Society for Neuroscience Abstracts
J.A. Onton & S. Makeig, Swartz Center for Computational Neuroscience, Institute for Neural Computation, University of California San Diego, La Jolla, CA
EEG-biofeedback, also known as neurofeedback in clinical settings or brain-computer interface (BCI) operation in basic research, usually involves self-regulation of spectral power in a certain frequency range at a specified electrode, the intention being to gain control of an underlying EEG process with power in the frequency range of interest. However, single scalp electrodes receive signals from many EEG sources at once. Thus, gaining control of scalp-recorded activity does not imply control of any particular cortical EEG source. This is particularly important in the clinical setting where therapeutic benefits are often hypothesized to result from successful control of activity in specific cortical regions. The present study was designed to more precisely determine where control of cortical activity in the alpha (8-12 Hz) range was achieved and, further, what other spectral modulations were involved in successful control. We used on-line spatial filtering to extract, in real time, spectral power for a single independent component (IC) generated in right somato-motor cortex. This IC was found by running independent component analysis (ICA) on data collected just prior to ICA-based self-regulation. In each of three sessions, the subject performed a total of 48 min of alpha regulation, alternating between 3-min blocks that rewarded increases or decreases in alpha power relative to mean basline power, respectively. During post-hoc analysis, we decomposed second-by-second spectral power (3-30 Hz) across several ICs at once and computed 15 spectral modulation factors active during the task. When the subject successfully increased alpha power in the right somato-motor region, gamma power simultaneously decreased in several frontal and parietal ICs. The converse was true with successful decreases in alpha power. By the subject's subjective report, high-alpha (and low-gamma) was associated with a relaxed and relatively thought-free mental state, while low-alpha (and high-gamma) produced a feeling of enthusiastic attentiveness. Future self-regulation studies using multiple subjects, ICs, and frequency ranges should reveal relationships between subjective mental state, willed EEG self-control, and EEG source activities.