Detecting Parkinson's Disease at an Early Stage Through Machine Learning Analysis of Brain MRI Images

Parkinson's disease manifests as a condition characterized by uncoordinated movements, and its symptoms gradually develop over time. These symptoms primarily affect essential physiological activities such as walking, talking, and speaking. Additionally, emotional changes occur, leading to psychological challenges. In this research study, machine learning models, including Naïve Bayes, SVM, and KNN classifiers, were employed to classify brain MRI images from patients into “Normal” and “patients with Parkinson”. The dataset consisted of 60 brain MRI scans of normal and Parkinson. The outcome of the proposed work represents that the SVM model peaked its performance with the best accuracy at 98.3%, while Naïve Bayes and KNN achieved 80% and 92.5%, respectively. Ultimately, the SVM Classifier was identified as the most effective in classifying normal and Parkinson's brains in MRI images, demonstrating the potential of machine learning in mimicking human radiologists’ abilities to detect abnormalities in radiographs.