The Brain Computer Interaction Technology (BCIT) project focuses on demonstrating novel forms of interactions to enhance human-system teams teaming. With a focus on incorporating advances from multiple research fields, including Translational Neuroscience, Cybernetics, and Data Sciences, we aim to demonstrate proof-of-concept technologies that use neural signals and advanced machine learning methods to improve human interactions with computers, autonomous systems, their environment, and even with other humans. Our BCIT research is focused solely on the development of non-medical performance-enhancing technologies, and aims to improve the reliability, robustness, and effectiveness of human-technology interactions with healthy, high-performing Soldiers. The long term goal is for BCIT to allow novel human-system teaming approaches that effectively utilize the dynamic adaptability of humans and the processing power of autonomy. Our perspective is outlined in “Brain–Computer Interface Technologies in the Coming Decades,” (Lance et al 2012) published in the 100th anniversary issue of the Proceedings of the IEEE.
Brain Computer Interaction Technology POC: David Slayback, Brent Lance
Human-AI Image Labeler (HAIL): One of the strategic goals of CAST researchers is to effectively integrate humans with intelligent agents. The HAIL system integrates a network of analysis agents consisting of humans using Rapid Serial Visual Presentation (RSVP)-based BCIs, human manual labelers, and autonomous Computer Vision (CV) agents in order to analyze a database of images as accurately as a human manual labeler but in a fraction of the time. For more information, please see the paper “Cortically Coupled Computing: A New Paradigm for Synergistic Human–Machine Interaction,” (Saproo et al, 2016) published in IEEE Computer as part of the cover feature on Emerging Computing Paradigms.
BCI-Gem Game: CAST scientists are specifically interested in overcoming one of the most critical challenges for BCIT: addressing the effect of multi-timescale human state dynamics on system performance. To overcome this challenge, the BCI-Gem game has been developed for studying how human state change affects BCIT performance and how humans adapt to BCIT during long-term usage. The BCI-Gem game is a 3-match game-similar to Bejeweled, Candy Crush, or Puzzle Quest-that has BCI paradigms embedded within the game. For more information, please see the paper “Towards Serious Games for Improved BCI,” (Lance et al, 2015).
Human Interest Detector (HID): CAST researchers in the MIND laboratory are currently developing HID, a passive brain monitoring system that attempts to detect operator interest in visual scenes. HID has three purposes: 1) provide a high bandwidth, novel stream of information from the operator based on the concept of Soldier as a sensor, 2) obtain critical insights into the effectiveness of a operator using opportunistic human-sensing, and 3) enable plug-and-play testing of novel closed-loop BCI technologies in a variety of paradigms. By combining eye-movement-based EEG analysis with CV agents, HID captures and prioritizes interest in images or visual clips from an operator without requiring the operator to intentionally indicate interest level. Using an opportunistic approach, as opposed to developing a system that specifically requires operators to indicate interest levels, potentially allows for HID to be effectively used in concert with other BCIT's across multiple tasks.