Addressing low frequency movement artifacts in EEG signal recorded during center-out reaching tasks

The successful application of noninvasive brain-computer interfaces (BCI) to neurological rehabilitation requires examination of low frequency movement artifacts and development of accurate new methods for their correction. To this end, this study applies an adaptive trend extraction method to electroencephalogram (EEG) signals recorded during active and passive center-out reaching tasks. Distinct patterns are discovered, which correlate to arm kinematics, but are shown to be largely artifactual in nature. Notably, these patterns are found to be similar to features currently used for discrimination of movement direction, indicating a necessity for caution and precise signal processing methods when utilizing low frequency content of EEG signals in such applications.