Respiratory Navigation

The subtle fragrance of a blooming rose that momentarily halts us in our tracks or the arresting odor of burning toast that instantly telegraphs a breakfast mishap serves as a universal reminder of the rich complexity that underlies the sense of smell. This intricate sensory domain serves as a compelling focal point within the larger landscape of the brain, body, and environment interactions that have long been a cornerstone of research in evolutionary robotics and whole animal modeling. Within this context, olfaction occupies a unique position. It captivates scientific interest not merely due to its puzzling nature, but also owing to its intimate interconnection with the volatile and ever-evolving environmental landscape. This system not only underscores the need for organismal adaptability but also exemplifies how sensory modalities are perpetually shaped by—and in turn shape—their surroundings. The study of olfaction in the field therefore serves as a poignant model for understanding how complex systems adapt and evolve in synchrony with dynamic external stimuli.

One interesting component, idiosyncratic to mammalian chemoreception, is respiratory spatial navigation. Respiratory spatial navigation presents an interesting case, as the agent must navigate both a turbulent and fluctuating chemical environment through the modulation of its own internal oscillatory physiology. This is seen most clearly in that, when a sufficient concentration of odorant is in the air, the organism will "sniff", or modulate its respiratory rhythm in such a way as to draw in air in quick rapid successions. This kind of dynamic and active sampling has led some to argue that sniffing is a clear example of active perception. As the modulation of variables such as sniff speed, frequency, and depth can profoundly influence the temporal dynamics of odorant concentrations, which in turn, impacts odor perception. This capacity, akin to adjusting the focus on a camera lens, allows organisms to 'tune' their olfactory system dynamically depending on both environmental and physiological context.

The code that produces these simulations is publicly available at the following Github repository: see here.