Towards cognitive brain-computer interfaces for patients with amyotrophic lateral sclerosis

Abstract

Brain-Computer Interfaces (BCIs) often rely on low-level cognitive processes known to be impaired in late stages of amyotrophic lateral sclerosis (ALS). We propose a BCI for ALS patients based on self-regulation of neuronal oscillations in the superior parietal lobule, which is less affected by ALS than motor and sensory cortices. We describe a case of self-regulation of band power in gamma range (55-85 Hz) based on feedback from the parietal cortex by an ALS patient, resulting in a mean offline two-class decoding accuracy of 79.2% across four sessions. Despite a good offline decoding accuracy, a source localisation analysis revealed that gamma-power modulation was not spatially localized, suggesting confounding by non-cortical artifacts. Theta-power in contrast, showed a strong localized response in the precuneus. As such, this may be an alternative possibility of using self-regulation of neuronal oscillations for cognitive BCI.