Research

Formalizing aspects of brain-body interactions

An intricate interplay exists between the brain and the body, acting over widely different timescales. From fast interoceptive spikes down to ultradian, circadian, and infradian hormonal cycles, various types of internal signals have been shown to affect behavior deeply. This suggests that the nervous system evolved together with the endocrine and immune systems to orchestrate physiological balance.

We are currently developing models at different scales describing physiological needs vs. brain interactions.

1. At the highest level, we have deep learning agents that interact with the world. These agents possess an idealized body and have internal needs, and are trained to minimize a loss representing physiological imbalance. This allows us to explore the effect of different physiological timescales on learning and behavior, and the strategies found to deal with conflicting needs.
2. At an intermediate level, we have models of sensory systems that funnel the amount of information they carry as a function of some underlying physiological need; think of olfaction, which is given more bandwidth when you're hungry. This allows us to compare optimal strategies found from theoretical principles with the information passing through real animals' sensory systems as a function of their underlying physiological state.
3. At a lower level, we have models of spiking neurons, which modulate synaptic connectivity, excitability, and neural plasticity as a function of some underlying neuromodulators. These substances can be endogenously released or can come from the outside, e.g., as a function of some physiological process. This approach allows us to investigate, from a normative perspective, various regulatory mechanisms that fine-tune neural circuits to meet specific physiological demands.