Webcam Based Kinematics Recording System for Brain Machine Interfacing Studies
Abstract
Brain machine interfacing research consists of recording neural data from neurons adjacent to multi-electrode arrays to link with physical behavior. It is critical to match data from such recordings with motion data in order to assess the mapping of neural data to the behavior. Information on upper limb location is used to understand this mapping for reaching and grasping. Given that neural recordings are sensitive to electromagnetic interference, this project aims to replace a current motion tracking system, TrakSTAR, which works through electromagnetic signals, with a new system that introduces little interference. By implementing this new system, consisting of two webcams simultaneously recording motion of the subject’s upper limb, we hypothesize that more accurate motion data can be acquired without compromising neural recordings. The webcams’ video data will be analyzed using a custom MATLAB code tracking a specific color. Using motion data from different trials, the graphs produced from the webcam system will be compared to those from TrakSTAR in order to validate the hypothesis. To further validate the webcam system, the electromagnetic interference in neural recordings produced by TrakSTAR during those trials will be compared the interference produced by the webcams. In the future, to improve the tracking accuracy of the webcams, an algorithm will be developed to aid the prediction of motion in “blind spots”. This will reduce the current estimation errors resulting from available interpolation functions in MATLAB. Such improvements will help facilitate the analysis of neural data from neural interfaces.
Poster
Automating a Behavioral Reaching Task to Reduce Variability in Neural Recordings
