Navigate Up
Sign In
  • CRM
  • CRM
  • CRM
CRM > English > Research > Research Groups > Computational Neuroscience
Description
Computational neuroscience is a sub-field of neuroscience proper in which computational models are used to learn something about how the nervous system works. It is a broad field, encompassing many different types of models, from statistical or probabilistic models, to differential equations. As most experimental work in neuroscience already requires some degree of modeling, if only at the level of data analysis, there is no clear divide between experimental and computational neuroscience.  This means that close collaboration between theorists and experimentalists is important.  At the very least, modeling work must be constrained by experimental data.

In the Computational Neuroscience group at the CRM, we focus mainly on the dynamics of cortical microcircuits, that is ensembles of hundreds or thousands of neurons in the cerebral cortex.  In particular, we study the role of the recurrent connectivity in shaping spontaneous activity in models of cortical microcircuits.  This is a timely topic because data on cortical connectivity has been increasing over the past decade, as well as improved measurements of the simultaneous activity of large numbers of neurons.  A future goal would be to identify which aspects of network connectivity are most important for cortical processing in models, and then direct experimentalists to look for similar patterns in the brain.  We study models of memory formation and memory consolidation in order to explore the computational limits of biological memory systems and shed light on the physiological mechanisms involved in memory in the animal brain.

We also work on developing computational models  of cortical circuits to shed light on the neural network dynamics  underlying an animal's behavior during elementary cognitive tasks such as working memory and perceptual decision making.  Modelling efforts  are complemented by analysis of typically high-dimensional neural data  obtained by collaborators (e.g. simultaneous recordings from large populations  of neurons or human neuroimaging data) involving state-of-the-art statistical  and machine learning tools.

The Computational Neuroscience Group was started in May 2012. In May 2017, Klaus Wimmer joined the group as a Ramón y Cajal researcher and co-PI. His focus is on investigating the neural network dynamics  underlying elementary cognitive functions such as working memory and perceptual decision  making. He complements the study of neural network models with analysis of experimental  data (obtained from neural recordings in behaving primates and from human neuroimaging in  the laboratories of experimental collaborators). In particular, he will work on extending  current local-circuit computational models of decision making processes towards a network  of interacting circuits that will allow to study the contribution of different brain  areas in the parietal and prefrontal cortex to decision build-up and memory maintenance.