An Efficient Approach Based on Medical Images for Parkinsons Disease Identification and Management

Abstract

A progressive neurodegenerative condition that profoundly affects motor and cognitive abilities is Parkinson\'s disease (PD). Timely action and improved patient outcomes depend on an early and correct diagnosis. Two different models are used in this study\'s deep learning-based method for Parkinson\'s disease detection: Long Short-Term Memory (LSTM) networks for structured patient data classification and Convolutional Neural Networks (CNN) for handwriting picture analysis. In order to uncover patterns of motor impairment, the CNN model uses spatial features extracted from handwriting images, and the LSTM model analyses sequential patient data to find PD-related distinctive trends. Handwriting samples from both healthy and Parkinson\'s patients are included in the dataset, as are organized medical records with motor and cognitive test results.To improve model performance, preprocessing methods such data standardization, normalization, and image scaling are used.

Introduction

Parkinson’s disease (PD) is a neurodegenerative disorder marked by motor symptoms like tremors, rigidity, and bradykinesia, alongside non-motor symptoms such as mood disorders and cognitive decline. Diagnosing PD early remains challenging due to the absence of reliable tests, often relying on clinical observations and patient history.

Recent advancements in artificial intelligence (AI) have shown promise in enhancing early PD detection. Deep learning techniques, particularly Convolutional Neural Networks (CNNs) and Long Short-Term Memory (LSTM) networks, have been effectively applied to analyze medical images and sequential clinical data. CNNs excel at extracting spatial features from MRI scans and handwriting images, while LSTMs are adept at identifying temporal patterns in structured patient data. Preprocessing methods like feature scaling, image resizing, and normalization are employed to improve model performance.

A proposed hybrid system combines CNN and LSTM models to diagnose PD by analyzing MRI scans, handwriting images, and structured patient data. The CNN model processes image data to detect spatial patterns indicative of neurodegeneration, while the LSTM model analyzes sequential clinical records to capture temporal relationships. The system undergoes training using cross-entropy loss and the Adam optimizer, with evaluation metrics including accuracy, precision, recall, and F1-score. To ensure robustness and generalization, the system is tested on diverse datasets.

Conclusion

This work proposes a deep learning approach for identifying Parkinson\'s disease using two distinct models: CNN for analyzing spiral and wave handwriting images and LSTM for evaluating sequential motor assessment data. The CNN model achieved an accuracy of 99.2% in recognizing spatial patterns in handwriting, while the LSTM model achieved 98% accuracy for temporal trends in clinical data. Both models demonstrated superiority over traditional classifiers, illustrating the potential of deep learning for the accurate and early diagnosis of Parkinson\'s disease.

References

[1] Sangeetha, S., Baskar, K., Kalaivaani, P.C.D., and Kumaravel, T.: Deep Learning-Based Early Parkinson’s Disease Detection from Brain MRI Image. Proceedings of the 7th International Conference on Intelligent Computing and Control Systems (ICICCS), IEEE, pp. 490–495, India (2023). [2] Kumar, K., and Ghosh, R.: Parkinson’s Disease Diagnosis Using Recurrent Neural Network Based Deep Learning Model by Analyzing Online Handwriting. Multimedia Tools and Applications, Springer, vol. 83, no. 4, pp. 11687–11715 (2024). [3] Govindu, A., and Palwe, S.: Early Detection of Parkinson\'s Disease Using Machine Learning. Procedia Computer Science, Elsevier, vol. 218, pp. 249–261 (2023). [4] Jovanovic, L., Damaševi?ius, R., Matic, R., et al.: Detecting Parkinson’s Disease from Shoe-Mounted Accelerometer Sensors Using CNNs Optimized with Metaheuristics. PeerJ Computer Science, vol. 10, p. e2031 (2024). [5] Vyas, T., Yadav, R., Solanki, C., et al.: Deep Learning-Based Scheme to Diagnose Parkinson\'s Disease. Expert Systems, Wiley, vol. 39, no. 3, p. e12739 (2022). [6] Sahu, L., Sharma, R., Sahu, I., et al.: Efficient Detection of Parkinson\'s Disease Using Deep Learning Techniques Over Medical Data. Expert Systems, Wiley, vol. 39, no. 3, p. e12787 (2022). [7] Dentamaro, V., Impedovo, D., Musti, L., et al.: Enhancing Early Parkinson’s Detection through Multimodal Deep Learning and Explainable AI. Scientific Reports, Nature, vol. 14, no. 1, p. 20941 (2024). [8] Shaban, M.: Deep Learning for Parkinson’s Disease Diagnosis: A Short Survey. Computers, MDPI, vol. 12, no. 3, p. 58 (2023). [9] Islam, M.A., Majumder, M.Z.H., Hussein, M.A., et al.: Review of Machine Learning and Deep Learning Algorithms for Parkinson\'s Detection Using Handwriting and Voice Datasets. Heliyon, Elsevier, vol. 10, no. 3 (2024). [10] Sayed, M.A., Tayaba, M., Islam, M.D., et al.: Parkinson\'s Disease Detection Through Vocal Biomarkers and Advanced Machine Learning Algorithms. arXiv preprint arXiv:2311.05435 (2023).

Copyright

Copyright © 2025 K. Lishalini, M. Jeyavani, P. Vidhya Saraswathi. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.