Independent Basal Ganglia Neural Populations Encode Speech Production and Ambient-Noise Levels

This repository contains analysis code supporting the manuscript:

Independent Basal Ganglia Neural Populations Encode Speech Production and Ambient-Noise Levels
by Latané Bullock, Matteo Vissani, Alan Bush, Jackie S. Kim, Lori L. Holt, Julie Fiez, Robert S. Turner, Todd M. Herrington, Jeffrey Schweitzer, Frank H. Guenther, R. Mark Richardson, submitted to Nature Communications

You can find the manuscript here: https://www.biorxiv.org/content/10.1101/2025.11.18.689157v1.

This README file was partially generated by Claude Sonnet 4.6 and subsequently reviewed and edited by the authors.

---

1. System Requirements

Operating systems

The analysis pipeline has been tested on:

• macOS (Ventura / Sonoma)
• Windows 10 (via mapped network drives)

Linux compatibility is expected but has not been formally tested.

R version

• R ≥ 4.2 (tested on R 4.2–4.4)

Required R packages

• tidyverse
• glue
• readr
• magrittr
• jsonlite
• RColorBrewer
• lme4
• lmerTest
• ggpubr
• plotly
• broom
• arrow
• testit
• boot

Non-standard hardware

• No specialized hardware is required.
• ≥32 GB RAM recommended for large parquet files.

---

2. Installation Guide

Instructions

1. Install R from https://cran.r-project.org
2. Clone this repository:

   git clone https://github.com/Brain-Modulation-Lab/Paper_Lombard-ambient-noise-GP-STN.git

3. Install required packages:

   install.packages(c(
     "tidyverse","glue","readr","magrittr","jsonlite",
     "RColorBrewer","lme4","lmerTest","ggpubr",
     "plotly","broom","arrow","testit","boot"
   ))

Typical install time

• Package installation: 5–15 minutes
• Repository setup: <1 minute

---

3. Demo

Main analysis script

20230524-subctx-lfp-group-PLB/
└── A04v5_analyze_all_trials_bandpower_table.R

Instructions to run

• Sections must be run and hand-tweaked according to the desired figure output.

Expected output

• Figures and summary tables will be saved to the ./fig/ directory

Expected runtime

• Demo data: 5–15 minutes
• Full dataset: 1–3 hours (depending on permutation settings)

---

4. Instructions for Use

Neural LFP analysis (main)

20230524-subctx-lfp-group-PLB/
└── A04v5_analyze_all_trials_bandpower_table.R   ← group-level bandpower & figures

1. Update paths and parameters in the main script:
   • PATH_ROOT
   • TASK
   • SESSION
   • MODALITY
2. Ensure required parquet and annotation files are present.
3. Run the script as above.

Behavioral analysis — Lombard effect

20210922-beh-lombard-effect-PLB/
└── B01_group_analyze.R   ← group-level Lombard effect statistics and figures

Run B01_group_analyze.R after ensuring group-level behavioral tables are present.

Behavioral analysis — speech timing overview

20220922-behavior-overview-PLB/
├── B00_combine_tables.R              ← combine per-subject behavioral tables
├── B01_trials_sentences_distribution.R  ← trial and sentence count distributions
├── B02_plot_sp_duration.R            ← sentence duration plots
└── B03_plot_word_by_word_durations.R ← word-level duration plots

Run scripts in order (B00 → B01 → B02 → B03). B00 must be run first to produce the combined table consumed by downstream scripts.

---

Contact

Please contact the corresponding author listed in the manuscript for questions regarding code or data access.