Periodic Reporting for period 1 - AUDICON (Neural mechanisms of spectral context effects on auditory processing)
Reporting period: 2015-03-01 to 2017-02-28
The auditory system has evolved to process natural sounds efficiently and robustly, even under adverse acoustic conditions. Humans can detect, localize, and identify speech embedded in background noise in reverberant environments. The general aim of this research was to uncover the neural mechanisms underlying such processing and study their functions in realistic environments. In particular, we were concerned with hearing in adverse conditions i.e. when a sound scene is corrupted with many other distracting sounds. To that end we studied how auditory sensory processing adapts to the acoustical context in which it happens. Understanding the neural origin of such effects is of fundamental importance in neuroscience and will guide the design of human auditory brainstem prostheses. Hearing impairment is a public health priority and each research that moves forward the understanding of the healthy auditory is critical.
The overall scientific objective was to determine the site of neural emergence and understand the neural mechanism of a perceptual effect where a sound pops out out of a mixture because of the context in which it happens. The goals of this project were reached. On the scientific side, the results convincingly support the hypothesis. Dissemination have taken place during invited talk and conferences and a paper his on its way. On the learning side, the fellow is now trained as a neurophysiologist and can design and undertake independently electrophysiology experiments in the mammalian auditory system.
Our approach was multi-disciplinary, combining computational modelling and neurophysiology. Using stimuli used in human psychology to induce the perceptual effect of interest we recorded the electrical activity from more than 500 neurons in the rodent brainstem and auditory nerve. We discovered that the response of those neurons parallel the human perception in the same acoustic situation. Some neurons were labelled to confirm their position in the neural circuit. We used computational modelling to prove that the known neural architecture of the brainstem could explain the emergence of those perceptual effects. In particular, the dynamical interaction between excitation and inhibitory input to the brainstem produce a pop out effect in the neural output stream of the cochlear nucleus.
The main results can therefore be summarized as follow: 1) changes in sound perception due to the acoustic context have neural correlates in the mammalian auditory brainstem, i.e. neural responses parallel psychophysical measurements. 2) Those effects are present in the output cell of the cochlear nucleus but not in its input, i.e. they emerge at that neural center. 3) this emergence can be explained by the dynamical interaction between excitatory and inhibitory neural streams. Those results have been presented at four conferences. The fellow was invited at the university of Salamanca and Cambridge to give presentations about his findings. The main findings will make into a peer-reviewed journal.
Most of our percepts emerge in the cortex, i.e. they involve high-level neural mechanisms. We show here that a relatively complex perceptual effect -the fact that a sound can pop out a mixture because of the sounds preceding it- has neural correlates not further than 2 synapses away from the sensor, the cochlea. Unlike the visual system, the auditory system contains complex sub-cortical neural architecture. In particular, the center under investigation, the dorsal cochlear nucleus, exhibits very cortical-like circuit with neural plasticity and complex wiring. Understanding the function of such structure is of fundamental in auditory neuroscience. It is also known that this center changes its configuration after hearing impairment. Understanding how the processing is modified in pathological conditions could guide the design of algorithms used in auditory prosthesis.