The performance impact of data augmentation in CSP-based motor-imagery systems for BCI applications. (September 2020)

• Record Type: Journal Article
• Authors: Gubert, Paulo Henrique; Costa, Márcio Holsbach; Silva, Cleison Daniel; Trofino-Neto, Alexande
• Is Part Of: Biomedical signal processing and control. Volume 62(2020)
• Publication Date: 2020-09
• Subjects: Brain-computer interface -- machine learning -- BCI -- CSP -- LDA; Signal processing -- Periodicals; Biomedical engineering -- Periodicals; Signal Processing, Computer-Assisted -- Periodicals; Image Processing, Computer-Assisted -- Periodicals; Biomedical Engineering -- Periodicals; 610.28
• DOI: 10.1016/j.bspc.2020.102152
• Languages: English
• ISSNs: 1746-8094

Abstract

Highlights: Time-delay data-augmentation in motor-imagery BCI systems is investigated. The performance of data-augmented CSP-based methods are analyzed and compared. Simulation with signals from 2 public datasets (left and right-hand) were performed. Results show classification improvement of 5%, and 31.7% for a single individual. The time-delay augmentation improves classification accuracy of CSP-based methods. Abstract: Objective: This work investigates the performance impact of time-delay data-augmentation in motor-imagery classifiers for brain-computer-interface (BCI) applications. Methods: The considered strategy is an extension of the Common Spectral Spatial Patterns (CSSP) method, which consists of accommodating available additional information about intra- and inter-electrode correlation into the information matrix employed by the conventional Common Spatial Patterns (CSP) method. Results: Simulation with electroencephalographic signals obtained from two public motor-imagery datasets, in a context of differentiation between binary (left and right-hand) movements, result in an overall classification improvement of 5%, and up to 31.7% for a single individual, when compared to the use of the conventional correlation matrix in CSP-based methods. These results are supported through statistical inference and the analysis of the Receiving Operating Characteristic curve. Conclusion: The results obtained indicate that intra and inter-electrode correlation have relevant information, which have been underestimated by the literature in motor-imagery based BCI applications. Significance: The analyzed time-delay data-augmentation method improves the motor-imagery BCI classification accuracy with a foreseeable increase in the computation complexity.