e-epi-decoding

This repository will provide the official code for the paper:

"Anti-fibrotic Electronic Epineurium for High-fidelity Chronic Peripheral Nerve Interface in Freely Moving Animals"

In this work, we demonstrate behavior decoding using signals recorded from our novel electronic epineurium (e-epi) interface.

Directory setup

From the repository root:

• e-epi-decoding/
  • main.py
  • run.sh
  • requirements.txt
  • data_csv/ (create this directory and place the CSVs here)

Expected data directory structure

Download the CSV files from the data provided by the paper and place them under e-epi-decoding/data_csv/.

Install required packages

git clone https://github.com/NICALab/e-epi-decoding.git
cd e-epi-decoding
pip install -r requirements.txt

Reproduce results from paper

cd e-epi-decoding
bash run.sh

This will train the model and write checkpoints + the test confusion matrix figure under:

• e-epi-decoding/checkpoints/nerve_behavior_decoding/main/main_<timestamp>_stride<...>_window<...>_bash<...>_splitseed<...>/