Early Detection of Parkinson's Disease Using AI Techniques and Image Analysis

Area of Science:

• Neuroscience
• Medical Imaging
• Artificial Intelligence

Background:

• Advancements in artificial intelligence (AI) and image processing offer significant benefits for Parkinson's disease (PD) diagnosis.
• This study investigates methods for processing hand-drawn Archimedean spirals to detect early signs of PD.

Purpose of the Study:

• To explore and evaluate AI-driven approaches for analyzing hand-drawn spirals for Parkinson's disease detection.
• To integrate the most effective approach into the NeuroPredict platform for neurodegenerative disease management.

Main Methods:

• Feature extraction techniques converting hand-drawn spirals into frequency spectra for analysis.
• Utilizing frequency-based features and drawing characteristics (thickness, pressure) as inputs for classification algorithms.
• Developing a method that does not require specialized hardware, using only pencil and paper.

Main Results:

• The selected AI approach achieved an overall accuracy of 95.24% in detecting signs of Parkinson's disease.
• The method successfully acquires new test data using conventional pencil and paper.
• No specialized devices like graphic tablets or digital pens are necessary for data acquisition.

Conclusions:

• AI significantly enhances diagnostic precision for neurodegenerative diseases like Parkinson's.
• The developed spiral analysis technique holds clinical relevance for improving early PD detection.
• The findings support the integration of AI tools in neurodegenerative disease management platforms.