Events Detail

Speaker:
Armin Bahl, Harvard University

Armin Bahl is a Postdoctoral Fellow at the Department of Molecular and Cellular Biology of Harvard University. He is interested in how nervous systems compute and control behavior by understading the neural basis of perceptual decision-making, using the larval zebrafish as a model. Through a combination of behavioral experiments, calcium imaging, anatomical analyses, and computational modeling, he hopes to unravel some general principles of the behavior that can also be tested in other animal models.

Neural basis of perceptual decision-making in larval zebrafish

To make appropriate decisions, animals need to accumulate sensory evidence. Even though the behavior can be well explained by simple integrator models with a decision threshold, the neuronal network implementation of such processes remains poorly understood. Here, we approach this problem by adapting an assay based on random dot motion kinematograms, usually used in primate studies, to larval zebrafish. Characterizing accuracy and delay of individual swimming decisions, we find that larvae temporally integrate and remember motion evidence over several seconds and that the behavior is best explained by leaky integration towards a threshold. Using brain-wide two-photon functional imaging at cellular resolution, we identify several neuronal clusters in the anterior hindbrain presumably involved in the underlying computations. Relating activity in these structures to individual behavioral choices, allows us to propose a biophysically plausible circuit model whose core elements are composed of two separate clusters that represent accumulated sensory evidence and decision threshold respectively, and that compete in a push-pull configuration for activating a downstream motor command. This arrangement provides an intriguingly simple implementation of the evidence integration mechanism and thresholding operation without the need for special cell-intrinsic biophysics. We are currently testing some of the structural predictions of our model through the study of neurotransmitter systems and single-neuron anatomy and we plan to further dissect the circuit by targeted electrophysiology, two-photon ablations and optogenetics, as well as functionally guided transcriptomics. Combining these techniques with more behavioral experiments will advance our understanding of the general neuronal mechanisms underlying evidence integration and decision-making.