Early Prediction of Parkinson’s Disease from Brain MRI Images Using Convolutional Neural Network

Abstract

The early diagnosis of Parkinson’s Disease (PD) is a challenging practice for doctors. Currently, there are no separate diagnostics and tests to be done to predict onset PD. However, the PD can be predicted through repeated clinical trials and tests. Sometimes, early prediction of PD can become tedious based on trials and tests. The computer-aided prediction will help medical professionals predict PD accurately during one’s onset stages to improve the PD patients’ quality of life. Hence, early prediction of PD is essential. In this article, Convolution Neural Networks (CNN) is proposed to classify PD patients and healthy individuals. The brain MRI images are given as input for the proposed methodology. The CNN deep neural network will first extract the features from the images. Then, it will classify the PD patients and healthy individuals from the extracted features. The automatic feature extraction will improve the accuracy of the classifier and reduce human error. The brain MRI images are taken from the PPMI dataset for experimentation. The sensitivity, specificity, and accuracy are calculated to assess the performance of the proposed methodology. The loss is also calculated to verify the performance of the classifier. It is observed that the CNN classifier has produced a higher accuracy of more than 98% in classifying PD patients and healthy individuals when compared to multi-layer perceptron deep learning.

Classification of Parkinson's Disease in Brain Mri Images Using Deep Residual Convolutional Neural Network

“…Over 90 genes have been linked to a significant hereditary component of disease risk. Additionally, large populations have shown that some potentially modifiable environmental (such as water pollutants, and pesticides) and other factors (such as coffee, smoking, head trauma, or exercise) have a role in the development of Parkinson's disease (Luo et al, 2022;Prema Arokia Mary et al, 2021).…”