Restricted Boltzmann Machines in Sensory Motor Rhythm Brain-Computer Interfacing: A Study on Inter-Subject Transfer and Co-Adaptation

Naive users often perceive calibration of a Sensory Motor Rhythm (SMR) based Brain-Computer Interfaces (BCI) as tedious and lengthy. The lack of feedback during training is assumed to be a major cause. I.e. if one had already a reasonable model to start with, feedback training could be started immediately. One concept to address this issue is learning a general model and adapting it to new observations. In this study we applied this concept by utilizing a generative model entitled Restricted Boltzmann Machine (RBM). We investigated its feature extraction capabilities by fitting a RBM to recordings of 9 subjects. Generalization was assessed in an online coadaptive study, covering 12 volunteers (10 naive). An overall median accuracy of 88.9% (83.5% naive) with a standard-error of 6.5% (6.6% naive) was achieved for a classical hand versus feet motor imagery task. The online co-adaptive training itself lasted approximately 25 minutes. Feedback was already presented after a one minute setup run, whose purpose was to estimate initial statistics and to train an online artifact detection system.