Implanted brain-computer interface functionality during nighttime in late-stage amyotrophic lateral sclerosis

Brain-computer interfaces (BCIs) hold promise as assistive communication technology for people with severe paralysis. Although such BCIs should be available 24/7, feasibility of nocturnal BCI use has not been investigated. Here, we addressed this question using data from an electrocorticography-BCI user with amyotrophic lateral sclerosis. We investigated nocturnal dynamics of neural signal features used for BCI control. Additionally, we assessed nocturnal performance of a decoder trained on daytime data, by quantifying the number of unintentional BCI activations at night. Finally, we developed a nightmode functionality and assessed its performance. Mean and variance of low and high frequency band power were significantly higher at night than during the day. When applied to night data, daytime decoders caused unintentional BCI activations in 100% of nights (245 unintended click-commands and 13 unintended caregiver-calls per hour). The specifically developed nightmode functionality, however, functioned error-free in 79% of nights over a period of ± 1.5 years, allowing the user to reliably call the caregiver. Reliable nighttime use of a BCI requires strategies to adjust to circadian and sleep-related signal changes. This demonstration of a reliable nightmode and its long-term use by an individual with amyotrophic lateral sclerosis underscores the importance of 24/7 BCI reliability.