An Effective Fusing Approach by Combining Connectivity Network Pattern and Temporal-Spatial Analysis for EEG-Based...

Area of Science:

• Neuroscience
• Biomedical Engineering
• Rehabilitation Medicine

Background:

• Motor-modality-based brain-computer interfaces (BCI) show promise for neural rehabilitation in stroke patients.
• Traditional temporal-spatial analysis is commonly used for pattern recognition in BCI tasks.
• There is a need for improved feature selection methods to enhance BCI performance for stroke rehabilitation.

Purpose of the Study:

• To introduce a novel connectivity network analysis for electroencephalography (EEG)-based feature selection.
• To evaluate the combination of temporal-spatial analysis and network analysis for improving BCI classification.
• To demonstrate the effectiveness of the proposed method for stroke patient rehabilitation.

Main Methods:

• Developed a novel connectivity network analysis for EEG feature selection.
• Extracted network features capturing spatial activities and inter-regional interactions during motor tasks.
• Combined temporal-spatial analysis with the proposed network analysis for feature selection.
• Evaluated the performance of the combined approach in BCI classification for stroke patients.

Main Results:

• The proposed connectivity network analysis captured spatial activities and mined interactive patterns among cerebral regions.
• The combination of temporal-spatial and network analysis improved BCI classification performance, achieving 81.7% accuracy.
• Classification accuracies were raised for the majority of patients (6 out of 7).

Conclusions:

• The novel connectivity network analysis is a meaningful advancement for EEG-based feature selection in BCI.
• The proposed method effectively enhances BCI classification performance for stroke rehabilitation.
• This approach holds potential for developing more effective BCI training programs for stroke patients.