torcheeg.datasets

Emotion Recognition Datasets

DEAPDataset	A multimodal dataset for the analysis of human affective states.
DREAMERDataset	A multi-modal database consisting of electroencephalogram and electrocardiogram signals recorded during affect elicitation by means of audio-visual stimuli.
SEEDDataset	The SJTU Emotion EEG Dataset (SEED), is a collection of EEG datasets provided by the BCMI laboratory, which is led by Prof.
SEEDFeatureDataset	The SJTU Emotion EEG Dataset (SEED), is a collection of EEG datasets provided by the BCMI laboratory, which is led by Prof.
SEEDIVDataset	The SEED-IV dataset provided by the BCMI laboratory, which is led by Prof.
SEEDIVFeatureDataset	The SEED-IV dataset provided by the BCMI laboratory, which is led by Prof.
AMIGOSDataset	A dataset for Multimodal research of affect, personality traits and mood on Individuals and GrOupS (AMIGOS).
MAHNOBDataset	MAHNOB-HCI is a multimodal database recorded in response to affective stimuli with the goal of emotion recognition and implicit tagging research.
BCI2022Dataset	The 2022 EMOTION_BCI competition aims at tackling the cross-subject emotion recognition challenge and provides participants with a batch of EEG data from 80 participants with known emotional state information.
MPEDFeatureDataset	The Multi-Modal Physiological Emotion Database for Discrete Emotion (MPED), a multi-modal physiological emotion database, which collects four modal physiological signals, i.e., electroencephalogram (EEG), galvanic skin response, respiration, and electrocardiogram (ECG).

Personal Identification Datasets

M3CVDataset

A reliable EEG-based biometric system should be able to withstand changes in an individual's mental state (cross-task test) and still be able to successfully identify an individual after several days (cross-session test).

Motor Imagery Datasets

BCICIV2aDataset

A dataset for motor imagery, BCI Competition 2008 Graz data set A (BCICIV_2a).

Mother of all BCI Benchmarks

moabb.MOABBDataset

Mother of all BCI Benchmarks (MoABB) aims at building a comprehensive benchmark of popular Brain-Computer Interface (BCI) algorithms applied on an extensive list of freely available EEG datasets.

Steady-state Visual Evoked Potential Datasets

TSUBenckmarkDataset

The benchmark dataset for SSVEP-Based brain-computer interfaces (TSUBenckmark) is provided by the Tsinghua BCI Lab.

Customized Datasets

FolderDataset	Read EEG samples and their corresponding labels from a fixed folder structure.
CSVFolderDataset	Read meta information from CSV file and read EEG data from folder according to the meta information.

Hooks

torcheeg.datasets.before_trial_normalize(data: ndarray, eps: float = 1e-06, axis=0)

A common hook function used to normalize the signal of the whole trial before dividing it into chunks.

It is used as follows:

from functools import partial
dataset = DEAPDataset(
        ...
        before_trial=before_trial_normalize,
        num_worker=4)

If you want to pass in parameters, use partial to generate a new function:

from functools import partial
dataset = DEAPDataset(
        ...
        before_trial=partial(before_trial_normalize, eps=1e-5),
        num_worker=4)

Parameters:

• data (np.ndarray) – The input EEG signals or features of a trial.

• eps (float) – The term added to the denominator to improve numerical stability (default: 1e-6)

Returns:

The normalized results of a trial.

Return type:

np.ndarray

torcheeg.datasets.after_trial_normalize(data: ndarray, eps: float = 1e-06)

A common hook function used to normalize the signal of the whole trial after dividing it into chunks and transforming the divided chunks.

It is used as follows:

from functools import partial
dataset = DEAPDataset(
        ...
        after_trial=after_trial_normalize,
        num_worker=4)

If you want to pass in parameters, use partial to generate a new function:

from functools import partial
dataset = DEAPDataset(
        ...
        after_trial=partial(after_trial_normalize, eps=1e-5),
        num_worker=4)

Parameters:

• data (np.ndarray) – The input EEG signals or features of a trial.

• eps (float) – The term added to the denominator to improve numerical stability (default: 1e-6)

Returns:

The normalized results of a trial.

Return type:

np.ndarray

torcheeg.datasets.after_trial_moving_avg(data: list, window_size: int = 4)

A common hook function for smoothing the signal of each chunk in a trial after pre-processing.

It is used as follows:

from functools import partial
dataset = DEAPDataset(
        ...
        after_trial=after_trial_moving_avg,
        num_worker=4)

If you want to pass in parameters, use partial to generate a new function:

from functools import partial
dataset = DEAPDataset(
        ...
        after_trial=partial(after_trial_moving_avg, eps=1e-5),
        num_worker=4)

Parameters:

• data (np.ndarray) – A list of dictionaries, one of which corresponds to an EEG signal in trial. Each dictionary consists of two key-value paris, eeg and key. The value of eeg is the representation of the EEG signal and the value of key is its ID in the IO.

• window_size (int) – The window size of moving average. (default: 4)

Returns:

The smoothing results of a trial. It is a list of dictionaries, one of which corresponds to an EEG signal in trial. Each dictionary consists of two key-value paris, eeg and key. The value of eeg is the representation of the EEG signal and the value of key is its ID in the IO.

Return type:

list