Related Experiment Videos
Joint regression and classification via relational regularization for Parkinson's disease diagnosis
Haijun Lei1, Zhongwei Huang1, Tao Han1
1Guangdong Province Key Laboratory of Popular High Performance Computers, Key Laboratory of Service Computing and Applications, College of Computer Science and Software Engineering, Shenzhen University, Shenzhen, Guangdong, China.
Summary
Early diagnosis of Parkinson's disease (PD) is crucial. This study introduces a novel multi-modal neuroimaging approach for accurate PD diagnosis and symptom prediction, outperforming existing methods.
Area of Science:
- Neuroscience
- Medical Imaging
- Machine Learning
Background:
- Parkinson's disease (PD) is a progressive neurodegenerative disorder.
- Early and accurate diagnosis of PD is vital for patient management and treatment.
- Current diagnostic methods may lack sensitivity in early stages.
Purpose of the Study:
- To develop a joint regression and classification scheme for PD diagnosis using multi-modal neuroimaging data.
- To introduce a novel feature selection method integrating relational learning.
- To improve the accuracy of clinical score prediction and PD classification.
Main Methods:
- A unified multi-task feature selection model incorporating feature, response, and subject relationships.
- Exploitation of five regression variables: depression, sleep, olfaction, cognition scores, and a clinical label.
- Application of the method to baseline multi-modal neuroimaging data from the Parkinson's Progression Markers Initiative (PPMI) dataset.
Main Results:
- Multi-modal data significantly enhances classification performance compared to single-modal data.
- The proposed method improves clinical score prediction accuracy.
- The approach outperforms state-of-the-art methods in PD diagnosis and prediction.
- Identified brain regions offer potential for further medical analysis.
Conclusions:
- The proposed joint regression and classification scheme is effective for PD diagnosis.
- Relational learning-based feature selection enhances diagnostic accuracy.
- Multi-modal neuroimaging data provides a robust foundation for PD assessment.