Reproducibility Report for Study Frontostriatal salience network expansion in individuals in depression by Lynch et al., (2024, Nature)

Author

Reproducibility Project Eugenia Giampetruzzi (eugiampe@stanford.edu)

Published

October 21, 2025

Introduction

The original study used precision functional mapping to generate participant-specific brain network parcellations. This method combined multi-echo fMRI preprocessing, surface-based connectivity analysis, and Infomap community detection. In multiple samples of depressed patients, the authors found that salience network expansion is a trait-like topological feature in adults with depression.

This report outlines a compuational reproduction of the method in Lynch et al., for generating participant-specific network parcellations and quantifying salience network cortical size. Using publicly available data from a healthy control participant in the Weill Cornell multi-echo dataset (OpenNeuro ds005118), I will reproduce their preprocessing and mapping workflow (HCP surface pipelines, ME-ICA/tedana denoising, surface-based functional connectivity, Infomap community detection, and algorithmic assignment to canonical networks). I will then compare my outputs directly to the published results for this participant.

I will follow the authors’ publicly available code for preprocessing multi-echo fMRI data (https://github.com/cjl2007/Liston-Laboratory-MultiEchofMRI-Pipeline) and performing the precision functional mapping & network size calculations (https://github.com/cjl2007/PFM-Depression). The pipeline requires a number of external neuroimaging tools (Freesurfer, FSL, Connectome Workbench, Infomap, Matlab , tedana Python). I will document the pipeline and results in this .rmd file, and upload .txt and .py with additional detail.

Justification for choice of study

My FYP will apply precision functional mapping to identify trait-like brain network topology in a longitudinal sample of emotionally dysregulated adolescents.

Reproducing the methods from this study gives me an opportunity to 1) correctly configure the many necessary neuroimaging tools, 2) practice implementing the preprocessing and mapping pipeline and, 3) learn how to interpret key metrics such as salience network cortical size.

Anticipated challenges

Do you anticipate running into any challenges when attempting to reproduce these result(s)? If so please, list them here.

Potential challenges: 1) Installing and configuring all of the tools the pipeline + downloading huge CIFTI files on my very old macbook air with unclear OS compatability & limited storage, 2) I’ve only ever worked with single-echo rsfmri data, have to learn extra steps for multi-echo (echo combination, ME-ICA denoising), 3) I’ve never used connectome workbench or infomap which require specialized command-line syntax,

Methods

Description of the steps required to reproduce the results

Please describe all the steps necessary to reproduce the key result(s) of this study.

  1. Data Acquisition A. Download the multi-echo resting-state fMRI dataset from OpenNeuro (ds005118) for sub-ME01. B. Obtain accompanying metadata (task design, echo times, field maps, etc.) necessary for multi-echo combination and denoising.

  2. Pre-processing A. Anatomical: Use the HCP minimal pre-processing pipelines to perform bias correction, skull stripping, registration to MNI, and surface reconstruction with FreeSurfer B. Functional: Perform denoising using ME-ICA via the tedana Python package. Apply motion correction, temporal filtering, and smoothing. Project pre-processed data onto the cortical surface (CIFTI) using Connectome Workbench. Concatenate runs and apply temporal scrubbing (tmask) to remove motion-contaminated volumes.

  3. Functional Connectivity Mapping A. Compute vertex-wise functional connectivity matrices by correlating each cortical vertex’s time series with every other vertex. B. Fisher’s r to z to normalize the correlation values.

  4. Community Detection (Precision Functional Mapping) A. Run Infomap community detection on the individual’s connectivity matrix to identify functional communities. B. Assign each community to a canonical resting-state network (e.g., salience, DMN, FNP) based on spatial overlap with the Yeo 7-network template.

  5. Network Size Quantification Compute the surface area of each identified network on the individual’s cortical surface. Make SN size a percentage of total cortical area.

  6. Comparison to Reference Map A. Original author-provided reference parcellation (Bipartite_PhysicalCommunities+FinalLabeling.dlabel.nii) for the same participant. B. Compare the reproduced parcellation with the reference map using spatial correlation and quantitative size metrics.

Differences from original study

Explicitly describe known differences in the analysis pipeline between the original paper and yours (e.g., computing environment). The goal, of course, is to minimize those differences, but differences may occur. Also, note whether such differences are anticipated to influence your ability to reproduce the original results.

  1. Original used linux-based HPC cluster, I’m using a macOS local machine. It may impact run time.
  2. Using most recent FSL, FreeSurfer, and MATLAB versions for 2025. Study was in 2023.
  3. Going to have to manually modify variables and macOS command-line paths. Their environment was configured for Cornell HPC.

Project Progress Check 1

Measure of success

Please describe the outcome measure for the success or failure of your reproduction and how this outcome will be computed.

  1. Topological Agreement: r ≥ 0.9 between my network parcellation map and the author-provided reference map (Bipartite_PhysicalCommunities+FinalLabeling.dlabel.nii).
  2. Cortical surface area of the salience network: percentage (Δ%) of total cortical surface area should differ from by ≤ 0.1 percentage points.

Pipeline progress

Earlier in this report, you described the steps necessary to reproduce the key result(s) of this study. Please describe your progress on each of these steps (e.g., data preprocessing, model fitting, model evaluation).

  1. Data Acquisition: cloned GitHub repositories for preprocessing (Liston-Laboratory-MultiEchofMRI-Pipeline) and analysis (PFM-Depression), downloaded the dataset (OpenNeuro ds005118).
  2. Pre-processing: verified the MATLAB analysis scripts (pfm_tutorial.m) & in process of installing dependencies (Freesurfer, FSL, Connectome Workbench, Infomap, tedana).

*next steps have not been started:

  1. Functional Connectivity Mapping
  2. Community Detection (Precision Functional Mapping)
  3. Network Size Quantification
  4. Comparison to Reference Map

Results

Data preparation

Data preparation following the analysis plan.

Key analysis

The analyses as specified in the analysis plan.

Side-by-side graph with original graph is ideal here

Exploratory analyses

Any follow-up analyses desired (not required).

Discussion

Summary of Reproduction Attempt

Open the discussion section with a paragraph summarizing the primary result from the key analysis and assess whether you successfully reproduced it, partially reproduced it, or failed to reproduce it.

Commentary

Add open-ended commentary (if any) reflecting (a) insights from follow-up exploratory analysis of the dataset, (b) assessment of the meaning of the successful or unsuccessful reproducibility attempt - e.g., for a failure to reproduce the original findings, are the differences between original and present analyses ones that definitely, plausibly, or are unlikely to have been moderators of the result, and (c) discussion of any objections or challenges raised by the current and original authors about the reproducibility attempt (if you contacted them). None of these need to be long.