EEG Semester Project: Average Task Value (Python/MNE)

This repository contains materials and a Python/MNE pipeline to reproduce and extend the EEG findings of the paper "Task‑level value affects trial‑level reward processing" using the OpenNeuro dataset ds004147.

Highlights

• Dataset: BIDS-structured BrainVision EEG under data/raw/ds004147.
• Original MATLAB reference: 原/averagetaskvalue-main (paper authors' code).
• Our pipeline: Python + MNE, with ICLabel/autoreject options, baseline regression, and group-level stats.

Quickstart

1. Create environment (conda recommended):

   • conda create -n eeg_project python=3.10 -y
   • conda activate eeg_project
   • pip install -r requirements.txt

2. Place data under data/raw/ds004147 (BIDS). Update src/config.py if needed.

3. Run the minimal sanity-check script (generates two figures):

   • python scripts/m1_quickcheck.py --sub 28

4. (Optional, course-aligned) Run mne-bids-pipeline with pipeline/config.py to generate derivatives and reports.

Project layout

• data/raw/ds004147/ BIDS dataset.
• src/ Python source modules (preprocessing, events, erp, tfr, stats).
• scripts/ Reproducible entry points (milestone-oriented helpers).
• notebooks/ Optional exploration notebooks.
• figures/ Generated plots (gitignored by default).
• results/ CSV summaries and stats (gitignored by default).
• pipeline/ mne-bids-pipeline config and docs.
• derivatives/ Outputs generated by pipelines (auto-created).

References

• Paper/preprint: https://doi.org/10.1101/2021.09.16.460600
• Dataset: https://openneuro.org/datasets/ds004147/
• MATLAB code: https://github.com/chassall/averagetaskvalue