Brain-Computer Interfaces for patients with Amyotrophic Lateral Sclerosis

Abstract

Electroencephalographic (EEG) brain-Computer Interfaces (BCIs) hold promise to restore communication with completely locked-in (CLIS) patients with Amy- otrophic Lateral Sclerosis (ALS). However, these patients cannot use existing EEG- based BCIs, possibly because such systems rely on brain processes that are im- paired in ALS. We propose to use for BCI for ALS patients high cognitive processes connected to consciousness, because ALS patients should be able to use such BCI as long as they are fully conscious. We introduce a BCI based on neurofeedback from precuneus, brain area linked to consciousness. We describe two cases of successful use of the BCI by ALS patients, with stable online performance over the course of disease progression. Additionally, we show that training time can be improved by replacing the neurofeedback with direct instructions, contrasting self-referential and neutral thoughts. We further investigate self-referential think- ing in ALS and find differences in the EEG correlates of self-referential thinking between ALS and healthy controls. This finding raises the question of awareness and consciousness in CLIS ALS. We propose a method that may serve as basis for consciousness detection in CLIS ALS patients: EEG-based identification of the Default Mode Network (DMN), brain resting-state network closely linked to consciousness.