Crowding Conservation

This repository contains the code to reproduce Figures 3A, 3B, 4, and 5 from the manuscript:

Human V4 size predicts crowding distance
Jan W. Kurzawski, Brenda Qiu, Noah Benson, Denis Pelli, Najib Majaj, Jonathan Winawer

Overview

The code in this repository is designed to:

• Recreate key figures from the study.
• Analyze the relationship between human V4 size and visual crowding distance.

Figures

• Figure 3A & 3B: Visualize two example subjects
• Figure 4: Provides test-retest metrics across subjects.
• Figure 5: Examines predictive models for crowding based on V4 size.

Data

The data/ directory contains the following structure:

data/
├── crowdingData/
├── surfaceData/
├── jov2021.mat
└── surface_area_data.csv

Raw fMRI data is available in BIDS format on OpenNeuro: OpenNeuro Dataset ds005639 (Version 1.0.0).

• crowdingData/: Contains data related to crowding measurements.
  • CriticalSpacing_sub-wlsubj044_Ecc=10_Ses=1.mat: Contains critical spacing data for subject wlsubj044 at an eccentricity of 10 degrees during session 1. Critical spacing refers to the minimum spacing at which targets can be distinguished without interference from neighboring stimuli.
• surfaceData/: Contains V1-V4 surface area measurements.
  • lh.R1_rois_sub-wlsubj044: ROI (Region of Interest) data for the left hemisphere from researcher 1 (R1). There are two researchers overall.
  • lh.surface_sub-wlsubj044_midgray: Surface area data of the midgray surface for subject wlsubj044.
• jov2021.mat: MATLAB data file used for the preliminary analysis. It can be used with makeFig5.m
• surface_area_data.csv: CSV file with surface area data for subjects.

Extra Functions

The extra/ directory contains additional functions used in the analysis:

extra/
├── fread3.m
├── mycmap.mat
└── read_curv.m

• fread3.m: Function to read binary files in a specific format.
• mycmap.mat: Colormap data used for visualizing results.
• read_curv.m: Function to read curvature data from surface files.

How to Use

1. This repository is a standalone MATLAB project and does not require additional files or software.
2. Clone this repository:

   git clone git@github.com:jk619/crowdingConservation.git

3. Run the provided scripts to generate the figures.

Citation

If you use this code, please cite the original manuscript:

Kurzawski, J., Qiu, B., Benson, N., Pelli, D., Majaj, N., & Winawer, J. (Year). Human V4 size predicts crowding distance.

Contact

For questions or issues, please contact Jan Kurzawski or open an issue on this repository.

Poster

You can view or download the poster here: Download VSS 2024 Poster