Modeling of precise onset spike timing for echolocation in the big brown bat, Eptesicus fuscus.

Abstract

The neuronal circuits enabling highly accurate sonar performance in bats are not well understood. In this research, a computational neuroscience model was developed based on Meddis’ auditory peripheral model and a biologically plausible network of integrate-and-fire neurons. The stimuli presented to the model include acoustic simulations of target echoes using wideband FM echolocation calls recorded from Eptesicus fuscus. Synaptic connectivity of the neural network was varied to explore its effect on the precision of onset spike timing. Results show that precisely timed spike codes can relay sufficient information of echo time-frequency representations to account for performance observed in laboratory experiments. This research represents a portion of our overarching goal to reverse engineer the bat’s critical auditory functions into a refined sonar signal processing algorithm. A necessary first component of this process is to understand the functionality of relevant circuitry in the auditory midbrain. Such biomimetic processing schemes may be useful in active sonar applications suffering from a high degree of environmental clutter, where man-made systems typically demonstrate poor performance.