Incorporating modern neuroscience findings to improve brain-computer interfaces: tracking auditory attention.
This repository contains code for a study comparing sensor- and source-based BCIs. Specifically, it shows the theoretical and quantitative advantages (using both simulated and real data) associated with using the source space instead of the sensor space in a BCI context. We demonstrate this by classifying when subjects switched spatial auditory attention with data from a previous task.
- Wronkiewicz, M., Larson, E., and Lee, A. KC (2016). Incorporating modern neuroscience findings to improve brain-computer interfaces: tracking auditory attention. Journal of Neural Engineering
The code makes use of at least these specialized libraries:
- MNE-Python v0.11
Raw data is processed with
Figures 1 and 2 were created using meshes obtained via MRI scans and Blender.
The activation simulation in Figure 4 is created with
The synthetic data for Figure 5 was created using
switchPredSim.py, reorganized to link into previously written plotting code with
reformulate_sim_scores.py, and plotted with
The script for computing spherical inverse models is
Code for sensor and source based classification are in
The script used for the statistics and plotting of Figure 6 is
General functions and parameters are in
Unfortunately, the raw data are not included because:
- it contains (HIPAA-violating) identifying information and
- the raw data are many tens of gigabytes