Optimizing Computational Neuronal Models
The underlying biophysical properties of cortical neurons such as different ion channels distribution determine their behavior. Currently, neuroscientists do not have a method to accurately determine the distribution of different ion channels in different parts of a neuron. This creates a big gap in understanding neuronal mechanisms in health and diseases. In this project we are developing a framework for predicting a neuron’s ion channels distribution by fitting models to empirical voltage recordings of cortical neurons. However, the immense computational complexity in addition to the nonconvex nature of this fitting problem makes it very difficult to generate highly accurate models. Using our framework, we were able to constrain simple models very accurately and we are now testing on more complex models. We are mainly focused in studying how mutations found in autism spectrum disorders affect the neuronal functions of individual neurons in the cortex of mice.
Message to Sponsor
- Major: Electrical Engineering & Computer Sciences
- Mentor: Roy Ben-Shalom, Kevin Bender | Neuroscience at UCSF