Auditory brainstem models: adapting cochlear nuclei improve spatial encoding by the medial superior olive in reverberation

Abstract

12Listeners perceive sound-energy as originating from the direction of its source, even as the 13 direct sound is followed only milliseconds later by reflected sound from multiple different 14 directions. Auditory neural adaptation, before spatial information is binaurally encoded in the 15 brainstem, is suggested as a contributing mechanism, yet the form of sufficient adaptive 16 mechanisms remains unspecified. Though a strong preference for spatial information 17 conveyed in early-arriving sound is observed behaviourally and neurophysiologically, 18 investigation by simultaneously recording from binaural brainstem neurons and their monaural 19 input neurons while controlling adaptation is yet unrealised. Therefore, here we 20 computationally model connected neurons across the auditory brainstem, exploring two 21 biophysical mechanisms that can account for the brainstem's emphasis of early-arriving spatial 22 information. Matching responses in the medial superior olive (MSO), of neurons encoding 23 interaural time differences, we demonstrate that accurate localisation in reverberant 24 environments is enhanced by realistic pre-and post-synaptic specialisations. Short-term 25 synaptic plasticity at the inputs from the auditory nerve to spherical bushy cells (SBCs) in the 26 ventral cochlear nuclei that project bilaterally to the MSO, and KV1 membrane-channels in the 27 SBCs and MSO neurons, promote efficient coding of spatial information, by suppressing 28 responses to late-arriving sound, particularly at frequencies where reverberant sound is 29 relatively intense in outdoor environments. Suiting weak reverberation outdoors at very low 30 frequency, including 200 Hz, weak KV1 activity promotes accurate localisation. Applied to the 31 lateralisation of human speech near 700 Hz in a virtual reverberant room, identical synaptic 32 depression and strong KV1 activity at this higher frequency enhance correct, over spurious, 33 localisation cues at the earliest stages of spatial processing. 34 3 Author Summary 35 Aurally locating a talker in a reverberant room is often challenging for hearing-impaired 36 listeners. Sound arriving directly from the talker isn't easily distinguished from reverberant 37 sound reflecting off walls milliseconds later. Nevertheless, normal-hearing listeners easily 38 localize sound-sources, aurally "glimpsing" the reliable spatial information of direct sound 39 during early, rising portions of modulated sounds, while ignoring later-arriving, confounding 40 information corrupted by reverberation. Understanding the successful underlying mechanisms 41 may suggest approaches to improve auditory spatial information provided to listeners with 42 hearing aids and cochlear implants. In the auditory brainstem, binaural neurons of the medial 43 superior olive (MSO) preferentially respond to early-arriving spatial information. As 44 simultaneous recordings from MSO neurons and their monaural input neurons while controlling 45adaptation are yet unrealised, to understand the contributing mechanisms, here...

Cortical adaptation to sound reverberation

“…What are the neurophysiological mechanisms that support listening in reverberant environments? Previous studies have examined subcortical processes that facilitate localization of reverberant sounds (Yin, 1994;Litovsky and Yin, 1998;Fitzpatrick et al, 1999;Spitzer et al, 2004;Tollin et al, 2004;Pecka et al, 2007;Devore et al, 2009;Kuwada et al, 2012;Kim et al, 2015;Brughera et al, 2020), and how subcortical processing of synthetic periodic sounds is disrupted by reverberation (Sayles and Winter, 2008) and partially restored by compensatory mechanisms (Slama and Delgutte, 2015). Much less is known about the neural processing of speech and other complex natural sounds in the presence of reverberation.…”