Objective and automatic classification of Parkinson disease with Leap Motion controller

Area of Science:

• Biomedical Engineering
• Neurology
• Rehabilitation Technology

Background:

• Parkinson disease (PD) significantly impacts motor function, necessitating objective assessment tools.
• The Movement Disorder Society-Unified Parkinson's Disease Rating Scale (MDS-UPDRS) part III is a standard clinical assessment.
• Objective quantification of motor dysfunction in PD is crucial for treatment monitoring.

Purpose of the Study:

• To evaluate the Leap Motion Controller's (LMC) capability in assessing motor dysfunction in Parkinson disease patients.
• To compare LMC-derived movement data against clinical MDS-UPDRS III ratings.
• To determine the feasibility of using LMC for objective PD motor assessment.

Main Methods:

• 16 patients with PD and 12 healthy controls participated.
• Standardized hand exercises (finger tapping, hand opening/closing, pronation/supination, postural tremor) were performed.
• Time and frequency domain features were extracted from LMC data for analysis.

Main Results:

• Machine learning models (Logistic Regression, Naive Bayes, SVM) were employed for classification.
• Naive Bayes achieved the highest accuracy (81.4%) in distinguishing PD patients from controls.
• LMC data showed statistical and clinical significance for finger tapping and hand opening/closing exercises.

Conclusions:

• The LMC demonstrated limited utility for assessing postural tremor and forearm pronation/supination in PD.
• The LMC shows potential for objective assessment of finger tapping and hand opening/closing in PD.
• Further validation with updated software is recommended for broader clinical application.