Cuban Human Brain Mapping Project

Description

The Cuban Human Brain Mapping Project (CHBMP) is an open, multimodal, population-based normative dataset of 282 functionally healthy young and middle-aged adults collected at the Cuban Center for Neuroscience (Centro de Neurociencias de Cuba), Havana, over the 2004—2008 acquisition period (Hernandez-Gonzalez et al., 2011; Valdes-Sosa et al., 2021). The 282 shared participants are a subset of a larger stratified random sample of 2,019 individuals drawn from the La Lisa municipality of La Habana, Cuba. Each participant contributes resting-state high-density EEG, structural and diffusion MRI, a battery of cognitive/psychological tests, and demographic records (age, gender, education, ethnicity, handedness, weight). The project was conceived as a quantitative-EEG normative reference and as a bridge between inexpensive (EEG) and expensive (MRI) neuroimaging, and it is distributed via the MNI/CONP neuroinformatics ecosystem.

Citation

• Paper DOI: 10.1038/s41597-021-00829-7
• Dataset DOI: TODO(confirm)
• Reference: Valdes-Sosa, P. A., Galan-Garcia, L., Bosch-Bayard, J., et al. (2021). The Cuban Human Brain Mapping Project, a young and middle age population-based EEG, MRI, and cognition dataset. Scientific Data 8, 45.

Catalog citation block

Valdes-Sosa, P. A., Galan-Garcia, L., Bosch-Bayard, J., Bringas-Vega, M. L., Aubert-Vazquez, E., Rodriguez-Gil, I., Das, S., Madjar, C., Virtanen, J., Lippé, S., Hernandez-Gonzalez, G., et al. (2021). The Cuban Human Brain Mapping Project, a young and middle age population-based EEG, MRI, and cognition dataset. Scientific Data, 8, 45. doi:10.1038/s41597-021-00829-7.

Hernandez-Gonzalez, G., Bringas-Vega, M. L., Galán-Garcia, L., Bosch-Bayard, J., Lorenzo-Ceballos, Y., Melie-Garcia, L., Valdes-Urrutia, L., Cobas-Ruiz, M., Valdes-Sosa, P. A., & the CHBMP (2011). Multimodal Quantitative Neuroimaging Databases and Methods: The Cuban Human Brain Mapping Project. Clinical EEG and Neuroscience, 42(3), 149—159. doi:10.1177/155005941104200303.

Dataset record: CONP / LORIS, The Cuban Human Brain Mapping Project (EEG, MRI, and Cognition dataset), CC BY-NC-SA. Portal: https://portal.conp.ca/dataset?id=projects/CHBMP. Metadata stub: doi:10.5281/zenodo.3945385.

Download and access

Field	Value
Official source	https://portal.conp.ca/dataset?id=projects/CHBMP
Access class	registration
Size	~27 GB across ~4,521 files (CONP portal)
Format	edf
BIDS	true
Mirrors	https://chbmp-open.loris.ca (LORIS); https://github.com/conpdatasets/CHBMP (DataLad); Zenodo 10.5281/zenodo.3945385 (metadata stub only)

Access steps:

1. Register on the LORIS portal (chbmp-open.loris.ca) and wait for account approval
2. Clone the CONP DataLad dataset (github.com/conpdatasets/CHBMP)
3. Retrieve annexed content with datalad get after LORIS authentication

Source & access

The canonical data descriptor is published in Scientific Data (Valdes-Sosa et al., 2021, doi:10.1038/s41597-021-00829-7; open access at PMC7865011). The dataset itself is hosted as a DataLad/git-annex dataset through the Canadian Open Neuroscience Platform (CONP); see the CONP portal record and the conpdatasets/CHBMP GitHub repository. Actual download requires registration on the LORIS portal at chbmp-open.loris.ca, after which the annexed content is retrieved with datalad get. The CONP portal lists the dataset as ~27 GB across ~4,521 files under a CC BY-NC-SA license. Note that the Zenodo record 10.5281/zenodo.3945385 (“CHBMP data release v1.0”, July 2020) is only a ~183 kB metadata stub (README, DATS.json, logo) and does not contain the imaging data — the imaging data resides in the LORIS/CONP DataLad dataset described above.

Data structure

Data are distributed in BIDS layout (EEG-BIDS + MRI-BIDS), with imaging stored as EDF (EEG) and NIfTI (MRI), accompanied by TSV/JSON sidecars and CSV/XLS tables for demographics and cognition. The original native EEG was recorded in the MEDICID neurometric system’s proprietary *.plg format and converted to EDF/BIDS for release. Subjects are laid out as sub-CBM{XXXXX}/ses-V01/{anat,dwi,eeg,fmap}/. The release is a DataLad dataset (git-annex backend): cloning fetches only pointers, and datalad get pulls the annexed EDF/NIfTI content after LORIS authentication.

License

• Name: CC-BY-NC-SA
• Snippets on this site: no — the site ships no snippet or precomputed product from this dataset; lessons link the official access page.

Acquisition

Field	Value
Device	MEDICID 5 (Neuronic S.A.) with customized 10-10 electrode cap; 64- or 120-channel montage
Device class	research-cap
Channels	120 (registry value: 64-120 (larger montage taken))
Sampling rate (Hz)	200
Online filters	0.5–50 Hz band-pass plus 60 Hz notch
Reference	linked earlobes
Mains frequency (Hz)	60
Paradigms	rest-ec, rest-eo, hyperventilation, recovery
Participants	282 (282 cohort total; 250 with EEG (170 × 64-ch + 80 × 120-ch))
Population	282 functionally healthy adults 18–68; EEG for 250 (170 × 64-ch, 80 × 120-ch)
Clinical groups	none
Sessions	1
Durations	EC ~10 min, EO ~5 min, hyperventilation ~3 min, recovery ~2 min

Acquisition

Resting-state EEG was recorded on the digital MEDICID 5 electroencephalograph (Neuronic S.A.) with differential amplifiers, using a customized electrode cap arranged according to the 10—10 International System. The reference was linked earlobes (not the common average), online filtering was a 0.5—50 Hz band-pass plus a 60 Hz notch, and the sampling rate was 200 Hz (Nyquist 100 Hz) (Valdes-Sosa et al., 2021). Two montages were used: a 64-channel montage (170 subjects) and a 120-channel montage (80+ subjects per the descriptor). Total recording duration was at least ~30 minutes per subject, split across four conditions: eyes closed (~10 min), eyes open (~5 min), hyperventilation (~3 min), and recovery (~2 min). Structural/diffusion MRI was acquired on a 1.5 T Siemens MAGNETOM Symphony: a 3D T1-weighted MPRAGE at 1×1×1 mm and diffusion-weighted imaging (DWI), with B0 fieldmaps.

Participants

N = 282 functionally healthy participants, ages 18—68 years (mean 31.9 ± 9.3). Sex breakdown is 195 males (≈69.1%) and 87 females (≈30.9%) (Valdes-Sosa et al., 2021). The cohort is the shared subset of a stratified random sample of 2,019 La Lisa residents; participants were screened to be neurologically and psychiatrically healthy (“functionally healthy”). Not all modalities are present for every subject: resting-state EEG is available for N = 250 (170 with the 64-channel montage, 80 with the 120-channel montage; descriptor Table 6), MRI for N = 203, MMSE for N = 156, WAIS-III for N = 167, and the computerized go/no-go reaction-time test for N = 56. So the 282 figure is the total cohort, of which 250 have EEG.

Tasks / conditions

The recording is a single resting-state session with four sequential conditions — eyes closed (EC), eyes open (EO), hyperventilation (HV), and recovery — rather than a cognitive task paradigm. The separately administered cognitive battery (MMSE, WAIS-III, go/no-go reaction time) is delivered outside the EEG recording and is shared as tabular/score data, not as time-locked EEG events.

Used on this site

L7.7

Loader

Fetches only the stated subset. Recorded loader: DataLad after LORIS registration — see snippet.

# Registration-gated (LORIS). Cloning fetches pointers only; content needs an account.
datalad install https://github.com/conpdatasets/CHBMP
cd CHBMP
datalad get sub-CBM00001   # after LORIS authentication; one subject only

Caveats worth knowing

• Event markers are in Spanish.
• Channel counts vary per subject (58–121).
• Three subject IDs are missing from the sequence.

Notable caveats

• Reference is linked earlobes, not common average — re-reference explicitly if a common-average or REST montage is required (Valdes-Sosa et al., 2021).
• Online band-pass is 0.5—50 Hz with a 60 Hz notch. The 50 Hz low-pass plus the 60 Hz notch effectively removes most usable signal above ~45—50 Hz, so high-gamma analyses are not supportable despite the 100 Hz Nyquist.
• Variable montage (observed 62/120/121/58—59 ch) — the channel count is not constant across subjects.
• Event markers are in Spanish — e.g. “ojos cerrados” (eyes closed) and “ojos abiertos” (eyes open); pipelines parsing marker labels must match the Spanish strings.
• Three subject IDs are absent from the otherwise sequential numbering: 00039, 00143, and 00251 — iterators assuming a gapless ID range will fail.
• Registration-gated: data are not anonymously downloadable; LORIS account approval is required before datalad get succeeds.
• Zenodo DOI is a stub: 10.5281/zenodo.3945385 is metadata-only; cite the Scientific Data descriptor and the LORIS/CONP record for the actual data.

Related datasets

• ERP CORE (Compendium of Open Resources and Experiments)
• EEG Motor Movement/Imagery Dataset (EEGMMIDB)
• LEMON (MPI-Leipzig Mind-Brain-Body)
• Dortmund Vital Study resting EEG
• SRM Resting-state EEG

# Loader recorded in the catalog: DataLad after LORIS registration — see snippet
# TODO(confirm) the exact call and the subset to fetch. Official source: https://portal.conp.ca/dataset?id=projects/CHBMP