EMFit
EMFit is a matlab toolbox for hierarchical fitting of computational / RL models to trial-by-trial decision data. It provides an extensive set of example likelihood functions for several different tasks, and extensive tools for model validation and comparison. The code is open-source and hosted on github.
EMFit page on githubTCPW
The Transcontinental Computational Psychiatry Workgroup hosts a regular online series of talks on computational psychiatry. We encourage speakers to combine a tutorial part on their methods with example applications in the domain of mental health.
TCPW WebsiteInterested in computational psychiatry?
If you are interested in computational psychiatry, you would greatly benefit from both mathematical / theoretical and clinical expertise. A good starting point would be:
General introduction
Clinical
- Basic understanding of major diagnoses
- Some understanding of the lived experience, e.g. through clinical shadowing / experience, or from reading autobiographies such as
- Kay Jamison – An Unquiet Mind
- Elyn Saks – The Centre Cannot Hold
- Basic understanding of psychotherapy
- Basic understanding of pharmacotherapy
Theoretical / Mathematical
You need some experience with and understanding of differential equations, linear algebra, Bayesian inference and mathematical statistics and reinforcement learning:
- General textbook: Riley, Hobson and Bence – Mathematical Methods for Engineering and Physics
- Linear algebra: Fraleigh and Beauregard – Linear Algebra
- Stats and machine learning: David MacKay – Information Theory, Inference, and Learning Algorithms
- Reinforcement learning: Sutton and Barto – Reinforcement Learning