In Columbia’s Brain Computer Interface class, run by Dr. Nima Mesgarani in the Electrical Engineering Department, our last lab assignment involved taking EEG recordings from one of our lab members (s/o to Jacob and Liang) during either left or right imagined movement. These EEG recordings are processed to reveal subtle differences in the motor cortex – resulting from just imagining moving your left or right hand! These differences are more pronounced when actual movement happens, but nonetheless there is a clear left versus right distinction that can be seen.
First, a brief overview of the background and experimental paradigm, which can also be found on the Neural Acoustic Processing Lab website as well.
The basis of this experiment is observing the desynchronization of Mu waves, which form patterns in the motor cortex when the cortex is idle. The disruption of these waves reduces the overall power of the EEG signal in the Mu frequency band. The left motor cortex controls the right hand and the right motor cortex controls the left hand, making it possible to see bilateral differences in power – explored in this experiment.
To focus on the specific relevant region of the brain, the EEG electrodes are arranged to provide maximum resolution near motor cortex areas. The specific electrode locations are shown below.
EEG data is acquired using g.Tech hardware and MATLAB SIMULINK software, initially filtered during acquisition from 0.5 – 30 Hz. Software that prompts the subject to imagine left or right movement in a predetermined but randomized order (20 on each side) provides indices which were later used to sort the EEG data into left and right categories. For this portion of the experimentation, the collected EEG data was broadly filtered for Mu waves (8-20 Hz) and power was calculated within each trial length, averaged for all the left side and right side separately, and the difference between the left and right side power is plotted at each electrode.
After processing, we were happy to see that our subject showed lower values on the left side of the scalp and higher values on the right side – as expected of the desynchronization when the right side signal is subtracted from the left side signal. This image is shown below.
Other interesting experiments we explored are the P300 Event Related Potential, the Auditory Oddball Paradigm, Auditory Steady State Response (ASSR), and Steady State Visually Evoked Responses (SSVER).
I also wanted to sincerely thank our TA, Vinay, who was knowledgeable and patient as we learned about these amazing topics.
Learn more at the Neural Acoustic Processing Lab BCI class website.