Individualized Electrode Subset Improves the Calibration Accuracy of an EEG P300-design Brain-Computer Interface for People with Severe Cerebral Palsy

Abstract Objective This study examined the effect of individualized electroencephalogram (EEG) electrode location selection for non-invasive P300-design brain-computer interfaces (BCIs) in people with varying severity of cerebral palsy (CP). Approach A forward selection algorithm was used to select the best performing 8 electrodes (of an available 32) to construct an individualized electrode subset for each participant. BCI accuracy of the individualized subset was compared to accuracy of a widely used default subset. Main Results Electrode selection significantly improved BCI calibration accuracy for the group with severe CP. Significant group effect was not found for the group of typically developing controls and the group with mild CP. However, several individuals with mild CP showed improved performance. Using the individualized electrode subsets, there was no significant difference in accuracy between calibration and evaluation data in the mild CP group, but there was a reduction in accuracy from calibration to evaluation in controls. Significance The findings suggested that electrode selection can accommodate developmental neurological impairments in people with severe CP, while the default electrode locations are sufficient for many people with milder impairments from CP and typically developing individuals.