The sounds we hear in almost every environment are reflected by nearby objects, producing many delayed and distorted copies of the original sound, known as reverberation. Our brains are usually able to filter this out, allowing us to recognise the source of each sound regardless of the environment, to the extent that people with normal hearing rarely notice it at all. However, reverberation is known to cause severe difficulties for many people with hearing impairments, and for speech recognition algorithms such as those designed to assist those who cannot hear.
Despite the ubiquity of reverberation, the question of how the brain’s auditory system copes with reverberation has been largely unanswered. Recent research has observed that the effects of room reverberation seem to be partially removed from the neural responses to sounds in the auditory cortex. However, so far it has been largely unknown how the brain could ‘remove’ reverberation in this way.
A new paper from the King and Walker groups has demonstrated, for the first time, how the brain may accomplish this ‘reverberation cancellation’. Their data shows that auditory cortical neurons adjust their filtering properties to adapt to changing acoustics of reverberant environments and reduce the effects of reverberation.
The team comprising Dr Aleksandar Ivanov, Professor Andrew King, Dr Ben Willmore, Associate Professor Kerry Walker, and Dr Nicol Harper first developed a simple computer model to predict what the brain should do in different reverberation conditions if it was aiming to filter out sound reverberations. They found that the model neurons filtered out reverberations by extending the inhibitory component of their receptive filters for more reverberant spaces - that is the model neurons were inhibited by sound further into the recent past in more reverberant spaces. Then they observed the responses of auditory cortical neurons in ferrets in a number of simulated reverberant environments. Notably, they saw what the model predicted, that the inhibitory fields of these neurons grew longer the more reverberant the environment as a result of an adaptive process, allowing similar cortical responses to sound sources regardless of the nature of the reverberation.
Prof Kerry Walker said: “The neurons in auditory cortex shift the timing of their receptive fields in a way that minimizes the effects of reverberation on their representations of sounds. They can essentially represent the original sound source without all the echoes. This allows us to recognise sounds and understand speech across different settings, such as a reverberant cathedral and a small bedroom. Understanding these mechanisms is particularly important for helping people with hearing impairments, as they often struggle to hear in highly reverberant environments.”
Dr Nicol Harper said: “Reverberation is a severe problem for individuals with hearing impairments and they tend to cope with it more poorly than those with normal hearing. By honing in on the mechanisms by which the brain copes with reverberation, perhaps we can tailor hearing prosthetics and maybe other future treatments to better enable the brain’s innate reverberation removal capacities, or perhaps we can incorporate synthetic versions of these reverberation removal processes into the prosthetics.
“It is remarkable that our simple model made concrete predictions about the inhibitory receptive fields of auditory neurons that were then revealed by our electrophysiological recordings. Our modelling, analysis, experimental methods and findings may pave the way for a fuller understanding of how the brain copes with reverberation. The neuroscience of hearing in reverb has been surprisingly little explored, but we have now laid out a new avenue in this field that we think could really open out this area of study.”
The full paper “Cortical adaptation to sound reverberation” is available to read in eLife.