orhansoyuhos / sEEG-Forced-Two-Choice-Task-Classification

This project is from the NeuralStorm Workshop in 2023. On the final day, we had a small competition to classify choices made in a sEEG Forced Two-Choice Task. Wenqing and I from CCLAB teamed up and tied for first place in the competition! We primarily utilized summary statistics derived from several frequency bands to predict the choices made.

sEEG Forced Two-Choice Task Classification

This repository contains the code and findings of a project carried out during the NeuralStorm Workshop in 2023. On the final day of the workshop, a competition was held to classify choices made in a sEEG Forced Two-Choice Task. Teaming up with Wenqing from CCLAB, we secured a tie for the first place in the competition!

Overview

We primarily focused on utilizing summary statistics derived from several frequency bands to predict the choices made. Among the different classification algorithms evaluated, Linear Discriminant Analysis (LDA) yielded the best results.

Dataset

The dataset comprises intracranial electroencephalography (iEEG) recordings collected during a sEEG Forced Two-Choice Task. The dataset is available on OpenNeuro: Dataset Link

Competition

The competition was hosted on Kaggle. More details can be found here: Competition Link

Preprocessing

The data preprocessing involved normalizing the signal data, extracting power from different frequency bands, and deriving summary statistics.