Frequently Asked Questions

You can start by learning that:

• The Healthy Brain Network is the Open Science Initiative of the Child Mind Institute. It aims to create a large, transdiagnostic, cross-sectional dataset that captures the diversity of mental health and learning disorders in developmental psychopathology.
• Its sampling strategy focuses on recruiting individuals based on perceived clinical concern, resulting in a high proportion of those with mental health or learning disorders, while also including diverse New York City – as well as some few other US-based - communities.
• HBN provides participants with comprehensive no-cost diagnostic evaluations, treatment recommendations, and access to educational support, offering valuable resources to families while also compensating them for their time and expenses.
• HBN collects data from multiple modalities, including phenotypic data from more than 140 assessments, magnetic resonance imaging (MRI) data, electroencephalography (EEG) data, eyetracking data, actigraphy data, and biological samples. You can refer to the HBN summary infographic at https://fcon_1000.projects.nitrc.org/indi/cmi_healthy_brain_network/index.html

We suggest that you take a moment to:

• Read about the protocols and the available datasets at https://fcon_1000.projects.nitrc.org/indi/cmi_healthy_brain_network/Data.html.
• Carefully review this FAQs, where you may find the answer to your question.
• Visit the HBN support forum at https://www.nitrc.org/forum/?group_id=296

Yes! Neuroimaging data (MRI, EEG, eye-tracking) and minimal demographic information (age, sex at birth, handedness score) is openly available without a Data Usage Agreement (DUA); complete phenotypic data is available after completing a DUA which requires the signature of the principal investigator and an institutional official (see https://fcon_1000.projects.nitrc.org/indi/cmi_healthy_brain_network/Phenotypic.html#DUA).

HBN phenotypic data can be accessed by completing a Data Use Agreement (DUA) and the DUA Form at https://redcap.link/HBN_DUA. After signing the standard DUA, phenotypic data can be accessed through LORIS at https://data.healthybrainnetwork.org/main.php.

Further instructions for this process are available at https://fcon_1000.projects.nitrc.org/indi/cmi_healthy_brain_network/Phenotypic.html#DUA and https://fcon_1000.projects.nitrc.org/indi/cmi_healthy_brain_network/Phenotypic.html#Loris

No, edits to the standard DUA are not permitted. The standard DUA has to be signed as-is to obtain access to the HBN phenotypic datasets.

Both are data management and sharing platforms used by HBN, but since Release 11 (November 2022), we are only maintaining the LORIS instance for phenotypic data, while COINS is kept for historical reasons. Users are strongly suggested using LORIS for phenotypic data download.

The authorized official needs to be a person who is authorized to sign agreements that bind the institution and must be an individual different from the principal investigator. They are typically a member of the Grants and Contracts Office or someone who has the authority to sign legal documents on behalf of the entire department, institution, or organization. This signature is required, and we will not accept a DUA that is missing this signature. Furthermore, their official title has to be specified in the DUA under "Recipient Institution" (e.g., Head of Department, University Research Officer). Titles such as "Professor" are not sufficient, and DUAs that do not include the required institutional designation will not be processed.

Multiple people can be added under the same DUA, if they are part of the same group and same institution. The DUA form will ask to list additional users and their institutional email. Collaborators outside your institution must complete a separate DUA signed by the institutional official of their institution and cannot be added under your DUA. In the DUA form, the PI will be requested to acknowledge that, by requesting such additions, they remain responsible for ensuring that all terms and conditions of this DUA are complied with by each additional user.

Yes, the DUA is non-transferable between PIs and institutions.

Recipients granted permission to access the Biobank Data shall receive an account that is valid for a period of seven (7) years. After 7 years, the LORIS account will be revoked until a new DUA is sent through the DUA form at https://redcap.link/HBN_DUA.

Yes, start-ups and companies can complete a Data Use Agreement (DUA) to gain access to the HBN dataset through the same form at https://redcap.link/HBN_DUA. However, it is important to note that not all participants have consented to their data being used for commercial purposes. Information on commercial use for each participant can be found in the Metadata file and in the “Basic_Demos” instrument on LORIS. To identify participants whose data cannot be used for commercial purposes, you can review the column "Commercial_Use", where "No" indicates "no commerical use allowed". Requesters from non-degree-granting organizations will be asked to provide further information about their organization in the DUA form.

Neuroimaging data can be accessed directly without a Data Use Agreement. Instructions on how to access the data is available at https://fcon_1000.projects.nitrc.org/indi/cmi_healthy_brain_network/MRI_EEG.html#Direct%20Down. To access data through INDI/FCP, users must Register with the 1000 Functional Connectomes Project website on NITRC (see https://fcon_1000.projects.nitrc.org/indi/req_access.html).

The curation of the actigraphy data is currently underway, and we aim to release the data in the next data release.

Voice and video data are not currently released.

At this time, we are working with NIH’s genetic repository (https://www.nimhgenetics.org/, study number 187) to provide blood samples from a subset of participants from HBN. The repository does not perform sequencing; it makes the samples available for other (approved, funded) researchers to do so. We are currently not planning on releasing or running analysis on data for any other sample (hair, saliva, teeth). Those samples will be stored in a repository indefinitely.

Yes, combining datasets is compliant with the DUA if it is strictly for analysis purposes only, but the user is prohibited from openly releasing the combined dataset.

Please send an email to CMIDataUsage@childmind.org, specifying:

• your affiliation
• the reason why you cannot obtain signatures for the standard DUA
• your background
• whether you are part of an education program
• a detailed list of phenotypic datasets you would like to access (see https://fcon_1000.projects.nitrc.org/indi/cmi_healthy_brain_network/Pheno_Protocol.html for details)
• what you aim to investigate, once you access the data
• whether you are under 18 years of age

A complete data download in one click is not currently optimized in LORIS, so users who want to download the complete dataset have to select each assessment one by one. Trying to download the entire dataset at once may result in failure or a pending message. Please note that the Define Filters option is not currently available on this LORIS instance.

Complete demographic data is available on LORIS (e.g., “Basic_Demos” and “PreInt” instruments) and a signed DUA is required to access them. Some demographic data (sex, age, handedness) is also openly available in the Metadata File.

More details can be found as part of the Phenotypic Protocol and data dictionaries can be downaloded at Data Overview.

HBN does not include longitudinal data, only cross-sectional data are collected. However, please note that a small subset of data is collected at multiple time points (e.g., Daily Medications).

Diagnosis information is available on LORIS after completing a Data Usage Agreement (DUA). Download the “Diagnosis_ClinicianConsensus” or “Diagnosis_KSADS” instruments and data dictionaries. The instrument “Diagnosis_ClinicianConsensus” is a combination of the KSADS results as well as information from other assessments. It includes up to 10 diagnoses for each participant, in no specific order. Each diagnosis further includes a level of certainty (e.g., confirmed). The levels of certainty characterize how definitive the diagnosis is based on the Healthy Brain Network evaluation.

Please note that the HBN evaluation is not comprehensive for Autism Spectrum Disorder (ASD).

Levels of certainty: https://fcon_1000.projects.nitrc.org/indi/cmi_healthy_brain_network/Phenotypic.html#Diagnosis

Information on diagnosis process and ASD considerations: https://fcon_1000.projects.nitrc.org/indi/cmi_healthy_brain_network/File/_pheno/HBN_DiagnosticProcess_v3.pdf

Up to 10 diagnoses are reported in no specific order; therefore, Dx_01 is not necessarily the main diagnosis. To pivot the diagnosis data to a more accessible format, you can use the HBN-DDP python package at https://github.com/childmindresearch/hbn-ddp.

Participants with no diagnosis show a 1 (= “No” for the question "Was the participant given a diagnosis?") for the variable “NoDX” in the instrument called “Diagnosis_ClinicianConsensus” on LORIS. They also show "No Diagnosis Given" for specific diagnoses (e.g., DX_01).

Please note that “No Diagnosis Given: Incomplete Eval” means that the participant was not given a diagnosis because the evaluation was incomplete. This is distinct from "No Diagnosis Given", which applies to participants who completed the HBN study and did not receive any diagnosis.

Detailed information is generally found in the test manual, website, or reference paper for each standard clinical assessment. Standardized conversions (like T-scores or standard scores) apply age and/or sex adjustments in accordance with the test manuals. We encourage the user to consult the test manual for standard clinical assessments. If information about scoring or standardization is not included in the data dictionary, the test manual may be proprietary and therefore cannot be shared.

No. To protect participants’ privacy and comply with regulatory requirements, exact dates are not made openly available. Instead, years and seasons are provided, and each assessment includes the variable “Days_Baseline”, which indicates the number of days that have elapsed since the participant’s enrollment in the study.

Sibling information is available on LORIS in the instrument called “Siblings” which lists the siblings of participants. However, this information is not available for every participant, and a more complete sibling dataset is planned for future data releases.

Metadata about phenptypic, EEG, and MRI session completion can be found in the Metadata file, which indicates whether participants completed phenotypic, MRI, and EEG sessions (0 = No, 1 = Yes), includes information on the number of runs or tasks completed, and provides additional metadata for each participant.

You can use the script available here to download MRI or EEG/eyetracking datasets for selected participants.

Yes, HBN MRI and EEG data is available in BIDS format at https://fcon_1000.projects.nitrc.org/indi/cmi_healthy_brain_network/MRI_EEG.html#Direct%20Down

Eye-tracking data is located in the EEG folder, specifically within each participant’s folder in a directory called "Eyetracking".

Currently, structural and functional MRI preprocessed data is available as part of the Reproducible Brain Charts Initiative (https://reprobrainchart.github.io/). This includes structural data preprocessed using Freesurfer and functional data preprocessed using C-PAC. Visit https://reprobrainchart.github.io/docs/get_data to access the preprocessed data. Furthermore, C-PAC offers open access pipelines to process MRI data (https://fcp-indi.github.io/).

Previous efforts of preprocessing MRI data using Mindboggle and CPAC are available in S3 under the directory /fcp-indi/data/Projects/HBN/derivatives/Freesurfer_version6.0.0 and /fcp-indi/data/Projects/HBN/CPAC_preprocessed_Derivatives. While HBN has continued releasing MRI data, this preprocessing effort was terminated earlier than Release 11, thus no more data will be released which were preprocessed through these pipelines. These preprocessed data are also no longer supported, and we recommend using the processed data from the Reproducible Brain Charts Initiative.

The HBN EEG dataset is BIDS-formatted and "analysis-ready". While some preprocessed data exists from an older project, it was processed with project-specific parameters. It is strongly recommended to preprocess the EEG raw data again according to specific research questions.

The scan parameters files are available at https://fcon_1000.projects.nitrc.org/indi/cmi_healthy_brain_network/MRI_Protocol.html#MRIscanparam

Information about the channel locations for EEG data can be found at https://fcon_1000.projects.nitrc.org/indi/cmi_healthy_brain_network/MRI_EEG.html#MRIEEG%20Overview

It is recommended to use the last run acquired (e.g., run-02), as the first run may have been interrupted due to head motion or incomplete scanning.

For phenotypic and EEG data collection, the sites are:

1. Staten Island
2. Mobile Research Vehicle
3. Midtown
4. Harlem
5. Staten Island Rutgers University Medical Center

For MRI data collection, the sites are:

1. Staten Island
2. RUBIC
3. CBIC
4. CUNY

Task-rest refers to resting state, and task-peer refers to the Predictive Eye Estimation Regression protocol (https://fcon_1000.projects.nitrc.org/indi/cmi_healthy_brain_network/MRI_Protocol.html).

Yes, functional connectomes were derived using the Configurable Pipeline for the Analysis of Connectomes (C-PAC). This dataset, and other C-PAC preprocessed derivatives, are available at https://github.com/ReproBrainChart/HBN_CPAC/tree/main/cpac_RBCv0

Please consult https://fcon_1000.projects.nitrc.org/indi/cmi_healthy_brain_network/Citation.html

You can find bookmarked papers at https://fcon_1000.projects.nitrc.org/indi/cmi_healthy_brain_network/Publications.html

Yes! Check out the Open Science Initiative page at https://fcon_1000.projects.nitrc.org/indi/cmi_healthy_brain_network/OpenScience.html. This includes a list of the data science competitions that have been organized over the years using the HBN dataset.