Multitype hand writing as a digital marker for Parkinson's disease

Area of Science:

• Neurology
• Biomedical Engineering
• Medical Imaging

Background:

• Handwriting impairment is a key symptom in movement disorders (MDs), including Parkinson's disease (PD) and Parkinsonian syndrome (PDS).
• Early and accurate diagnosis of MDs is crucial for effective patient management.

Purpose of the Study:

• To develop and validate custom-designed digital handwriting (DHW) tools for enhanced diagnosis of MDs.
• To assess the correlation between DHW performance and VMAT2 expression measured by PET imaging.

Main Methods:

• Recruited participants with MDs and healthy controls for DHW tasks (drawing, writing in Chinese, English, Korean).
• Utilized AV133 PET-CT/MRI to quantify vesicular monoamine transporter 2 (VMAT2) expression.
• Developed an attention-based 1D-CNN model to analyze DHW indicators for PD diagnosis.

Main Results:

• The DHW model achieved excellent discrimination between MDs and controls (AUC=0.982).
• Specific DHW task combinations differentiated PD from PDS (AUC=0.727 and 0.717).
• High diagnostic agreement (85.45%) was found between DHW and PET-VMAT2 results; DHW scores correlated with VMAT2 expression in the caudate nucleus.

Conclusions:

• Custom-designed DHW tasks serve as effective digital markers for diagnosing Parkinson's disease.
• DHW offers a promising, non-invasive approach for MD diagnosis and monitoring.