A deep learning and metaheuristic optimization algorithm based on Parkinson's disease classification from MRI images

Area of Science:

• Neurology
• Medical Imaging
• Artificial Intelligence

Background:

• Parkinson's disease (PD) is a common neurodegenerative disorder with a diagnostic challenge due to late symptom manifestation.
• Early detection of PD is crucial for timely intervention and improved patient outcomes.
• Current diagnostic methods often lack the sensitivity for pre-symptomatic detection.

Purpose of the Study:

• To develop an advanced Artificial Intelligence (AI) and Deep Learning (DL) model for early classification of Parkinson's disease (PD) using MRI scans.
• To create a robust medical decision-support system to enhance diagnostic precision for PD.
• To improve prompt clinical intervention strategies through accurate early detection.

Main Methods:

• A modified EfficientNet Deep Learning (DL) model was employed.
• Reinforcement learning optimization was integrated for dynamic parameter adjustment.
• The model was trained to differentiate MRI scans of PD patients and healthy individuals, minimizing misclassification rates.

Main Results:

• The proposed model achieved 98% overall accuracy in classifying PD and healthy individuals.
• High precision, recall, and F1-score values were obtained for both classes.
• Specific performance metrics included 95% precision, 96% recall, and 98% F1-score for PD patients; and 93% precision, 97% recall, and 96% F1-score for healthy individuals.

Conclusions:

• The AI-driven EfficientNet model significantly enhances diagnostic performance for early PD detection compared to standard methods.
• The integration of DL and reinforcement learning offers a powerful approach for predictive analytics in neurology.
• This study highlights the transformative potential of AI in improving patient outcomes for neurodegenerative diseases like Parkinson's.