Neurobiology of Navigation

Navigation is a key behavior for animals and humans. It is a cognitively-demanding, complex behavior – which requires planning, strategy-shifting, and ongoing integration of sensory information with memory. We investigate the neural basis of navigation, spatial memory, and spatial learning – and we ask: What are the neural codes that underlie real-world navigation? In particular, we study the neural codes for 3D space, 1-kilometer distances, and complex navigation.

Our discoveries include

• We found that in flight, 3D hippocampal place cells have nearly spherical 3D place fields. 
• Recordings in the bat presubiculum revealed 3D head-direction cells, which could serve as a 3D compass; and surprisingly, this compass followed a toroidal coordinate system – providing an interesting biological solution to the discontinuity and non-commutativity problems associated with a standard spherical coordinate system.
• Recordings in entorhinal cortex of flying bats reveled 3D grid cells that exhibited fixed local distances between firing-fields, but no global hexagonal lattice – arguing against many of the prevailing theories on the function of grid cells, which rely on lattice-like periodicity.
• We discovered a new population of neurons in the hippocampus that are tuned to the egocentric direction and distance to navigational goals – a vectorial representation of spatial goals, which could provide a neural mechanism for goal-directed navigation.
• In two studies (one and two) we showed nonoscillatory phase-coding in bat place-cells and grid-cells, without any theta oscillations – suggesting that phase-coding, but not oscillations, is conserved across species.
• We discovered a surprising optimization principle in the sonar system of Egyptian fruit bats.
• We performed the first GPS tracking of bat navigation outdoors, using tiny GPS dataloggers – this behavioral study provided evidence for a cognitive map in bats.
• We constructed a huge behavioral setup – a 200-meter long tunnel – and found that in bats flying in a very large environment, hippocampal CA1 neurons exhibit a surprising neural code for space – a multifield multiscale code – where the fields of the same neuron differed up to 20-fold in size; theoretical analysis showed that this multiscale code reduced dramatically the decoding-errors.
• In pairs of bats flying in the long tunnel, hippocampal neurons exhibited extreme dynamism of the neural code – rapidly switching between coding position when flying alone, to jointly coding position x distance when encountering another bat.