Source codes used for data analysis in Cerebellar neural populations orchestrate dopamine reward signaling with single-trial precision, including MATLAB and Python scripts for DCN-VTA fiber photometry analysis, single neuron reward detection, IPS detection, bursting detection, spike-phase coupling, and SVM (python script). Additionally, R scripts were used for modeling trial-by-trial Q values and RPE in the reward foraging task. README: 1. Dataset structure: Raw data and code are organized in the following folder hierarchy: f02_XXX/. % (figure number_experiment) ├── A01_XXX.m % main analysis scripts (A01, A02...) └── Example/ ├── XXX.mat % demo dataset └── ... Each folder corresponds to one analysis module for figure. 2. System Requirements: Hardware We test codes on windows 11: CPU:Intel i9-12900KF, RAM:128GB, GPU(if needed):NVIDIA RTX 3070 Ti Software a. MATLAB:R2020b b. R version 4.4.0 with RStudio (for behavioural modelling only) c. Python 3.7 via Anaconda Navigator (for DCN and VTA single unit decoding). Dependencies: Numpy, Scipy, Pandas, Matplotlib, sklearn, HappyML(https://github.com/cnchi/HappyML) No non-standard hardware is required. 3. Setup and steps: a. Download the source code from: Code: https://doi.org/10.5281/zenodo.19331923 b. Unzip packages and place them in a working directory (example data are already inside). c. Launch MATLAB (or Rstudio/Spyder, depending of pipeline) d. Ensure the supporting functions are added to the MATLAB path. e. Open the main script: A01_XXX.m d. Run the script. The script will automatically load example files from the Example folder and perform all the processes. Typically, you don't need to change any parameters. 4. Output: The script will produce corresponding XXX.mat file, either in the same directory as the script, or in an Output folder. 5. (Optional) Reproducibility: You can find all dataset in https://doi.org/10.5281/zenodo.19345221