Applied Neuroscience

Reverse Engineering the Brain (Wednesday, October 2, 2013)

Speaker: Dr. Dmitri “Mitya” Chklovskii
Janelia Farm Research Campus
Howard Hughes Medical Institute

The complexity of human behavior is matched by the complexity of biological hardware, which consists of 10¹¹ neurons connected by 10¹⁵ synapses. To understand how the brain performs its function despite rigid constraints on its size and energy consumption, we are using two complementary approaches. First, we are mapping the connectome, or the wiring diagram of the brain, on the synapse level. The main challenge here is to process a large number of high-resolution micrographs to identify individual neurons and synapses. By applying modern computer-vision and machine-learning methods, we have made major progress on this front and have reconstructed a visual motion detection circuit in insects. Second, we are attempting to understand the function of individual neurons from the computational perspective. We have found that neurons specialize in processing sparse signals online, not unlike the mathematical algorithms at the cutting edge of adaptive signal processing. Combining the two approaches will allow us to model neuronal computation and build biologically inspired artificial intelligence systems.