Non-invasive Detection of Cortical Control Signals for Brain-Computer Interfaces.

摘要

In the sensorimotor areas of the cortex, motor execution and imagery have been found to be associated with increases in the high gamma (HG) frequency band, as well as decreases in the beta, and mu frequency bands. Because these cortical rhythms can be intentionally modulated by motor imagery or execution, they have often been used in brain-computer interface (BCI) studies as a control signal. However, the non-invasive recording of the modulation of the HG frequency band is still a relatively novel concept and most HG studies have been limited to invasive recording methods, due to the low signal-to-noise ratio of the neuronal ensemble activity at these frequencies. Consequently, the systematic study of HG effects has been limited to narrow populations (typically epilepsy patients) and cortical locations, which are usually determined by clinical needs rather than by the experimental paradigm one wishes to study. This thesis serves as a proof-of-concept that the HG frequency band can be studied non-invasively, thus opening up this band to a much wider subject population and range of experimental paradigms. In addition, this thesis provides evidence that an implanted subdermal recording electrode system may provide a reliable, long-term, portable method for recording these motor related control signals for BCI control.