Biological survival poses some of the hardest computational challenges that we can think of. Escaping a predator within a fraction of a second, learning novel strategies without error margin, remembering near-misses for a lifetime while forgetting unimportant detail – these are difficult tasks for any intelligent agent, biological and artificial. To shed light on how some of these problems might be solved in neural or digital circuits, Dominik Bach's symposium brought together speakers from molecular, circuit, systems, and computational neuroscience.
A full success: “The symposium showed the entire breadth of the topic: from abstract/mathematical planning algorithms to molecular mechanisms and the development of novel technology,” says Dominik Bach. “It provided fantastic intellectual stimulation and sparked many engaging conversations between scientists working in this field.”
Bianca A. Silva from the Consiglio Nazionale di Ricerche (Italy) talked about the Neural mechanisms of memory update, Kevin Briggman from the Max Planck Institute for Neurobiology of Behavior (Bonn) spoke on anatomically constrained models of visual circuit computations, and Vanessa Stempel from the Max Planck Institute for Brain Research (Frankfurt) discussed midbrain circuits for flexible instinctive behaviors. Jason Kerr of the Max Planck Institute for Neurobiology of Behavior gave his talk on tracking skeletal kinetics and imaging cortical activity in the freely moving animal, and Benjamin Grewe from ETH Zurich (Switzerland) spoke about sensory and behavioral substrates of avoidance learning in prefrontal cortex population activity. Moreover, Peter Dayan from the Max Planck Institute for Biological Cybernetics (Tübingen, Germany) addressed the topic of risk, repetition, and rehearsal.
The composition of the speakers reflected the internationality of Bach's research group, which is based in London and Zurich in addition to Bonn. The keynote of the symposium was delivered by Dominik Bach himself with his inaugural lecture, in which he presented data from experiments designed to reveal biologically implemented decision algorithms. In the experiments, humans were exposed to various threats in a fully-immersive virtual reality, in which they could escape and run for shelter. “These data challenge the view that escape behavior is instinctive or hard-wired,” explained the psychologist, physician, and mathematician. “Instead, the underlying algorithm appears goal-directed, and dynamically updates decisions as the environment changes.” In contrast, information-seeking behavior might rely on simpler computations, he said. A special virtual reality lab, where more such simulations can take place, is currently under construction at the University of Bonn.
