Accurate single-trial detection of movement intention made possible using adaptive wavelet transform

Abstract

The outlook of brain-computer interfacing (BCI) is very bright. The real-time, accurate detection of a motor movement task is critical in BCI systems. The poor signal-to-noise-ratio (SNR) of EEG signals and the ambiguity of noise generator sources in brain renders this task quite challenging. In this paper, we demonstrate a novel algorithm for precise detection of the onset of a motor movement through identification of event-related-desynchronization (ERD) patterns. Using an adaptive matched filter technique implemented based on an optimized continues Wavelet transform by selecting an appropriate basis, we can detect single-trial ERDs. Moreover, we use a maximum-likelihood (ML), electrooculography (EOG) artifact removal method to remove eye-related artifacts to significantly improve the detection performance. We have applied this technique to our locally recorded Emotiv(®) data set of 6 healthy subjects, where an average detection selectivity of 85 ± 6% and sensitivity of 88 ± 7.7% is achieved with a temporal precision in the range of -1250 to 367 ms in onset detections of single-trials.

Recognizing Tonal and Nontonal Mandarin Sentences for EEG-Based Brain–Computer Interface

“…This study aims to understand the neural processing before and during speaking. We investigated the brain activities during the BS (baseline), Before Speaking, and Speaking a sentence in [72], and with the aid of dimensionality reduction, our proposed BCI method can be used for accurate and real-time classification of the tonal versus nontonal language. This study will have the following potential impacts.…”