# 2020 International BCI Competition Track 3: 64-Channel Imagined Speech EEG of Five Words and Phrases from 15 Participants (OSF)

## Facts

- Provider: Jeong et al. (2020 International BCI Competition organizers); Hugging Face mirror
- License: CC-BY-4.0 (https://api.osf.io/v2/nodes/pq7vb/?embed=license)
- Size: 52 files
- Format: TSV (in sampled files)
- Stored at: Hugging Face, https://huggingface.co/datasets/raei/Jeong-2022-InternationalBCICompetition2020Review-track3
- Access: public, readable without an account
- Category: Neuroscience
- Source: https://doi.org/10.3389/fnhum.2022.898300
- Page: https://scigantic.com/archive/0e1083cb-cbec-5e53-b8dc-2b00390a90e9
- Keywords: Neuroscience, EEG, TSV, Hugging Face

## About

Track 3 (Data Set-C) of the 2020 International Brain-Computer Interface Competition, reviewed in Jeong et al., Frontiers in Human Neuroscience 2022: scalp EEG recorded while 15 participants (S1 to S15, aged 20 to 30) silently imagined saying five English words or phrases: hello, help me, stop, thank you and yes. Recordings have 64 EEG channels at 256 Hz, epoched from -500 to 2600 ms (795 samples, amplitudes in microvolts). Each participant has 300 training trials (60 per class), 50 validation and 50 test trials, 6,000 trials in all, in MATLAB files Data_Sample01.mat to Data_Sample15.mat under Training set, Validation set and Test set; test ground truth is in the released answer sheet, and the competition protocol is subject-dependent classification. Training and validation files open with scipy.io.loadmat, test files (MATLAB v7.3) with h5py. Use for imagined speech decoding, silent speech BCI and EEG classification benchmarks. This mirror keeps the 48 source files from OSF project pq7vb byte for byte with a checksum manifest. The OSF project is licensed CC BY 4.0. Cite Jeong et al., Front. Hum. Neurosci. 16:898300, 2022.

## Load the data

```python
import pandas as pd
df = pd.read_csv("hf://datasets/raei/Jeong-2022-InternationalBCICompetition2020Review-track3/MANIFEST.tsv", sep='\t')
df.head()
```

## Data structure

File types in sampled files: TSV

Example file: `MANIFEST.tsv`

## Open it

- In a hosted notebook with the data mounted: https://scigantic.com/archive/0e1083cb-cbec-5e53-b8dc-2b00390a90e9
- Full record as JSON (schema card included): https://api.scigantic.com/api/archives/0e1083cb-cbec-5e53-b8dc-2b00390a90e9
- Over MCP: `get_schema_card` with id `0e1083cb-cbec-5e53-b8dc-2b00390a90e9` at https://api.scigantic.com/mcp

## Related topics

- [Neuroscience datasets](https://scigantic.com/topics/neuroscience)
- [EEG datasets](https://scigantic.com/topics/eeg)
- [TSV datasets](https://scigantic.com/topics/tsv)
