Neural logic of multisensory integration

Neural logic of multisensory integration

Andreas Braun)
Drosophila larvae respond to a variety of sensory cues (heat, light, smell, taste and touch). These cues can lead to complementary or conflicting information. We are interested in identifying the principles and circuits controlling multisensory integration in the larva.

Is sensory perception a probabilistic inference process in the Drosophila larva? While it is usually assumed that the nervous system encodes the average of a fluctuating stimulus, increasing evidence suggest that information pertaining to stimulus variance is also transmitted to the brain. In humans and monkeys, sensory cues are combined in a nearly optimal way based on their relative level of reliability. Presumably, our brain has evolved to perform Bayesian inference. Theory predicts that this type of high-level computation involves complex, yet tractable, neural operations like the summation of activity across populations of neurons. The exact mechanisms underlying these operations have yet to be elucidated. We aim to clarify how congruent and conflicting sensory signals are integrated in the larval brain.