The paper describing cpm is now published in PLOS Computational Biology. The toolbox has a peer-reviewed reference — and a way to cite it.

cpm is an open-source Python library for theory-driven modelling in computational psychiatry and cognitive neuroscience. It gives mechanistic, theory-driven modelling the kind of general-purpose, plug-and-play workflow that data-driven work already has: you specify the trial-level logic of your model, and cpm handles the simulation loop, likelihood, parameter estimation, and parallelisation. It covers reinforcement learning, prospect theory, and metacognition models, with hierarchical estimation via empirical and variational Bayes — all built around reproducibility and transparent modelling practices.

If you use cpm in your work, please cite:

Dome L, Hezemans FH, Kadri K, Wagner BJ, Webb A, Hauser TU (2026). cpm: A Python library for theory-driven modelling in computational psychiatry. PLOS Computational Biology 22(7): e1014481. https://doi.org/10.1371/journal.pcbi.1014481

📖 Documentation · 💻 GitHub · 📄 Paper

cpm is actively developed and we welcome contributions — models, methods, docs, and feedback all help the toolbox grow.