Multimodal Neuroelectric Interface Development: A Survey of Research at NASA Ames Research Center

We are developing electromyographic (EMG) and electroencephalographic (EEG) methods that bypass muscle activity and draw control signals for human-computer interfaces directly from the human nervous system. We have made progress in four areas: a) real-time pattern recognition algorithms for decoding sequences of forearm muscle activity associated with control gestures, b) signal-processing strategies for computer interfaces using EEG signals, c) a flexible computation framework for neuroelectric interface research, d) non-contact sensors, which measure EMG or EEG signals without resistive contact to the body. interfaces. We will describe an EMG-based flight stick, an EMG-based numeric keypad, an EEG-based interface for smooth, continuous control of a one dimension of motion in a graphic display, and comparison of algorithms for modeling the EEG patterns associated with real and imagined mouse motion or typing. Finally, we will present some new results on the development of non-contact electric field sensors for EMG and EEG recording. These sensors offer a less intrusive alternative to current sensing technology, which will make them more suitable for real-world applications. Our approach is to describe a body of developmental research, mostly still in progress, and to indicate methods that have potential for engineering development. Given the BCI focus of this special issue, descriptions of purely EMG-based interfaces will be brief. We will describe the EEG results and the new sensor developments in more detail.