Drori 2026

MATLAB code for reproducing the analyses, figures, and tables for Drori et al. 2026:

Drori E, Kurer N, Mezer AA.
Sensorimotor basal ganglia circuit asymmetry explains lateralized motor dysfunction in early Parkinson’s disease.
bioRxiv. 2026. https://doi.org/10.64898/2026.03.16.711841

The repository assumes that PPMI preprocessing and ROI-statistic extraction have already been completed, using the python-based package https://github.com/eliord/drori_ppmi_prep. The main entry point is:

run_drori2026

Public Release Setup

1. Clone this repository and add it to your MATLAB path by running run_drori2026 from the repository root.
2. Install the required MATLAB dependencies listed below, especially mrGrad.
3. Preprocess the PPMI data with drori_ppmi_prep: https://github.com/eliord/drori_ppmi_prep
4. Edit config/drori2026_config.m so local data roots and dependency paths point to your machine.
5. Run run_drori2026 from MATLAB.

The repository does not include PPMI data, preprocessed images, segmentations, ROI tables, or generated analysis outputs.

Repository Layout

run_drori2026.m                 Main reproduction script
config/                         Project configuration and local paths
cohort/                         Cohort filtering and demographics
features/                       Feature-table and histogram generation
io/                             Loading ROI statistics and mrGrad outputs
analysis/                       Statistical analyses and cached result generation
figures/                        Manuscript figure and table scripts
visualization/                  Plotting and image-rendering helpers
utils/                          General helper functions

Required Data

The code expects a local PPMI cohort directory with this general structure:

PPMI_<cohort>/
  PPMI_analysis/
  cohort_tables/
    <cohort>_cohort_meta.csv
    <cohort>_cohort_calc_motor.csv
    <cohort>_cohort_calc_disease_duration.csv
    <cohort>_cohort_moca.csv
    <cohort>_cohort_gba1_lrrk2.csv
    <cohort>_cohort_datscan.csv
    <cohort>_cohort_imaging_QA.csv
    mri/

Subject-level image paths are configured relative to each subject/session folder under PPMI_analysis/. See config/drori2026_config.m for the expected T1w, T2w, PDw, segmentation, ROI-statistic, and mrGrad paths.

The group_analysis/ directory is used for generated analysis outputs. In particular, group_analysis/mrGrad/ is created/populated when mrGrad gradient extraction is run.

Before running the pipeline, edit config/drori2026_config.m so that:

• cfg.dataset.dir points to your local PPMI cohort directory.
• cfg.code.* points to locally installed external MATLAB packages, if needed.
• Example subject IDs and visualization paths match data available on your system, if figure insets are enabled.

Dependencies

Required:

• MATLAB with Statistics and Machine Learning Toolbox
• Image Processing Toolbox
• mrGrad: https://github.com/MezerLab/mrGrad

Optional:

• boundedline for some plot visualizations: https://www.mathworks.com/matlabcentral/fileexchange/27485-boundedline-m
• ComBat for harmonization validation analyses: https://github.com/Jfortin1/ComBatHarmonization
• VistaSoft, AFQ, FSL, and MRtrix3 for optional cortex rendering in Figure 2

The optional Figure 2 cortex rendering is controlled in run_drori2026.m by:

CortexRender = 0;

Set it to 1 only if the rendering dependencies are available.

Running

From MATLAB, run:

run_drori2026

The script adds this repository and configured dependencies to the MATLAB path, then runs the manuscript analyses in order.

Useful switches near the top of run_drori2026.m:

export_figures = 1;
force_reanalysis = 0;
CortexRender = 0;
extra_visualization = 0;

Set force_reanalysis = 1 to regenerate cached .mat and .csv analysis outputs.

Outputs

Outputs are written under:

cfg.outdir

Typical subdirectories include:

analysis_output/
display_items/Figures/
display_items/Tables/
display_items/SupplementaryFigures/
display_items/SupplementaryTables/

Notes

• This code is project-specific reproduction code, not a general-purpose toolbox.
• Preprocessing is not performed by this repository; it assumes the required NIfTI files, segmentations, ROI statistics, cohort tables, and mrGrad outputs are already present or can be generated from the configured inputs.

Citation

If you use this code, cite:

Drori E, Kurer N, Mezer AA.
Sensorimotor basal ganglia circuit asymmetry explains lateralized motor dysfunction in early Parkinson's disease.
bioRxiv. 2026. https://doi.org/10.64898/2026.03.16.711841

Please also cite PPMI and any external tools used in your run, including drori_ppmi_prep, mrGrad, ComBat, AFQ, VistaSoft, FSL, or MRtrix3 where applicable.