Detection of magnetic field properties using distributed sensing: a computational neuroscience approach.

Diverse taxa use Earth's magnetic field to aid both short- and long-distance navigation. Study of these behaviors has led to a variety of postulated sensory and processing mechanisms that remain unconfirmed. Although several models have been proposed to explain and understand these mechanisms' underpinnings, they have not necessarily connected a putative sensory signal to the nervous system. Using mathematical software simulation, hardware testing and the computational neuroscience tool of dynamic neural fields, the present work implements a previously developed conceptual model for processing magnetite-based magnetosensory data. Results show that the conceptual model, originally constructed to stimulate thought and generate insights into future physiological experiments, may provide a valid approach to encoding magnetic field information. Specifically, magnetoreceptors that are each individually capable of sensing directional information can, as a population, encode magnetic intensity and direction. The findings hold promise both as a biological magnetoreception concept and for generating engineering innovations in sensing and processing.

Duke Scholars

Author Sonke Johnsen Biology