- R 100%
| Filename | Latest commit message | Latest commit date |
|---|---|---|
|
|
||
| 00_paper_functions.R | ||
| README.md | ||
| trial_by_trial.csv.zip | ||
NMDA receptor ablation in medial prefrontal cortex disrupts value updating and reward history integration
This repository contains the basic data and analysis code needed to reproduce the primary findings reported in the manuscript.
Authors: Evan Knep1, Angelica Velosa1, Dana Mueller1, Cathy Chen2, Sophia Vinogradov2, Matthew V. Chafee3, Becket Ebitz4, Sarah Heilbronner5, Patrick E. Rothwell3, Nicola Grissom1
Affiliations
1 Department of Psychology, University of Minnesota, Minneapolis, Minnesota
2 Department of Psychiatry, University of Minnesota, Minneapolis, Minnesota
3 Department of Neuroscience, University of Minnesota, Minneapolis, Minnesota
4 Department of Neurosciences, Université de Montréal, Québec, Canada
5 Department of Neurosurgery, Baylor College of Medicine, Houston, Texas
Repository contents
This repository contains the core data and R functions needed to reproduce the main behavioral, reinforcement-learning, and simulation-based analyses reported in the manuscript.
00_paper_functions.R
Core R functions used for manuscript analyses:
- computing behavioral summary measures from trial-level data
- calculating reward acquisition relative to chance
- calculating optimal choice rate, switching, win-stay, lose-shift, negative outcome weight, mutual information, richness, and Δprobability
- fitting and evaluating the additive-gains RLCK model
- generating RLCK-based simulations and posterior predictive checks
trial_by_trial.csv.zip
Compressed trial-level behavioral dataset used for the main analyses.
Each row corresponds to a single trial from the restless bandit task. The included columns are:
| Column | Description |
|---|---|
animal |
Animal identifier |
virus |
Viral manipulation group, lacz control or grin1 ablation |
sex |
18 male, 16 female |
surgery |
Pre- versus post-surgery phase |
weeks_post_surgery |
Week relative to surgery, animals began behavior on week 2 |
condition |
Experimental/pharmacological condition, no injection, saline, or MK-801 |
schedule |
Task schedule/session identifier (unique task "walk" each day) |
choice |
|
outcome |
|
left_prob |
Reward probability of the left choice on a given trial |
right_prob |
Reward probability of the right choice on a given trial |
The reward probabilities define the restless bandit environment and are used to compute chance reward rate, optimal choice rate, richness, and Δprobability.