Real-time EEG classification via coresets for BCI applications. (March 2020)

• Record Type: Journal Article
• Authors: Netzer, Eitan; Frid, Alex; Feldman, Dan
• Is Part Of: Engineering applications of artificial intelligence. Volume 89(2020)
• Publication Date: 2020-03
• Subjects: Machine learning -- Coreset -- Data structures -- On-line learning -- Electroencephalogram (EEG) -- Brain computer interface (BCI); Engineering -- Data processing -- Periodicals; Artificial intelligence -- Periodicals; Expert systems (Computer science) -- Periodicals; Ingénierie -- Informatique -- Périodiques; Intelligence artificielle -- Périodiques; Systèmes experts (Informatique) -- Périodiques; Artificial intelligence; Engineering -- Data processing; Expert systems (Computer science); Periodicals; 620.00285
• DOI: 10.1016/j.engappai.2019.103455
• Languages: English
• ISSNs: 0952-1976

Abstract

Abstract: A brain-computer interface (BCI) based on the motor imagery (MI) paradigm translates a subject's motor intention into a control signal by classifying the electroencephalogram (EEG) signals of different tasks. However, most existing systems use either (i) a high-quality algorithm to train the data off-line and run only the classification in real-time since the off-line algorithm is too slow, or (ii) low-quality heuristics that are sufficiently fast for real-time training but introduce relatively large classification error. In this work, we propose a novel processing pipeline that allows real-time and parallel learning of EEG signals using high-quality but potentially inefficient algorithms. This is done by forging a link between BCI and coresets, a technique that originated in computational geometry for handling streaming data via data summarization. We suggest an algorithm that maintains the representation of such coresets tailored to handle the EEG signal which enables (i) real-time and continuous computation of the common spatial pattern (CSP) feature extraction method on a coreset representation of the signal (instead of the signal itself), (ii) improvement of CSP algorithm efficiency with provable guarantees by applying the CSP algorithm on the coreset, and (iii) real-time addition of the data trials (EEG data windows) to the coreset. For simplicity, we focus on the CSP algorithm, which is a classic algorithm. Nevertheless, we expect that our coreset will be extended to other algorithms in future papers. In the experimental results, we show that our system can indeed learn EEG signals in real-time in, for example, a 64-channel setup with hundreds of time samples per second.