Our goal in this project is to reveal the organization of functional networks in the neocortex at very early stages in development and to dissect the circuit motives that produce these early patterns of neural activity. To accomplish these goals, we combine computational modeling and advanced data analysis tools with highly sensitive calcium imaging and pharmacological and optogenetic methods for up-and-down-regulating specific circuit components. This is in tight collaboration with the experimental lab of Gordon Smith at the University of Minnesota, Department of Neuroscience.  

Previous work: Smith, Hein, Whitney, Fitzpatrick, Kaschube, Nature Neuroscience, 2018

The overarching goal of the project is to identify cellular interactions that govern the growth of a body-wide array of chromatophores on the skin of cuttlefish used for camouflage. In this project we develop an accurate dynamical model of the growing chromatophore layout spanning multiple scales, from cellular description to the whole organism. We explore these questions of growth with a reaction-diffusion type model that involves lateral interactions, which are hypothesized to affect both the time of insertion of new chromatophores and the time at which inserted chromatophores transition in color. In parallel, we develop deep learning-assisted predictive growth models that will inform the dynamic model.

Previous work: Reiter et al., Nature, 2018

Brain development is often conceptualized by the relative roles of experience vs. experience-independent factors, but their interface has received less attention. A coarse layout of neural circuits is established early in life by experience-independent factors that involve molecular cues and endogenous neural activity. These coarse networks are refined considerably when the patterns of neural activity shift at the onset of sensory experience. However, a profound understanding of this process of refinement is still lacking, and so is basic knowledge about the relationship between the endogenous and sensory driven patterns of neural activity. In this project we explore this relationship through dynamic network models and machine learning assisted analysis of both endogenous and stimulus evoked patterns of neural activity in the early developing visual cortex.

Previous work: Smith, Hein, Whitney, Fitzpatrick, Kaschube, Nature Neuroscience, 2018