The Role of Conversational AI in Early Diagnosis of Parkinson’s Disease
The Role of Conversational AI in Early Diagnosis of Parkinson’s Disease
Mohammadesmail Nikfar1
1) Ph.D. Graduate in Mathematics, Doctor of Philosophy (Ph.D.) in Pure Mathematics, Mathematics Teacher, District 4 Department of Education, Tehran, Iran
IOI Code
:
ADAD-ACADDB
Publication :
The 20th International Conference on Health, Treatment, and Health Promotion, Madrid, Spain(https://hwconf.ir/20th)
Abstract :
Parkinson’s disease is a neurological disorder that affects motor control, leading to symptoms such as tremors and stiffness. Early diagnosis is essential for effective treatment; however, traditional methods are often time-consuming and costly. This study employs artificial intelligence and machine learning techniques, combined with voice analysis, to detect the early signs of the disease. We developed a hybrid model—integrating convolutional neural networks, recurrent neural networks, multiple kernel learning, and a multilayer perceptron—and applied it to a dataset consisting of 81 voice recordings. Acoustic features, including Mel-frequency cepstral coefficients (MFCCs), jitter, and shimmer, were analyzed. The model achieved a recall of 91.11%, precision of 92.5%, an F1-score of 91.13%, and an Area Under the Curve (AUC) of 0.9125. Furthermore, SHAP (SHapley Additive exPlanations) values were used to provide model explainability, identifying key features that influence the detection of Parkinson’s disease, thereby enhancing the interpretability and trustworthiness of the AI system. Additionally, a probability-based scoring system was developed, enabling patients and physicians to monitor the progression of the disease. This AI-driven approach offers a non-invasive, cost-effective, and rapid tool for the early diagnosis of Parkinson’s disease, facilitating personalized treatment through the use of vocal biomarkers.
Keywords : parkinson's disease deep learning acoustic biomarkers artificial intelligence