Neural correlates of learning in an electrocorticographic motor-imagery brain-computer interface

Abstract

Human subjects can learn to control a one-dimensional electrocorticographic (ECoG) brain-computer interface (BCI) using modulation of primary motor (M1) high-gamma activity (signal power in the 75–200 Hz range). However, the stability and dynamics of the signals over the course of new BCI skill acquisition have not been investigated. In this study, we report 3 characteristic periods in evolution of the high-gamma control signal during BCI training: initial, low task accuracy with corresponding low power modulation in the gamma spectrum, followed by a second period of improved task accuracy with increasing average power separation between activity and rest, and a final period of high task accuracy with stable (or decreasing) power separation and decreasing trial-to-trial variance. These findings may have implications in the design and implementation of BCI control algorithms.

Unsupervised machine-learning classification of electrophysiologically active electrodes during human cognitive task performance

“…Previous studies used power in the iEEG spectrum along with statistical tests to select active electrodes 7 , with an emphasis on the electrodes that showed increase in power 6,8 . Our method is able to automatically identify electrodes as active with higher induced power in the high-gamma band with no a priori knowledge or constraint.…”

“…We found that our metric of induced power in the high-gamma band showed the highest sensitivity and specificity performance in classifying active electrodes. Increased gamma and high-gamma band activity has been shown to play a role in visual tasks 47–49 , motor tasks 50,51 , and motor-imagery tasks 6,52 , and is thought to be a general biomarker of cognitive processing in the brain 53 . Therefore, metrics based on the gamma band power, including gamma consistency, would be useful for active electrode classification in a range of cognitive, motor, and perceptual tasks.…”

“…Also, those studies have mainly focused on finding active electrodes for task-related classification problems and have not addressed the physiological relevance of the identified electrodes, leading to manual or semiautomatic identification of active electrodes in iEEG 3,4,9,28 . Studies aimed at addressing the physiological relevance of identified electrodes with automatic electrode selection have typically used a combination of different statistical methods based on iEEG power or event-related potentials 6–8 .…”

Sensorimotor ECoG Signal Features for BCI Control: A Comparison Between People With Locked-In Syndrome and Able-Bodied Controls

“…Since ECoG allows one to capitalize on the detailed spatial organization of the sensorimotor cortex, multiple independent control signals may conceptually be extracted from this area using high-spatial-density ECoG grids (Slutzky et al, 2010; Branco et al, 2017). Importantly, research on implantable ECoG-based BCIs often focuses on movement-related increases in High-Frequency Band (HFB, >30 Hz) power (Crone et al, 1998; Chestek et al, 2013; Blakely et al, 2014; Branco et al, 2017), which are thought to reflect local processing and are considered to be more spatially focal than changes in LFB power (Miller et al, 2009; Hermes et al, 2012), potentially providing a more specific, and therefore more reliable, BCI control signal (Schalk and Leuthardt, 2011). Yet, it has also been shown that changes in LFB power may contribute to accurate decoding and reliable ECoG-BCI control (Schalk et al, 2007; Nakanishi et al, 2014; Vansteensel et al, 2016; Flint et al, 2017).…”

Effects of Light and Sound on the Prefrontal Cortex Activation and Emotional Function: A Functional Near-Infrared Spectroscopy Study

“…Initially, the BCI techniques were developed to assist handicapped and paralyzed individuals to act independently and communicate successfully with others (Wolpaw et al, 2002 ; Dornhege et al, 2007 ). However, at present, patients require special training to be able to handle BCI (Blakely et al, 2014 ; Simon et al, 2015 ), which might prove nearly impossible for some of the disabled patients. Recently, near infrared spectroscopy (NIRS) has attracted attention for its use as a basic instrument for BCI techniques (Chaudhary et al, 2015 ; Von Lühmann et al, 2015 ; Shin et al, 2016 ), because of its moderate temporal resolution, noninvasive continuous measurement facility, compactness, and low interference with other techniques such as EEG and fMRI.…”