All Roads Lead to Rome: Resting-State Criticality as a Neurophysiological Marker of Brain-Computer Interface (BCI) Performance in People with Amyotrophic Lateral Sclerosis (ALS) and Healthy Participants.

This exploratory study investigates the relationship between resting-state brain dynamics (EEG) and subsequent brain-computer interface (BCI) performance in people with amyotrophic lateral sclerosis (ALS) and healthy participants. Grounded in the critical brain hypothesis, we examined whether neurophysiological markers of criticality in resting-state EEG correlate with subsequent BCI performance. We analyzed neuronal avalanche dynamics, multiscale criticality measures, the spectral slope, and Lempel-Ziv complexity across two datasets: a longitudinal single-case study of a person in the locked-in state after a disease history with ALS (Dataset 1) and longitudinal data from 5 people with ALS and single sessions from 6 healthy participants (Dataset 2). Neuronal avalanche dynamics, the spectral slope, and Lempel-Ziv complexity consistently correlated with subsequent BCI performance in people with ALS. Conversely, healthy participants showed inverted correlation patterns for the neuronal avalanche dynamics. While group comparisons were not statistically significant, effect size analyses indicated supercritical brain dynamics in people with ALS. We integrate these findings into the B-ARC (BCI-Aptitude through Resting-State Criticality) framework proposing an inverted U-shaped relationship between the proximity of resting-state brain activity to criticality and BCI control. In this framework, healthy participants approach the critical optimum from a subcritical regime, while people with ALS appear to approach it from a putatively supercritical regime characterized by cortical hyperexcitability. These findings and the proposed framework provide potential resting-state predictors of BCI aptitude in people with ALS and may inform the detection of so-called