Penn LEAD: Penn Longitudinal Executive functioning in Adolescent Development

Penn LEAD is a longitudinal dataset consisting of clinical, cognitive, and multimodal neuroimaging data from 132 adolescents, designed to investigate transdiagnostic executive function during development.

Link to bioRxiv: https://www.biorxiv.org/content/10.1101/2025.11.10.687633v1.full

Raw data

Note that in some cases, there may be more runs than expected or noted in dataset contents below if a run was repeated. If a run was repeated, there might be an extra run in the dataset. Additionally, in cases where the data from one of the runs was poor quality, there might be a second and third run but no first run.

Any variant to the metadata that might differ from what is described below will be flagged in the file name for that scan in the acquisition entity (i.e., sub-*_ses-*_acq-VARIANT{variant_description}run-**.*)

Dataset contents Structural MRI Each participant has one T1-weighted normalized scan (176 slices, TR = 2500 ms, TE = 2.9 ms), one T2-weighted (not normalized) scan (176 slices, TR = 3200 ms, TE = 565 ms), and one T2-weighted normalized scan, all with the AP phase encoding direction. T1 scans are refaced and T2 scans are defaced. The processed sMRI data derivatives can be found here.

Functional MRI Each participant has two resting-state fMRI runs (522 volumes and 383 volumes, TR = 800 ms, TE = 30 ms, 60 slices) and one n-back task fMRI run with fractal images (522 volumes, TR = 800 ms, TE = 30 ms, 60 slices). The processed fMRI data derivatives can be found here and here.

Note that for one subject, they had to re-run the resting-state fMRI, since in the first run, a movie was accidentally playing in the first run. Hence, the task entity was renamed to task-restfilm.

Diffusion MRI One 103-volume multi-shell (MB = 3), single spin-echo DWI run was acquired for all participants using a monopolar sampling scheme (TR = 4200 ms, TE = 89 ms, voxel size = 1.7 mm isotropic). The processed DWI data derivatives can be found here and here.

Perfusion MRI One background-suppressed, unbalanced, 3D stack of spirals PCASL scan was acquired for all participants (labeling duration = 1500 ms, TR = 4298 ms, TE = 10 ms). The processed ASL data derivatives can be found here.

A reference scan was also acquired to assist in ASL calibration, consisting of an M0 scan that is retained in the perfusion folder with the m0scan suffix.

MEGRE We acquired a multi-echo gradient-echo sequence with four echos, for use in quantitative susceptibility mapping (QSM), at 1.5 mm isotropic resolution (TR = 35 ms, voxel size = 1.5 * 1.5 * 1.5 mm).

Note that we did not curate or preprocess the MEGRE data, but raw data can be found in the dataset.

Field maps For distortion correction, pairs of field maps with opposite phase-encoding polarity (anterior to posterior and posterior to anterior) were acquired for both diffusion-weighted and fMRI scans.

Phenotypic Data Clinical diagnostic data can be found in the participants.tsv file.

Cognitive Data Participants completed a number of cognitive tasks from the Penn CNB. These included:

Penn Abstraction, Inhibition and Working Memory Task Penn Face Memory Test Children's Penn Word Memory Test Short Computerized Finger-Tapping Task Penn Emotion Recognition Task for Children Digit Symbol Test Short Letter N-Back (2 back) Measured Emotion Differentiation Test Motor Praxis Test Penn Conditional Exclusion Task for Children Short Penn Continuous Performance Test - Number and Letter Version Penn Matrix Reasoning Test Variable Short Penn Line Orientation Test Children's Verbal Reasoning Test Short Visual Object Learning Test Trailmaking A Trailmaking B Effort Discounting Task Delayed Discounting Task Risk Discounting Task Age Differentiation Test The following subjects/sessions are missing cognitive data:

sub-21161, ses-1 sub-22510, ses-2 sub-22617, ses-2 sub-22618, ses-2 sub-23582, ses-1 sub-20812, ses-2 had a session 1 that was started, but incomplete, and a session 2 that began a couple of days later. As such, we only have CNB data from ses-1, and ses-3, but ses-2 did not have a separate CNB administered. Other data The acquisition date and time for each session was rounded to the 15th of the month and the nearest hour, and stored in the *_sessions.tsv files.

Additional notes Note that sub-20812 has two sessions (ses-1 and ses-2) that are only a few days apart. Ses-1 contains incomplete data and thus the scans were repeated a few days later for ses-2.

Derivatives: ASLPrep:

This repository contains data derivatives from running ASLPrep for perfusion data. We provide derivatives including standard ceberal blood flow maps, SCORE-processed CBF, SCRUB CBF, BASIL CBF, and BASIL CBF with partial volume correction, as well as quality control metrics.

The link to the accompanying raw dataset is here: https://openneuro.org/datasets/ds006688 [accession number ds006688].

The Docker Hub link for ASLPrep is here.

The quality control recommendations for ASL can be found on Github. Both QC recommendations for ASL and T1 structural data should be taken into account for ASL data.

Derivatives: XCP-D

Penn LEAD (Penn Longitudinal Executive functioning in Adolescent Development) is a data resource designed to investigate transdiagnostic executive function during development.

Note that we are running into some technical issues uploading the full XCP-D data to OpenNeuro. Data for all subjects will be uploaded as soon as possible as available. This repository contains data derivatives from running XCP-D. We provide derivatives including post-processed clean BOLD time series that are denoised, filtered, and censored, ALFF and ReHo maps, parcellated time series and functional connectivity matrices calculated from Pearson correlations, as well as quality control metrics and figures.

The link to the accompanying raw dataset is here: https://openneuro.org/datasets/ds006688 [accession number ds006688].

The Docker Hub link for XCP-D is here.

The quality control recommendations for fMRI can be found on Github. Both QC recommendations for fMRI and T1 structural data should be taken into account for fMRI data.

*Note that we manually deleted extra n-back task runs from the data upon consulting the scanner notes after running XCP-D. Hence, the html files for the following subjects that originally had an extra n-back run will incorrectly specify that there were two runs:

sub-20149, ses-1 sub-20212, ses-1 sub-20238, ses-2 sub-20352, ses-2 sub-20934, ses-1 sub-20964, ses-1 sub-20968, ses-2 sub-21035, ses-1 sub-21197, ses-1 sub-21516, ses-2 sub-21553, ses-1 sub-21738, ses-1 sub-23676, ses-1