Fatma Latifoğlu,Sultan Penekli,Fırat Orhanbulucu,Muhammad E. H. Chowdhury
Abstract:Abstract Parkinson’s disease (PD) is the second most common neurological disorder caused by damage to dopaminergic neurons. Therefore, it is important to develop systems for early and automatic diagnosis of PD. For this purpose, a study that will contribute to the development of systems for the automatic diagnosis of PD is presented. The Electroencephalography (EEG) signals were decomposed into sub-bands using adaptive decomposition methods, such as empirical mode decomposition, variational mode decomposition, and Vold-Kalman order filtering (VKF). Various features were extracted from the sub-band decomposed signals, and the significant ones were determined by Chi-squared test. These important features were applied as input to support vector machine (SVM), fitch neural network (FNN), k-nearest neighbours (KNN), and decision trees (DT), machine learning (ML) models and classification was performed. We analysed the performance of ML models by obtaining accuracy, sensitivity, specificity, positive predictive value, negative predictive values, F1-score, false-positive rate, kappa statistics, and area under the curve. The classification process was performed for two cases: PD ON-HC and PD OFF-HC groups. The most successful method in this study was the VKF method, which was applied for the first time in this field with the approach specified for both cases. In both instances, the SVM algorithm was employed as the ML model, with classifier performance criterion values close to 100%. The results obtained in this study seem to be successful compared to the results of recent research on the diagnosis of PD.
What problem does this paper attempt to address?
The main problem that this paper attempts to solve is to develop a system for the early and automatic diagnosis of Parkinson's disease (PD). Parkinson's disease is a common neurodegenerative disease caused by damage to dopaminergic neurons and usually affects people over 50 years old. Patients usually present with motor symptoms, such as slight tremors, slow movement, gait or postural disorders, as well as non - motor symptoms, such as depression, loss of smell, constipation or sleep problems. It may be difficult to assess Parkinson's disease clinically based on these symptoms, so faster and more reliable methods are required to achieve early diagnosis and start treatment as soon as possible.
To achieve this goal, the paper proposes a new method to detect Parkinson's disease using Vold - Kalman order filtering (VKF) and machine - learning algorithms. Specifically, the researchers used three different signal - processing techniques (Empirical Mode Decomposition EMD, Variational Mode Decomposition VMD, and Vold - Kalman order filtering VKF) to decompose electroencephalogram (EEG) signals into sub - bands and extract features from them. The important features were determined by the chi - square test, and then these features were used as inputs for machine - learning models such as Support Vector Machines (SVM), Feed - forward Neural Networks (FNN), k - Nearest Neighbors (KNN), and Decision Trees (DT) for classification. The performance of different machine - learning models was analyzed, including accuracy, sensitivity, specificity, positive predictive value, negative predictive value, F1 - score, false positive rate, Kappa statistic, and Area Under the Curve (AUC).
The research results show that among all the tested machine - learning algorithms, the features extracted using the VKF method perform the best. Especially in two cases (PD ON - HC group and PD OFF - HC group), when the SVM algorithm is used as the machine - learning model, the performance indicators of the classifier are close to 100%. This indicates that the new method combining VKF and machine - learning techniques has high accuracy and reliability in the early diagnosis of Parkinson's disease, surpassing the existing diagnostic methods. In addition, this study also explored the evaluation of drug efficacy, further enriching the understanding and management of Parkinson's disease.