moabb.datasets.RomaniBF2025ERP

class moabb.datasets.RomaniBF2025ERP(data_folder: str | None = None, subjects: List[str] | None = None, exclude_subjects: List[str] | None = ['P15', 'P18'], calibration_length: int = 60, n_targets: int = 10, t_target: int = 1, nt_target: int = 2, interval: tuple = [-0.1, 1.0], extra_runs: bool = True, include_inference: bool = False, load_failed: bool = False, montage: str = 'standard_1020', sessions=None)

Bases: BaseDataset

[source]

Dataset Snapshot

RomaniBF2025ERP

BrainForm: a Serious Game for BCI Training and Data Collection - gamified BCI training system designed for scalable data collection using consumer hardware

P300 / ERP, 2 classes (Target vs NonTarget)

AuthorsMichele Romani, Devis Zanoni, Elisabetta Farella, Luca Turchet

🇮🇹 University of Trento, IT·2025

P300 / ERP Code: RomaniBF2025ERP 22 subjects 2 sessions 8 ch 250 Hz 2 classes 0.9 s trials CC BY 4.0

Class Labels: Target, NonTarget

Quickstart Read Paper Compare Similar Report Issue

Overview

MOABB class for BrainForm event-related potentials (ERP) dataset.

The BrainForm dataset is a dataset collected using a serious game for brain-computer interface (BCI) training and data collection. It includes EEG recordings from multiple subjects performing a ERP task on 10 unique stimuli. The dataset is organized in BIDS format and contains calibration and inference sessions.

Each subject performed two calibration sessions, one with a checkerboard texture on the targets, the other one with a grain texture. Each calibration was followed by an inference session where the subject played the game. The game consisted of two separate tasks: in the first one, the subject had to hit moving aliens using the color matched target; in the second one, they need to follow a randomized sequence of colors by hitting the corresponding targets to unlock a door.

Calibration sessions consisted of 60 trials on a single training target, for a total of 600 events (60 trials x 10 unique targets). This means that by default, the data is unbalanced, with 60 target events and 540 non-target events per session. In inference sessions, the number of events varied depending on the subject's performance

A total of 16 subjects took an optional free-play run after the main protocol, where they could choose their favorite texture and play the game again. The extra sessions include a calibration and can be included by setting the `extra_runs` parameter. The full protocol is described in A study on cross-subject decoding using this dataset is presented in

A total of 2 subjects (15 and 18) did not complete the full protocol and are excluded by default.

For many subjects, multiple calibration attempts were not successful, but they can be included by setting the `exclude_failed` parameter to False.

Inference sessions can be included along with calibration by setting the `include_inference` parameter to True, but the triggers only indicate the stimulus onset and not ground truth labels.

You can test the dataset with the following code:

Show more ▾Open in Overview tab →

Citation & Impact

• Paper DOI10.48550/arXiv.2510.10169
• CitationsLoading…
• Public APICrossref | OpenAlex
• Page Views
  30d: 18 · all-time: 29

  #75 of 97 · Top 78% most viewed

  Updated: 2026-03-12 UTC

Stimulus Protocol

0.9s task window per trial · 2-class p300 / erp paradigm · 1 runs/session across 2 sessions

HED Event Tags

HED tags2/2 events annotated

Source: MOABB BIDS HED annotation mapping.

Experimental-stimulus

2

Sensory-event

2

Visual-presentation

2

Non-target

1

Target

1

Target

Sensory-eventExperimental-stimulusVisual-presentationTarget

NonTarget

Sensory-eventExperimental-stimulusVisual-presentationNon-target

HED tree view

Tree · Target

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

Tree · NonTarget

├─ Sensory-event
├─ Experimental-stimulus
├─ Visual-presentation
└─ Non-target

Channel Summary

Total channels8

EEG8 (EEG)

Montagestandard_1020

Sampling250 Hz

Referenceright mastoid

Notch / line50 Hz

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

Overview

MOABB class for BrainForm event-related potentials (ERP) dataset.

The BrainForm dataset [1] is a dataset collected using a serious game for brain-computer interface (BCI) training and data collection. It includes EEG recordings from multiple subjects performing a ERP task on 10 unique stimuli. The dataset is organized in BIDS format and contains calibration and inference sessions.

Each subject performed two calibration sessions, one with a checkerboard texture on the targets, the other one with a grain texture. Each calibration was followed by an inference session where the subject played the game. The game consisted of two separate tasks: in the first one, the subject had to hit moving aliens using the color matched target; in the second one, they need to follow a randomized sequence of colors by hitting the corresponding targets to unlock a door.

Calibration sessions consisted of 60 trials on a single training target, for a total of 600 events (60 trials x 10 unique targets). This means that by default, the data is unbalanced, with 60 target events and 540 non-target events per session. In inference sessions, the number of events varied depending on the subject’s performance

A total of 16 subjects took an optional free-play run after the main protocol, where they could choose their favorite texture and play the game again. The extra sessions include a calibration and can be included by setting the extra_runs parameter. The full protocol is described in [1]. A study on cross-subject decoding using this dataset is presented in [2].

A total of 2 subjects (15 and 18) did not complete the full protocol and are excluded by default.

For many subjects, multiple calibration attempts were not successful, but they can be included by setting the exclude_failed parameter to False.

Inference sessions can be included along with calibration by setting the include_inference parameter to True, but the triggers only indicate the stimulus onset and not ground truth labels.

You can test the dataset with the following code:

Examples

Loading the dataset and then create a subset of the available sessions and runs for subject P01 and P02: >>> paradigm = P300(resample=128) >>> dataset = RomaniBF2025ERP(include_inference=True, exclude_failed=False) >>> subset = paradigm.get_data(dataset, [0, 1])

Expected output: Sessions for subject 2: [‘0grain’, ‘1cb’, ‘2cbExtra’]

References

[1] (1,2)

M. Romani, D. Zanoni, E. Farella, and L. Turchet, “BrainForm: a Serious Game for BCI Training and Data Collection,” Oct. 14, 2025, arXiv: arXiv:2510.10169. doi: 10.48550/arXiv.2510.10169.

[2]

M. Romani, F. Paissan, A. Fossà, and E. Farella, “Explicit modelling of subject dependency in BCI decoding,” Sept. 27, 2025, arXiv: arXiv:2509.23247. doi: 10.48550/arXiv.2509.23247.

Code Examples

from moabb.datasets import RomaniBF2025ERP
dataset = RomaniBF2025ERP()
data = dataset.get_data(subjects=[0])
print(data[0])

Metadata

Dataset summary

#Subj	22
#Chan	8
#Trials / class	540 NT / 60 T
Trials length	1 s
Freq	250 Hz
#Sessions	up to 3

Participants

• Population: healthy

• Age: 21.87 years

• BCI experience: naive

Equipment

• Amplifier: g.tec Unicorn Hybrid Black

• Electrodes: EEG

• Montage: standard_1020

• Reference: right mastoid

Preprocessing

• Data state: raw

Data Access

• DOI: 10.48550/arXiv.2510.10169

• Data URL: https://zenodo.org/records/17225966

• Repository: GitHub

Experimental Protocol

• Paradigm: p300

• Tasks: Complex Task (5 colored laser beams), Speller Task (10 color targets)

• Feedback: visual

• Stimulus: flickering

Notes

Notes

• EEG signals were recorded using a g.tec Unicorn with a sampling rate of 250 Hz and conductive gel applied.

• Data were collected in Trento, Italy, where the power line frequency is 50 Hz.

• EEG was recorded from 8 scalp electrodes according to the international 10–20 system: “Fz”, “C3”, “Cz”, “C4”, “Pz”, “PO7”, “Oz”, “PO8” EEG signals were referenced to the right mastoid and grounded to the left mastoid.

• Events for the calibration are encoded as follows:

  • 1: Target

  • 2: NonTarget

• For inference sessions, events only indicate stimulus onset without ground truth labels (from 1 to 10).

Subjects indexing: {0: ‘P01’, 1: ‘P02’, 2: ‘P03’, 3: ‘P04’, 4: ‘P05’, 5: ‘P06’, 6: ‘P07’, 7: ‘P08’, 8: ‘P09’, 9: ‘P10’, 10: ‘P11’, 11: ‘P12’, 12: ‘P13’, 13: ‘P14’, 14: ‘P16’, 15: ‘P17’, 16: ‘P19’, 17: ‘P20’, 18: ‘P21’, 19: ‘P22’, 20: ‘P23’, 21: ‘P24’}

__init__(data_folder: str | None = None, subjects: List[str] | None = None, exclude_subjects: List[str] | None = ['P15', 'P18'], calibration_length: int = 60, n_targets: int = 10, t_target: int = 1, nt_target: int = 2, interval: tuple = [-0.1, 1.0], extra_runs: bool = True, include_inference: bool = False, load_failed: bool = False, montage: str = 'standard_1020', sessions=None)

Initialize the Brainform MOABB dataset.

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)

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.

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])

See also

CacheConfig

Cache configuration for get_data().

moabb.datasets.bids_interface.get_bids_root

Return the BIDS root path.

Notes

Added in version 1.5.

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

Return the path to the dataset. Required abstract method from BaseDataset.

Parameters:

• subject (int) – Subject identifier

• path (str, optional) – Not used for this dataset

• force_update (bool) – Not used for this dataset

• update_path (bool) – Not used for this dataset

• verbose (bool) – Not used for this dataset

Returns:

Path to the dataset folder

Return type:

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.

Parameters:

• subject_list (list of int | None) – List of subjects id to download, if None all subjects are downloaded.

• 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.

• accept (bool) – Accept licence term to download the data, if any. Default: False

• 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) → None | DataFrame

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 | pd.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

BaseDataset.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)

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 (Pipeline | None) – Optional processing pipeline to apply to the data. To generate an adequate pipeline, we recommend using moabb.utils.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.

Returns:

data – dict containing the raw data

Return type:

Dict

property metadata: DatasetMetadata | None

Return structured metadata for this dataset.

Returns the DatasetMetadata object from the centralized catalog, or None if metadata is not available for this dataset.

Returns:

The metadata object containing acquisition parameters, participant demographics, experiment details, and documentation. Returns None if no metadata is registered for this dataset.

Return type:

DatasetMetadata | None

Examples

>>> from moabb.datasets import BNCI2014_001
>>> dataset = BNCI2014_001()
>>> dataset.metadata.participants.n_subjects
9
>>> dataset.metadata.acquisition.sampling_rate
250.0