COMPUTATIONAL MODELING OF AUDITORY NOVELTY RESPONSES IN ANAESTHETIZED AND CONSCIOUS MICE

Abstract

Auditory cortex of urethane-anaesthetised mice displays electrophysiological activity from 200 to 800 ms in response to unexpected, environmentally salient sounds.These auditory novelty responses differ from those seen in conscious mice that occur over a shorter timescale.Understanding the mechanisms responsible for these differences is relevant for validating mouse models of human auditory novelty responses.We explore this by fitting a recurrent neural network to epidural field potential recordings from the primary auditory cortex of anaesthetised and conscious mice exposed to unexpected sounds (50 or 150 ms, 10 kHz, 80 dB) in a deviant-alone paradigm.Model outputs highlighted an inverse relationship between certain waveform features and state of consciousness: P1-N2 (20 to 150 ms) was present in the conscious group but absent in the anaesthetised group, while PLL-NLL (200 to 800 ms) was present in anaesthetised but absent in conscious animals.The best-fitting model was examined by individually turning off units to calculate their importance for generating model outputs.A sub-network of units involved in replicating the response of urethane-anaesthetised mice was identified by thresholding their importance.Selectively silencing these units deteriorated model outputs representing the long-latency features of anaesthetised mice while also reducing P1 peak amplitudes from model outputs associated with the conscious mouse response.This provisionally suggests that neural populations involved in generating long-latency PLL-NLL auditory novelty responses in anaesthetised mice are also involved in the earlier P1-N2 deflections observed from conscious mice.From these findings we can hypothesize that urethane raises the threshold for triggering the neural mechanisms underpinning P1-N2 in conscious mice, and prolongs their activity resulting in a slower waveform feature that manifests as PLL-NLL in urethaneanaesthetised mice.Future experimental work should test this hypothesis and the proposed link between neurophysiological processes underlying P1-N2 in conscious mice and PLL-NLL auditory novelty responses in anaesthetised mice.