PreciseKids — Individualized Topography

This repository contains MATLAB, Python, R, and shell scripts supporting the following work:

Individualized network topography in pre-adolescent children and adults using naturalistic precision fMRI

First-author and script creation: Shefali Rai Under Review BioRxiv DOI: https://doi.org/10.64898/2026.03.05.709899

For the individualized network topography analyses using naturalistic precision fMRI data, we have included scripts for the generation of individual-level network maps, comparison with HCP templates, surface area and density maps, and network similarity between adults and children.

---

Repository Structure

PreciseKids_IndividualizedTopography/
├── scripts/        # MATLAB, Python, R, and shell scripts
├── data/           # Supporting data files
└── README.md

---

Requirements

• MATLAB R2021b or later
• Python 3 (for vertex-wise assignment)
• R (for VertexwiseR TFCE analyses and plotting)
• Connectome Workbench (wb_view, wb_command) was used for all cortical surface visualization
• cifti-matlab and gifti-1.6 should be in your MATLAB path

---

Script Descriptions

Network Map Generation

MatchedCensoring_NetworkMaps_Allsubs.m Censors data across all video stimuli and sessions, then randomly draws usable and censored data that is matched in quantity across stimuli and sessions. Note: since sampling is still within-session, this is split-half iterative sampling rather than true test-retest.

createMaxDice_WinnerTakeAll.m Creates a group-average network map using a winner-take-all approach across participants. Used to generate the visual children and adult average maps.

HCP Template Comparisons

HCP_overlapbetweentemplates.m Overlap between the Dworetsky HCP adult template and the HCP 8-9 year old template.

HCP_overlaptemplate_creation.m Overlap map between the two HCP templates for visualization.

HCPOverlap_AllDiceValues_ExtractNetworkValues.m Dice overlap per network for each participant. This becomes the input for VertexwiseR TFCE analyses.

ExtractBinaryNetworks_HCPoverlapnetworks.m Extracts each network column as a binary mask and saves it separately as a .dscalar.nii file.

Network Proportion & Surface Area

NetworkPropDiff_Sample1.m For each vertex, computes the proportion of children and adults assigned to each network, then visualize the difference as a .dscalar map. Uses outputs from EachVertex_ProportionAssignment.py.

EachVertex_ProportionAssignment.py (Python) Child minus adult network assignment proportions at each vertex and outputs a differences.txt file used by NetworkPropDiff_Sample1.m.

SurfaceArea_Gradiation.m Computes surface area and HCP overlap network assignment for each participant to create density maps across all individuals — this is not averaged across participants.

---

Visualization information

• Open .dscalar.nii output files in Connectome Workbench (wb_view)
• We visualize with the inflated.32k_fs_LR.surf.gii for both hemispheres

---

Citation

If you use these scripts, for now please cite our bio archive doi: https://doi.org/10.64898/2026.03.05.709899