moabb.datasets.Zhou2020

class moabb.datasets.Zhou2020(subjects=None, sessions=None, *, cohort=None, return_all_modalities=False)

Bases: BaseDataset

[source]

Dataset Snapshot

Zhou2020

Imagery, 4 classes (left_hand vs right_hand vs feet vs rest)

AuthorsQing Zhou, Jiafan Lin, Lin Yao, Yueming Wang, Yan Han, Kedi Xu

🇨🇳 Zhejiang University, CN·2021

Imagery Code: Zhou2020 20 subjects 7 sessions 28 ch (26 EEG) 500 Hz 4 classes 5.0 s trials CC BY 4.0

Class Labels: left_hand, right_hand, feet, rest

Quickstart Read Paper Compare Similar Report Issue

Overview

7-day motor imagery BCI EEG dataset from Zhou et al 2021.

Dataset from the article Relative Power Correlates With the Decoding Performance of Motor Imagery Both Across Time and Subjects

It contains data recorded on 20 subjects over 7 sessions (one session every ~2 days over 2 weeks) with no feedback training. Two groups of subjects were recorded with a 64-channel Neuroscan SynAmps2 system at 500 Hz:

• - S-subjects (subjects 1-12): 41 EEG + 2 EOG + 2 mastoid channels
• A-subjects (subjects 13-20): 26 EEG + 2 EOG channels

Both cohorts share 26 fronto-central electrodes (F3-F4, FC5-FC6, C5-C6, CP5-CP6); the S-subjects add F5, F6, AF3, AF4, P5-P6, PO3, POz, PO4 and Oz. The data files carry no channel names: they come from Table 1 and Fig. 2 of the authors' readme_MIdataset.docx and from their MI_dataset.py, both distributed with the original IEEE DataPort deposit but not with the Zenodo copy used for download. The S-subject files also hold the M1/M2 mastoid channels (columns 28-29, as MI_dataset.py drops them), which Table 1 omits; the column placement was checked against the recordings' correlation structure (:gh:`1120`). hEOG and vEOG (type eog) and the mastoids (type misc) are loaded only with return_all_modalities.

Cohorts and channel sets. MOABB paradigms and evaluations stack the epochs of every subject of a dataset into one array, so all subjects of a dataset object must return the same channels. The cohort option selects which channel set that is:

• - cohort=None (default): all 20 subjects with the 26 channels both cohorts share. This is the only set that covers every subject, and the one the paper analyses. With return_all_modalities, hEOG and vEOG are added; the mastoids are not, as the A-subjects have none.
• cohort="S": subjects 1-12 only, with all 41 EEG channels (and hEOG, vEOG, M1, M2 with return_all_modalities).
• cohort="A": subjects 13-20 only, with the 26 EEG channels (and hEOG, vEOG with return_all_modalities).

Subject numbers stay 1-20 in every option: S1-S12 are subjects 1-12 and A1-A8 subjects 13-20. Asking a cohort for a subject of the other cohort raises a ValueError. Each option has its own code (Zhou2020, Zhou-s2020, Zhou-a2020), so cached epochs and results never mix channel sets; downloads are shared. METADATA of a cohort instance describes that variant (subjects, trials, channels); the sex and age figures cover all 20 subjects, so only the default carries them.

Four MI classes were recorded: left hand, right hand, both feet, and idle/rest. Each session contains ~6 runs of 40 trials each (10 per class), giving ~240 trials per session and ~1680 trials per subject.

The data is stored as Neuroscan NSsignal NPZ files with continuous recordings (band-pass 0.5-100 Hz, 50 Hz notch). Events are encoded using GDF/BioSig codes: 769 (left), 770 (right), 771 (feet), 780 (rest).

:param subjects: Subjects to load (1-20, restricted to the selected cohort). :type subjects: list of int | None :param sessions: Sessions to keep. :type sessions: list of int | str | None :param cohort: None returns all 20 subjects on the 26 shared channels; "S" subjects 1-12 with all 41 EEG channels; "A" subjects 13-20 with their 26 channels. :type cohort: None | "S" | "A" :param return_all_modalities: Also return hEOG/vEOG (eog) and, with cohort="S", the M1/M2 mastoids (misc). :type return_all_modalities: bool | dict

Show more ▾Open in Overview tab →

Citation & Impact

• Paper DOI10.3389/fnhum.2021.701091
• CitationsLoading…
• Public APICrossref | OpenAlex
• Page Views
  30d: 0 · all-time: 26

  #83 of 152 · Top 55% most viewed

  Updated: 2026-10-22 UTC

Stimulus Protocol

5s task window per trial · 4-class imagery paradigm · 6 runs/session across 7 sessions

HED Event Tags

HED tags4/4 events annotated

Source: MOABB BIDS HED annotation mapping.

Sensory-event

4

Agent-action

3

Experimental-stimulus

1

Rest

1

Visual-presentation

1

left_hand

Sensory-eventAgent-action

right_hand

Sensory-eventAgent-action

feet

Sensory-eventAgent-action

rest

Sensory-eventExperimental-stimulusVisual-presentationRest

HED tree view

Tree · left_hand

├─ Sensory-event
│  ├─ Experimental-stimulus
│  └─ Visual-presentation
└─ Agent-action
   └─ Imagine
      ├─ Move
      └─ Left
         └─ Hand

Tree · right_hand

├─ Sensory-event
│  ├─ Experimental-stimulus
│  └─ Visual-presentation
└─ Agent-action
   └─ Imagine
      ├─ Move
      └─ Right
         └─ Hand

Tree · feet

├─ Sensory-event
│  ├─ Experimental-stimulus
│  └─ Visual-presentation
└─ Agent-action
   └─ Imagine
      ├─ Move
      └─ Foot

Tree · rest

├─ Sensory-event
├─ Experimental-stimulus
├─ Visual-presentation
└─ Rest

Channel Summary

Total channels28

EEG26 (Ag/AgCl)

EOG2

Montage10-05

Sampling500 Hz

Referencevertex (Cz)

Filter{'bandpass': [0.5, 100], 'notch_hz': 50}

Notch / line50 Hz

This diagram is automatically generated from MOABB metadata. Please consult the original publication to confirm the experimental protocol details.

Overview

7-day motor imagery BCI EEG dataset from Zhou et al 2021.

Dataset from the article Relative Power Correlates With the Decoding Performance of Motor Imagery Both Across Time and Subjects [1].

It contains data recorded on 20 subjects over 7 sessions (one session every ~2 days over 2 weeks) with no feedback training. Two groups of subjects were recorded with a 64-channel Neuroscan SynAmps2 system at 500 Hz:

• S-subjects (subjects 1-12): 41 EEG + 2 EOG + 2 mastoid channels

• A-subjects (subjects 13-20): 26 EEG + 2 EOG channels

Both cohorts share 26 fronto-central electrodes (F3-F4, FC5-FC6, C5-C6, CP5-CP6); the S-subjects add F5, F6, AF3, AF4, P5-P6, PO3, POz, PO4 and Oz. The data files carry no channel names: they come from Table 1 and Fig. 2 of the authors’ readme_MIdataset.docx and from their MI_dataset.py, both distributed with the original IEEE DataPort deposit [2] but not with the Zenodo copy used for download. The S-subject files also hold the M1/M2 mastoid channels (columns 28-29, as MI_dataset.py drops them), which Table 1 omits; the column placement was checked against the recordings’ correlation structure (#1120). hEOG and vEOG (type eog) and the mastoids (type misc) are loaded only with return_all_modalities.

Cohorts and channel sets. MOABB paradigms and evaluations stack the epochs of every subject of a dataset into one array, so all subjects of a dataset object must return the same channels. The cohort option selects which channel set that is:

• cohort=None (default): all 20 subjects with the 26 channels both cohorts share. This is the only set that covers every subject, and the one the paper analyses. With return_all_modalities, hEOG and vEOG are added; the mastoids are not, as the A-subjects have none.

• cohort="S": subjects 1-12 only, with all 41 EEG channels (and hEOG, vEOG, M1, M2 with return_all_modalities).

• cohort="A": subjects 13-20 only, with the 26 EEG channels (and hEOG, vEOG with return_all_modalities).

Subject numbers stay 1-20 in every option: S1-S12 are subjects 1-12 and A1-A8 subjects 13-20. Asking a cohort for a subject of the other cohort raises a ValueError. Each option has its own code (Zhou2020, Zhou-s2020, Zhou-a2020), so cached epochs and results never mix channel sets; downloads are shared. METADATA of a cohort instance describes that variant (subjects, trials, channels); the sex and age figures cover all 20 subjects, so only the default carries them.

Four MI classes were recorded: left hand, right hand, both feet, and idle/rest. Each session contains ~6 runs of 40 trials each (10 per class), giving ~240 trials per session and ~1680 trials per subject.

The data is stored as Neuroscan NSsignal NPZ files with continuous recordings (band-pass 0.5-100 Hz, 50 Hz notch). Events are encoded using GDF/BioSig codes: 769 (left), 770 (right), 771 (feet), 780 (rest).

param subjects:

Subjects to load (1-20, restricted to the selected cohort).

type subjects:

list of int | None

param sessions:

Sessions to keep.

type sessions:

list of int | str | None

param cohort:

None returns all 20 subjects on the 26 shared channels; "S" subjects 1-12 with all 41 EEG channels; "A" subjects 13-20 with their 26 channels.

type cohort:

None | “S” | “A”

param return_all_modalities:

Also return hEOG/vEOG (eog) and, with cohort="S", the M1/M2 mastoids (misc).

type return_all_modalities:

bool | dict

References

[1]

Zhou, Q., Lin, J., Yao, L., Wang, Y., Han, Y., Xu, K. (2021). Relative Power Correlates With the Decoding Performance of Motor Imagery Both Across Time and Subjects. Frontiers in Human Neuroscience, 15, 701091. https://doi.org/10.3389/fnhum.2021.701091

[2]

Zhou, Q. et al. (2020). EEG dataset of 7-day Motor Imagery BCI. IEEE DataPort. https://doi.org/10.21227/f1c7-7x89

Code Examples

from moabb.datasets import Zhou2020
dataset = Zhou2020()
data = dataset.get_data(subjects=[1])
print(data[1])

Metadata

Dataset summary

#Subj	20
#Chan	26
#Classes	4
#Trials / class	420
Trials length	5 s
Freq	500 Hz
#Sessions	7
#Runs	6
Total_trials	33600

Participants

• Population: healthy

• Age: 23.2 (range: 21-27) years

• Handedness: right-handed

• BCI experience: mixed

Equipment

• Amplifier: Neuroscan SynAmps2

• Electrodes: Ag/AgCl

• Montage: standard_1005

• Reference: vertex (Cz)

Data Access

• DOI: 10.3389/fnhum.2021.701091

• Data URL: https://zenodo.org/records/18988317

• Repository: Zenodo

Experimental Protocol

• Paradigm: imagery

• Feedback: none

• Stimulus: arrow cues

__init__(subjects=None, sessions=None, *, cohort=None, return_all_modalities=False)

Initialize function for the BaseDataset.

property all_subjects

Full list of subjects available in this dataset (unfiltered).

convert_to_bids(path=None, subjects=None, overwrite=False, format='EDF', verbose=None, generate_figures=False)

Convert the dataset to BIDS format.

Saves the raw EEG data in a BIDS-compliant directory structure. Unlike the caching mechanism (see CacheConfig), the files produced here do not contain a processing-pipeline hash (desc-<hash>) in their names, making the output a clean, shareable BIDS dataset.

Parameters:

• path (str | Path | None) – Directory under which the BIDS dataset will be written. If None the default MNE data directory is used (same default as the rest of MOABB).

• subjects (list of int | None) – Subject numbers to convert. If None, all subjects in subject_list are converted.

• overwrite (bool) – If True, existing BIDS files for a subject are removed before saving. Default is False.

• format (str) – The file format for the raw EEG data. Supported values are "EDF" (default), "BrainVision", and "EEGLAB".

• verbose (str | None) – Verbosity level forwarded to MNE/MNE-BIDS.

• generate_figures (bool) – If True, generate interactive neural signature HTML figures in {bids_root}/derivatives/neural_signatures/. Requires plotly (pip install moabb[interactive]). Default is False.

Returns:

bids_root – Path to the root of the written BIDS dataset.

Return type:

pathlib.Path

Examples

>>> from moabb.datasets import AlexMI
>>> dataset = AlexMI()
>>> bids_root = dataset.convert_to_bids(path="/tmp/bids", subjects=[1])

Notes

Use CacheConfig to configure caching for get_data(). Use moabb.datasets.bids_interface.get_bids_root to get the BIDS root path.

Added in version 1.5.

data_path(subject, path=None, force_update=False, update_path=None, verbose=None)

Get path to local copy of a subject data.

Parameters:

• subject (int) – Number of subject to use

• path (None | str) – Location of where to look for the data storing location. If None, the environment variable or config parameter MNE_DATASETS_(dataset)_PATH is used. If it doesn’t exist, the “~/mne_data” directory is used. If the dataset is not found under the given path, the data will be automatically downloaded to the specified folder.

• force_update (bool) – Force update of the dataset even if a local copy exists.

• update_path (bool | None Deprecated) – If True, set the MNE_DATASETS_(dataset)_PATH in mne-python config to the given path. If None, the user is prompted.

• verbose (bool, str, int, or None) – If not None, override default verbose level (see mne.verbose()).

Returns:

path – Local path to the given data file. This path is contained inside a list of length one, for compatibility.

Return type:

list of str

download(subject_list=None, path=None, force_update=False, update_path=None, accept=False, verbose=None)

Download all data from the dataset.

This function is only useful to download all the dataset at once.

When the dataset declares a nemar_id and the download provider is not "upstream", the files come from NEMAR’s sourcedata/ – the original pre-BIDS distribution, byte-identical to what the upstream host serves. On any NEMAR failure this falls back to the dataset’s own downloader with a warning (unless the provider is pinned to "nemar"). See sourcedata_path() and moabb.set_download_provider().

Parameters:

• subject_list (list of int | None) – List of subjects id to download, if None all subjects are downloaded. On the NEMAR path each subject is resolved through the deposit’s sourcedata_provenance.json; deposits enriched before that manifest recorded subjects fetch the whole tree.

• path (None | str) – Location of where to look for the data storing location. If None, the environment variable or config parameter MNE_DATASETS_(dataset)_PATH is used. If it doesn’t exist, the “~/mne_data” directory is used. If the dataset is not found under the given path, the data will be automatically downloaded to the specified folder.

• force_update (bool) – Force update of the dataset even if a local copy exists.

• update_path (bool | None) – If True, set the MNE_DATASETS_(dataset)_PATH in mne-python config to the given path. If None, the user is prompted. Not used on the NEMAR path.

• accept (bool) – Accept licence term to download the data, if any. Default: False. Only relevant to the dataset’s own downloader; NEMAR mirrors are already public.

• verbose (bool, str, int, or None) – If not None, override default verbose level (see mne.verbose()).

get_additional_metadata(subject: str, session: str, run: str)

Load additional metadata for a specific subject, session, and run.

This method is intended to be overridden by subclasses to provide additional metadata specific to the dataset. The metadata is typically loaded from an events.tsv file or similar data source.

Parameters:

• subject (str) – The identifier for the subject.

• session (str) – The identifier for the session.

• run (str) – The identifier for the run.

Returns:

A DataFrame containing the additional metadata if available, otherwise None.

Return type:

None | pandas.DataFrame

get_block_repetition(paradigm, subjects, block_list, repetition_list)

Select data for all provided subjects, blocks and repetitions.

subject -> session -> run -> block -> repetition

See also

get_data

Parameters:

• subjects (List of int) – List of subject number

• block_list (List of int) – List of block number

• repetition_list (List of int) – List of repetition number inside a block

Returns:

data – dict containing the raw data

Return type:

Dict

get_data(subjects=None, cache_config=None, process_pipeline=None, n_jobs=1)

Return the data corresponding to a list of subjects.

The returned data is a dictionary with the following structure:

data = {"subject_id": {"session_id": {"run_id": run}}}

subjects are on top, then we have sessions, then runs. A sessions is a recording done in a single day, without removing the EEG cap. A session is constitued of at least one run. A run is a single contiguous recording. Some dataset break session in multiple runs.

Processing steps can optionally be applied to the data using the *_pipeline arguments. These pipelines are applied in the following order: raw_pipeline -> epochs_pipeline -> array_pipeline. If a *_pipeline argument is None, the step will be skipped. Therefore, the array_pipeline may either receive a mne.io.Raw or a mne.Epochs object as input depending on whether epochs_pipeline is None or not.

Parameters:

• subjects (List of int) – List of subject number

• cache_config (dict | CacheConfig) – Configuration for caching of datasets. See CacheConfig for details.

• process_pipeline (sklearn.pipeline.Pipeline | None) – Optional processing pipeline to apply to the data. To generate an adequate pipeline, we recommend using moabb.make_process_pipelines(). This pipeline will receive mne.io.BaseRaw objects. The steps names of this pipeline should be elements of StepType. According to their name, the steps should either return a mne.io.BaseRaw, a mne.Epochs, or a numpy.ndarray. This pipeline must be “fixed” because it will not be trained, i.e. no call to fit will be made.

• n_jobs (int) – Number of jobs to run in parallel over subjects (passed to joblib.Parallel). Default 1 (sequential). Per-subject processing (reading, filtering, resampling, epoching) is independent, so this gives a near-linear speedup for datasets with many subjects.

Returns:

data – dict containing the raw data

Return type:

Dict

property metadata

Structured metadata of this cohort variant.

sourcedata_path(subject=None, path=None, force_update=False, verbose=None)

Get the dataset’s original pre-BIDS distribution from NEMAR.

Where data_path() fetches the original files from the upstream host, this fetches the copy NEMAR mirrors under sourcedata/. The files keep their upstream names, so the two are interchangeable in content – but NEMAR stays reachable when the upstream host is slow, rate-limited, behind a bot gate, or retired.

Parameters:

• subject (int | str | None) – Restrict the download to one subject, resolved through the deposit’s sourcedata_provenance.json. Deposits enriched before that manifest recorded subjects fall back to the whole tree with a warning. When None the whole tree is fetched.

• path (None | str) – Base path where MOABB stores datasets.

• force_update (bool) – Re-fetch even when a local copy is present.

• verbose (bool, str, int, or None) – If not None, override default verbose level.

Returns:

Local path to the sourcedata directory.

Return type:

str

Raises:

• ValueError – If the dataset declares no nemar_id.

• moabb.datasets.download.NemarDownloadError – If the download fails or the deposit publishes no sourcedata/.

Examples using moabb.datasets.Zhou2020

Benchmarking with MOABB

Benchmarking with MOABB