Research in the laboratory of Konrad Kording uses ideas from economics to address problems of relevance to neuroscience and rehabilitation. In economics it is often assumed that agents in the market make efficient decisions. Similarly we may assume that the nervous system efficiently solves the problems that it encounters during everyday life, after all we may argue that the purpose of our nervous system is to allow us to thrive in our environment.

Bayesian decision theory is the systematic way of calculating how the nervous system may make good decisions in the presence of uncertainty. We live in an uncertain world and each decision may have many possible outcomes and choosing the best decision is thus complicated. Our laboratory builds new algorithms to deal with uncertainty and analyzes how people deal with uncertainty.

We use theory as well as computational and neural modeling to understand how information is processed in the nervous system, explaining data obtained in collaboration with electrophysiologists and in psychophysical experiments. One of the central objectives of our laboratory is to improve rehabilitation through better understanding of the economical principles underlying human movement.

Bayesian inference of functional connectivity from spikes

Using current multi-electrode recording techniques electrophysiologists can study the simultaneous spiking behavior of hundreds of neurons. With Lee Miller at Northwestern and Nicho Hatsopoulos at University of Chicago we are working on methods for estimating how these neurons interact with one another in motor cortex. Our current approach uses Bayesian inference and unsupervised learning techniques. These methods should provide useful new tools for studying structure and plasticity.
Ian Stevenson

Normative studies of human search strategies

We use experimental techniques borrowed from psychophysics studies of motor control and motor learning to test human search behavior. Our theoretical approach is to use statistical decision theory to develop normative models of human search strategies. These models render both testable predictions of search behavior and a useful framework for describing search strategies. More...
Gregory Dam

Data sharing in motor control

We are currently proposing to the NSF CRCNS data sharing initiative to share data and models about reaching across many labs. We are looking for labs who would like to participate in this initiative. This initiative (from our perspective) promises to allow parts of motor control to be even more collaborative, quantitative and to design more useful experiments.
Data Sharing Initiative

website tracker

Unless otherwise stated, the content of this page is licensed under Creative Commons Attribution-ShareAlike 3.0 License