mi-bci-analysis

MI BCI Analysis

Analysis of data collected during a 3-day Motor Imagery (MI) Brain-Computer Interface (BCI) experiment involving 8 healthy participants.

MATLAB Toolbox

biosig

disp('[config] - Adding biosig toolbox');
addpath(genpath('<absolute-path>\mi-bci-analysis\toolbox\biosig\biosig\t200_FileAccess'));
addpath(genpath('<absolute-path>\mi-bci-analysis\toolbox\biosig\biosig\t250_ArtifactPreProcessingQualityControl'));

% NOTE: Toolbox version may be different

eeglab

disp('[config] - Adding eeglab toolbox');
addpath(genpath('<absolute-path>\mi-bci-analysis\toolbox\eeglab\eeglab2024.2'));

% NOTE: Toolbox version may be different

Dataset

The data was recorded using a 16-channel EEG amplifier at a sampling rate of 512 Hz, where the electrodes were positioned according to the 10-20 international system.

Each participant completed at least two recording days:

  • Day 1: 3 “offline” runs (calibration, without real feedback) and 2 “online” runs
    (with real feedback);
  • Day 2 and Day 3: 2 “online” runs per day.

Resource

https://cloud.dei.unipd.it/index.php/s/DLJfJccgFnFiDZY

Instructions

  1. Create toolbox/ folder in root directory and insert biosig and eeglab toolboxes
  2. Create dataset/ folder in root directory and insert micontinuous dataset
  3. Run generation.m script
  4. Run analysis.m script
  5. Run selection.m script
  6. Run training.m script
  7. Run evaluation.m script

Tested Environments

  • Local machine, Windows 11, MATLAB R2024b
  • Local machine, MacOS Sequoia, MATLAB R2024b
  • Local machine, Windows 11, MATLAB R2023a
  • Local machine, Windows 11, MATLAB R2023b

Authors

Contributions

Member Workload Work
Federico Pivotto 25% Project structure, data generation, training workflow, evaluation workflow, feature selection, selection workflow, model training and evalutation chapter in report
Alessandro Bozzon 25% Topoplot computation, analysis workflow, feature selection, performance metrics, topoplot chapter in report
Riccardo Simion 25% Spectrogram computation, model training, feature selection, spectrogram chapter in report
Riccardo Zerbinati 25% Feature map computation, model evaluation, feature selection, feature map chapter in report


Visit original content creator repository

Comments

Leave a Reply

Your email address will not be published. Required fields are marked *