Audrey Sederberg

,

Credentials

PhD

Bio

Audrey Sederberg is an Assistant Professor in the Department of Neuroscience and a member of the Medical Discovery Team in Optical Imaging and Brain Science. Dr. Sederberg develops models and analyses for large-scale neuronal recordings, including two-photon imaging and high-density multi-electrode arrays, to further our understanding of the function of neuronal networks and the mechanisms controlling them. Dr. Sederberg received her Ph.D. in theoretical physics from Princeton University and completed postdoctoral work in neurobiology at the University of Chicago and at the Georgia Institute of Technology. She was most recently an associate research scientist at Emory University in the Theory and Modeling of Living Systems initiative.

Research Summary

Recent advances in high-throughput recording technology enable the acquisition of rich neuronal datasets. Not long ago, recording from a few dozen cells simultaneously was remarkable; now, labs routinely record from thousands of neurons. We also know far more about microcircuit anatomy — both single-cell properties and how they connect in local networks. However, we cannot yet predict from the anatomy of a microcircuit the basic statistics of activity that the circuit generates or how it supports microcircuit computations relevant to behavior. This basic understanding is a prerequisite toward understanding variability in neural circuit activity, including variability arising due to disease states.To this end, the Sederberg group uses a wide range of methods, including dynamical and statistical models, decoding approaches, and ideal observer analyses, to link microcircuit structure to the observed functions (e.g., selectivity for stimulus or choice) of dynamical network activity. We are focused on understanding these dynamics in networks of thousands of neurons. At this scale, it is no longer feasible to synthesize each neuron's activity one at a time, yet the richness of the data permits more subtle analyses than averaging over the entire population. In one set of current projects, we develop models that provide rigorous baseline hypotheses for what we expect to find in large neural datasets. In another set of projects, we develop multi-scale models to extract the individual and common elements of the dynamics of cortical networks in order to use this approach to assess the variability of cortical dynamics observed in different individuals, brain areas, and species and relate this variability to differences in the underlying microcircuit structure. The Sederberg group works closely with experimental labs and welcomes new opportunities for collaboration.