MindX: A Low-Cost Benchmark System for Real-Time Brain-Controlled Gaming

Xiaodong Qu, Jingwen Dou, Wensi Xie, Yiran Wang

Keywords: Gaming, BCI, Brain-Computer Interfaces, EEG, CNN, Transformer

Abstract: We introduce MindX, an open benchmark platform for real-time EEG-based game control using low-cost, consumer-grade hardware. The system features three directional-control games—Hide-and-Seek, Rhythm Game, and Snake Game—designed to elicit motor imagery and attention-related neural signals. A lightweight CNN+LDA model processes 4-channel EEG from a Muse S headset and issues directional predictions (left and right) every 300\~ms, with end-to-end latency under 350\~ms. A user-centered co-adaptive loop enables lightweight personalization based on gameplay feedback. In a pilot study with three users, MindX achieved 76\% accuracy, well above the 50\% baseline. The framework provides a reproducible and extensible testbed for evaluating real-time EEG decoding pipelines.

Primary Area: applications to neuroscience & cognitive science

Submission Number: 13637