Computational Cognitive Neuroscience Lab
Principal Investigator: Professor Anne Collins
The Computational Cognitive Neuroscience (CCN) Lab is a part of the Department of Psychology and the The Helen Wills Neuroscience Institute at UC Berkeley. We build and test computational models of learning and executive function.
Follow our Twitter: @ccnlab
Find us on Youtube: CCN Lab
We are always looking for new volunteers to contribute to our work. To participate in an experiment in our lab, email firstname.lastname@example.org or click on Contact Us above for more information about the studies we are recruiting for.
Our research focuses on learning and executive function. We believe that the brain's processes for learning and decision making are complex and multi-faceted, more like a symphony played by many musicians, each good at their own instrument, than a melody played by one or two instruments. We ask the following questions to understand how the brain learns and makes decisions:
- Who are the players? What systems in the brain interact to produce our behavior? How can we isolate their contribution to the overall melody?
- What are they good at? Which environments are they useful in, what are they optimized for?
- How do they work? How does the brain implement them? How does this implementation constrain behavior?
- How do they interact? Can we separate the contributions of each player?
To answer these questions, we use three classes of methods:
- Behavioral experiments. Careful experimental design and data analysis are essential to understand the behavior of learning.
- Computational modeling. The players that contribute to learning and decision making are hidden and how they work needs to be inferred. Computational modeling is essential to extracting these hidden processes. Computational modeling lets us precisely and quantitatively define theories, make explicit predictions, and investigate how well different information representations work in different environments. Computational modeling may also provide a link to the mechanistic implementation of processes.
- Neuroscience methods. Learning and decision making happen in the brain, and relating the cognitive processes to the brain's neural systems help us understand both, by providing constraints on possible mechanisms and the information processed by different brain networks.
You can see examples of specific projects tackling these questions, read published papers that provide some answers, and read more about the methods we use.
To watch Professor Collins' latest talk on generalization with hierarchical reinforcement learning for the CCN 2020 GAC Workshops, click here. Links to other lectures you can watch online can be found under Research Projects.