An Operative Expectation of Parkinson’s Ailment Using a Hybrid Machine Learning and Artificial Intelligence Systems

• Published in: SN computer science Vol. 5; no. 8; p. 1146
• Main Authors: Athithan, Senthil, Sachi, Savya, Singh, Ajay Kumar
• Format: Journal Article
• Language: English
• Published: Singapore Springer Nature Singapore 04.12.2024; Springer Nature B.V
• Subjects: Accuracy; Algorithms; Amyotrophic lateral sclerosis; Artificial intelligence; Artificial neural networks; Brain research; Computer Imaging; Computer Science; Computer Systems Organization and Communication Networks; Data Structures and Information Theory; Datasets; Decision trees; Disease; Dopamine; Electroencephalography; Electromyography; Energy consumption; Error analysis; Graphical user interface; Information Systems and Communication Service; Machine learning; Neurons; Original Research; Parkinson's disease; Pattern Recognition and Graphics; Research Advancements in Intelligent Computing; Software Engineering/Programming and Operating Systems; Spinal cord; Vision; Temporal Brain; Parkinson’s Disease; Neurological Disorder; Frontal Brain; EEG
• ISSN: 2661-8907; 2662-995X; 2661-8907
• DOI: 10.1007/s42979-024-03404-0

Human biosignals can be used to detect early signs of Parkinson’s disease (PD). Electroencephalography (EEG) measures brain signals, while electromyography (EMG) measures muscle signals, both of which are useful in studying functional and neuronal changes in Parkinson’s patients. Researchers extracted various time and frequency domain features from EEG and EMG signals of early-stage PD patients and control subjects (CS) to identify motor symptoms. A diagnostic tool based on a mathematical model utilizing an artificial neural network (ANN) and a graphical user interface was developed. The ANN-based classifier, using the EEG and EMG features as inputs, accurately identified early-stage PD patients from control subjects with a classification rate of approximately 100%. The combination of EEG and EMG characteristics resulted in the highest recognition rate of 99.1%. This model demonstrated near-perfect accuracy, with an R-value of 0.9976, indicating its precision. Additionally, it achieved a low validation mean square error of 0.0068 and an elapsed time of 1.32 s. Studies show that combining EEG and EMG features as input parameters effectively distinguishes PD patients from control subjects.