night-owls

Repository for the Night Owls Scan Club (NOSC) Project under Temple University IRB #24452.

NOSC is a multiband, multi-echo, intensively sampled fMRI study of the reward response.

NOSC is described in detail in Mattoni et al., 2025 https://www.biorxiv.org/content/10.1101/2025.09.26.678878v1.

Study design summary

For data use please cite Mattoni et al., 2025 (currently preprinted at https://www.biorxiv.org/content/10.1101/2025.09.26.678878v1) and the OpenNeuro dataset: https://openneuro.org/datasets/ds006707.

BIDS data are publicly available at: https://openneuro.org/datasets/ds006707

Information on scanning sessions, behavioral data, and outputs of L1 and LSS models (described in Mattoni et al., 2025) are available on OSF at: https://osf.io/df5t9/.

---

Preprocessing

Data were preprocessed in an HPC environment.

All preprocessing code is in /code.

Before upload to OpenNeuro:

• Raw DICOMS were BIDS-fied using prepdata.sh.
• events.tsv files were generated using events_generation.R and convertSharedReward2BIDSevents.m.

Field map generation

• warpkit-hpc.sh generates fieldmap files in /bids/sub-xx/ses-xx from multi-echo data.
• addIntendedFor_fieldmap-hpc.py edits .json files to include IntendedFor fields for the generated fieldmap files.

FMRIprep

fMRIPrep was run in a 2-step process to:

• Create a single anatomical image per subject.
• Avoid processing multiple sessions in parallel.

Scripts:

• fmriprep-hpc-anat.sh performs anatomical-only preprocessing (--anat-only, --longitudinal).
  • Creates one T1w image for all sessions per subject.
• gen_fmriprep-anat.sh creates functional fMRIPrep commands for each session.
  • Uses preprocessed anatomical data as an existing derivative.
• run_fmriprep_qsub submits all fMRIPrep commands created in /code/fmriprep-anat/.
• fmriprepOrganize.sh organizes fMRIPrep output in BIDS format.
  • Removes intermediate files from /derivatives/anat-only/.
• normEcho2.sh and smooth-3dBlurToFWHM normalize and smooth single and multi-echo data
  • resample_sub-101_ses-03_task-mid resamples one session to account for scanning error

Tedana

• tedana-hpc.sh estimates tedana confounds for fMRIPrepped data.

---

FSL Analyses

• gen3colfiles.sh converts events.tsv files into FSL-compatible events files.
• MakeConfounds.py adds fMRIPrep confounds to /derivatives/fsl/confounds_tedana.
• genTedanaMultiSes.py adds selected tedana confounds to /derivatives/fsl/confounds_tedana.

Model estimation:

• L1-stats-loop.sh estimates L1 models for each run using combinations of:
  • Space: MNI vs T1w
  • Echo: echo-2 vs multi
  • Confounds: base fMRIPrep vs fMRIPrep + tedana
• L1statsSingleTrial-${task}.sh runs LSS models for the respective task (mid or sharedreward).

Data extraction:

• extractData.sh and extractData-LSS.sh return derivatives used in Mattoni et al., 2025.