Neuromancer: A Brain-Computer Interface for Sustained Attention

Abstract

Sustaining one's attention is a perpetually di cult tasks for humans, not only because our minds tend to wander but also because we lose attention unpredictably, often not noticing our own attentional lapses until long after they have begun. In this experiment, we explore how EEG-derived neurofeedback provided to participants in real-time can help humans to not only to notice when they have lost attention, but also to create better mental representations of tasks. Participants responded to a task in which they were instructed to attend to pictures of faces and ignore superimposed pictures of places, or attend to pictures of places and ignore faces, while receiving real-time neurofeedback from a wireless EEG headset. We provided participants neurofeedback derived from how face-like or place-like their EEG data appeared when categorized by a classi er, and found that this modestly improved participants performance post-feedback. Additionally, we provide a software package, Neuromancer, for running EEG Brain Computer Interface experiments.