Project 3-5: Mobile Brain–Machine Interface for Integrated Information–Social /Cognitive Network Operations

Principal Investigator: Kwong T. Ng (New Mexico State University)

Much of the research in AHPCRC focuses on developing high-performance computational capabilities, better problem-solving algorithms, and software/hardware systems that operate efficiently and economically. This project focuses on another side of the system—the humans who ultimately use and apply the simulations and other information generated by their HPC systems.

 The researchers for this project use an electroencephalography (EEG)-based brain–machine interface (BMI) as a sensor of human cognition. This interface provides quantitative neural signatures that represent the state of a social/cognitive network. Human test subjects wear electrodes attached to the scalp to record the voltage signals generated by the neural electric sources associated with underlying mental activities. A beamforming inverse algorithm is used to reconstruct the neural source distributions in the brain, which then serve as the neural signatures. The information network can extract these signatures dynamically and use them to perform integrated information–cognitive network functions.  

The researchers are paying special attention to the design of algorithms and routines that can be implemented efficiently on multi-core processors and general purpose graphic processing units (GPGPUs). They plan to obtain fast, accurate scalp potential calculations to enable high-resolution neural source imaging, using a beamformer to produce tomographic maps of the source distribution. Focusing on specific brain locations is done by adjusting the EEG signal weights.