Analyzing Correlations in Neuronal Networks

The activity of neural populations underlies many of the brain's computations. As the simultaneous recording of multiple neurons has become widespread, researchers are able to observe how neurons act in concert during these computations. However, the origin and function of observed correlations of activity across a population remain difficult to determine. The aim of this project is to develop and apply mathematical analyses of correlations within population activity, with the goal of providing a link between the activity and network structure. The first half of the project is to continue development of a method to estimate network connectivity from simultaneously recorded neurons. The unique feature of this analysis is that it explicitly accounts for effects from unmeasured neurons, an important feature since one can simultaneously measure only a small fraction of neurons. The second half of the project is to develop computational tools, based on the kinetic theory approach, to study the propagation of correlations through networks. The tools that result from this research will help neuroscientists analyze the relationship between correlations and neural network structure, providing insight into the circuitry that underlies processing within the brain.

The goal of this work is to develop mathematical tools that neuroscientists can use to determine how neurons in the brain are connected to each other and how those connections influence the activity of the neurons. Understanding how neurons are connected and the consequences of those connections is an important first step toward discovering how neurons communicate with each other to process information. As such knowledge may help scientists better understand the effects of the degradation of neural connections, such as in neurodegenerative disease, it may eventually lead to strategies to mitigate or reverse these effects.