Code for running the analyses in "Neural Manifolds of Human Intracranial Recordings During Naturalistic Arm Movements" by Z. Steine-Hanson, R. Rao and B. Brunton
This paper and code explore neural manifolds in naturalistic ECoG data Methods heavily influenced by Natraj et al. 2022
To use this repo:
- Set up your environment:
- Create a matching virtual environment using ECoGDL_venv.yml Note: may not be up to date for some packages.
- Then, in the virtual environment, pip install the src folder using:
pip install -e .
While in the home directory for the project code.
You can test that this worked by importing src in a python interpretter. - You will also need to download these other git repos and pip install them:
- For loading in mat files: https://github.com/skjerns/mat7.3
- Create a new experiment parameter json file, and update the parameters you'd like to change. An example can be found in experiment_params.
I like to name the files such that:- the number indicates which dataset is used
- the alphabetic character indicates an over-arching experiment type, such as varying frequency bands for filtering
- the roman numeral indicates variations for that experiment
- For example: exp_params_1a_i.json uses the 4-class naturalistic movement data, with motor roi projection, for the first test
- Decide if you want to run PCA on each movement separetly, or together. Most likely you will want to run them separately.