An Improved Method for Diagnosis of Parkinson's Disease using Deep Learning Models Enhanced with Metaheuristic...

Area of Science:

• Neurology
• Artificial Intelligence
• Medical Imaging

Background:

• Early diagnosis of Parkinson's disease (PD) remains a clinical challenge due to its slow progression.
• Machine learning and deep learning models are increasingly utilized for PD detection.

Approach:

• This study introduces five metaheuristic-enhanced deep learning models for early PD detection.
• Grey Wolf Optimization (GWO) was employed to fine-tune model hyperparameters for improved performance.
• The models were evaluated on T1, T2-weighted and SPECT DaTscan datasets.

Key Points:

• The proposed models, including GWO-VGG16, GWO-DenseNet, GWO-DenseNet + LSTM, GWO-InceptionV3, and GWO-VGG16 + InceptionV3, demonstrated high efficacy.
• All models achieved accuracy rates exceeding 99%.
• GWO-VGG16 + InceptionV3 and GWO-DenseNet achieved an AUC-ROC score of 99.99% on the T1, T2-weighted dataset.

Conclusions:

• Optimized deep learning models show significant promise for accurate and early Parkinson's disease detection.
• The GWO-enhanced models achieved near-perfect performance on standard medical imaging datasets.
• This approach offers a potential advancement in diagnosing PD, facilitating earlier intervention.