Biology Seminar: Daniel Pollak

How can a null result be provocative? Magnetoreception, widely believed to be a real bodily sense in a range of species, carries the intoxicating promise of noninvasive neuromodulation, but years of searching for the sensor have turned up little, and basic questions remain open. One such question is how magnetic fields are represented in the brain. For vision and hearing we understand this reasonably well, with neurons tuned to specific features of the stimulus. I set out to look for the magnetic equivalent of visual or auditory neurons. Using brain activity I recorded from zebra finches, pigeons and also collaborative brain recordings in zebrafish, owls, and quails, we found no evidence of any that responded to magnetic fields. I'll talk about what to do with a provocative null result.

Hunting in the lab: The second project concerns a behavior that is common in the wild yet invisible to most of us: mice hunting prey. With experimental design engineered to study hunting, I identified a turning rule that guides it. I then investigated how activity in the superior colliculus, a midbrain region that converts sensory input into movement, transforms what a mouse sees into how it moves. I found that mice use one of two strategies and switch between them depending on the geometry of the hunt: they go where the target is when it's at a distance, and they go where the target is going to be when it's closer. Mice act on a sensorimotor delay of about 100 ms, and population activity in the colliculus is sufficient to decode several behavioral variables including turning rate, ground speed, and distance to target. The quality of this decoding is state-specific; while the mouse is hunting, this decoding is accurate, and while it is not hunting, the decoding largely fails.