Mapping space from the self outward
Spatial memory is stored in an allocentric, world-centered frame, but behavior is executed egocentrically — actions are directed relative to the body. Our lab studies the neurons and circuits that transform between these reference frames, guided by computational models of egocentric–allocentric transformation. We record from freely moving rodents with single-unit electrophysiology and one-photon calcium imaging, pairing neural data with quantitative behavior.
Egocentric coding & reference-frame transformation
Behavior unfolds in an egocentric frame, yet the cognitive map is allocentric. We study where and how the brain converts between the two — from candidate transformation hubs in retrosplenial and parietal cortex and the subiculum to the action-selection output stage in striatum — building on our discovery of egocentric boundary cells in the dorsal medial striatum.
- Egocentric goal coding in DMS
- Top-down and bottom-up influence on goal coding
Boundaries, doorways & agency
Boundaries anchor spatial representations. We characterize the receptive field properties of egocentric boundary cells, including how their fields reshape as an animal moves through increasingly confined spaces. In parallel, we study how the neural code updates when a boundary becomes passable — when a wall becomes a doorway — and whether the animal's agency over that transition matters.
- Egocentric boundary cell receptive field properties
- Doorway coding & agency
Foraging & decision-making
When should an animal abandon a depleting resource? Using a two-patch foraging task grounded in the marginal value theorem, we find that rats systematically overstay the optimal leaving time. To identify what tunes foraging efficiency, we manipulate sex, social isolation, and gonadal hormones — finding, for example, that social isolation biases females toward safer, more efficient foraging. Recording from anterior cingulate cortex, we ask how the stay-or-leave decision is represented.
- Patch-leaving decision-making
- Patch-leaving coding in ACC
Social spatial coding
The egocentric framework extends naturally to social targets. During group foraging, we ask whether individuals adopt stable producer versus scrounger strategies — with scroungers finding food by monitoring their neighbors rather than searching for themselves. Separately, we test whether neurons encode the egocentric direction and distance to another animal, treating a conspecific as a dynamic spatial object.
- Social foraging: producer–scrounger
- Egocentric conspecific coding in RSC
Interested in this work?
We're recruiting graduate students, lab technicians, and postdoctoral fellows.