series feature extraction were conducted using Python libraries including pandas, numpy, scipy, and tsfresh. Additionally, machine learning classification models were constructed using scikit-learn to identify key motor signatures. A deeper analysis found that the PD group exhibited higher pause counts and frequencies compared to the control group. Additionally, the PD group displayed more unstable movements when using peripherals in both the speed at which they drew the spirals and in the amount of pressure they placed on it. These differences were found to be statistically significant through feature selection. Although an analysis of individual tremor metrics lacked statistical significance, analyzing patterns across multi-feature interactions revealed that movement jerkiness and pressure are key indicators for designing adaptive digital interfaces. These findings establish a framework for developing digital interfaces that respond to motor fluctuations, enhancing digital accessibility for individuals with PD.