Deep Learning-based Prediction of Parkinson's Disease Progression using Voice Recordings

The main underlying factor behind Parkinson's Disease (PD) is the degeneration of brain cells responsible for producing dopamine, a neurotransmitter crucial for interconnecting brain cells. These dopamine-producing cells play a vital role in controlling adaptability, regulating movement, and facilitating fluency. The concept of active aging has emerged to enhance various aspects of fitness as individuals age, ultimately improving their overall quality of life. However, most efforts have concentrated on promoting normal aging, with limited attention given to the elderly population afflicted with chronic conditions like PD. This research focuses on the application of deep learning techniques for early detection and diagnosis of Parkinson's disease (PD). In the first experiment, this study utilizes the voice samples from healthy individuals and PD patients as a minimum training set for the Enhanced Deep Neural Network Model (EDNN) classifier, resulting in an optimal prediction model capable of identifying early onset PD. Subsequently, this study also investigates the effectiveness of transfer learning using the Convolutional Neural Network (CNN) architecture Alex Net for diagnosing Parkinson's disease from MR images in a second experiment. The results demonstrate the successful implementation of the EDNN classifier, showcasing its potential for early detection of Parkinson's disease, allowing for timely intervention and treatment. Moreover, by employing transfer learning with AlexNet, this study presents a novel approach for accurately diagnosing Parkinson's disease from MR images, representing a significant advancement in medical image analysis.