“Putting some muscle into hearing: motor to auditory pathways for adaptive hearing and vocal learning” – Reflection
MAY 1, 2017 | LAURA GRIMA
Interactions between sensory and motor processes – it goes both ways…
It’s spring in Oxford, and you’re at a classical music concert in the Sheldonian Theatre. You settle down and get comfortable in your chair, ready to listen to a string quartet launch into Hadyn’s Opus 32. As they begin, activity in your auditory cortex changes in response to the incoming sensory stimulus; neurons increase or, in some cases, decrease their firing as the brain processes the music.
The violinist on stage is also listening to the same music as they play. However, whilst they may well be producing motor movements to generate the music, surely the pure sensory processing signal in the auditory cortex should be nearly identical to that of a passive listener in the audience? After all, the stimulus – the music – is the same regardless of who is listening to it, right?
In fact, this is not the case. It is intuitive to think of our brains as reactive, responding to sensory information in the environment as it comes in; this information is processed in the relevant sensory areas of the brain, and is then passed on to motor areas, leading to the performance of whatever behaviour is appropriate (such as running away from the sound of a rapidly approaching vehicle).
And yet a lot of neuroscience research suggests that, actually, often the brain is proactive in its processing; sometimes the flow of information processing goes the other way, with motor systems exerting an influence over sensory systems1.
Prof. Richard Mooney, from Duke University, has been working on understanding how such an interaction occurs between the auditory and motor cortices. His research shows that in the aforementioned example of the string quartet, the violinist’s motor movements can influence how the auditory cortex processes the music. In short, listening to someone else play music is different to listening to yourself playing music.
Movement suppression of auditory cortical activity: a predictive signal
One key question that Prof. Mooney’s research has addressed is how exactly sensory processing in the auditory cortex changes during various behaviours. Using mice, his lab have shown that, during a wide range of movements (including running, turning of the head, and grooming) auditory cortical activity is suppressed2. A similar response has been found in non-human primates during vocalisation3. But, interestingly, this suppression often occurs before the movement itself – in anticipation of the upcoming movement.
In fact, not only does auditory cortical activity dampen prior to the onset of movement, but as Prof. Mooney described in his recent talk for Cortex Club, they have also shown that this goes on to influence what is perceived by the animal. When mice have to lick at different rates in response to different sound intensities, they can do just fine – as long as they’re not moving: in other words, there is a movement-related cost to auditory perception.
This suggests that this suppression of activity in the auditory cortex is not caused by sensory information as it comes in, but is rather due to a copy of the upcoming motor command being sent to the auditory cortex – aiding prediction of the impending movement and its consequences. This is known as a corollary discharge, and examples of this predictive phenomenon have been found in other forms of sensory processing, particularly vision4.
Prof. Mooney’s lab has recently found further evidence that this movement-related suppression is strongly based on previously formed predictions, rather than just acting as a simple online signal that informs the animal that it is moving. They’ve used an ingenious experiment where mice run on a spinning disk – the clever twist being that the speed of the animal’s movement actually drives how quickly a tone beeps. After a bit of training, the mice learn what frequency of beeps occurs when running at different speeds; in other words, they form expectations of what movements produce which sounds. Interestingly, if the sound is expected, then that same aforementioned suppression of activity in the auditory cortex occurs – but if it is unexpected, it is less suppressed, again suggesting that this is a predictive mechanism.
Elucidating the circuit mechanisms: local and distal influences
To fully understand the relationship between movement and auditory cortical activity – and auditory perception – we also need an understanding of the mechanisms underlying these changes and how different areas can communicate with each other. And there is much evidence that these mechanisms occur on both local and more wide-reaching scales.
Prof. Mooney et al. have also been interested in understanding what the circuitry underlying this suppression of auditory cortical activity is. Using intracellular recordings they have shown that, at a local level, the activity of inhibitory circuits – specifically parvalbumin-positive interneurons, which can strongly modulate the activity of auditory cortical neurons postsynaptically – increases during the early stages of movement5. In fact, as with the auditory cortical suppression, the firing rate of these interneurons even increased before movement onset.
But what drives this increase in activity of inhibitory interneurons? There are well-documented links between the motor and auditory cortices (though it’s worth noting that the auditory cortex has multiple other innervations, including the basal forebrain6), but the Mooney lab have demonstrated that there is a subset of neurons in the motor cortex, specifically in M2, that synapse onto parvalbumin-positive interneurons7. And using optogenetics in resting mice to stimulate M2 terminals in the auditory cortex can actually mimic the changes seen in auditory cortical neurons – as though the animals are moving.
Attributing sensory input to self or other – who was that?
Whilst understanding the source and mechanisms of these signals is very interesting, it is also of course worth understanding why such signals arise in the first place- why does one subarea of the motor cortex have this influence on the activity of neurons in the auditory cortex? What function might this signal have?
One key advantage of such a mechanism is that it allows the brain to be able to tell the difference between sounds that are self-generated – like the violinist’s music – and sounds that come from other sources in the environment. Being more sensitive to externally-evoked auditory stimuli has a clear evolutionary advantage, as these are often sounds that you might want to pay attention to, such as the rustle of grass as a predator approaches…In other words, evasion over general auditory sensitivity8.
Models, predictions, and errors: learning from your mistakes
But what about the predictive nature of this signal? What purpose does this serve? In the brain more generally, predictive signals usually aid learning: the predictive signal itself acts as an internal model which is compared to a subsequent sensory input. The difference between the two – the error – is used to change the model and therefore change behavioural output.
Is this perhaps what is going on with the corollary discharge signal in the auditory cortex? This may well be the case: models of speech learning usually include such a signal as a mechanism of refining and improving sound-generating movements. Indeed, there is evidence of neurons in the auditory cortex homologue of songbirds that act as ‘error’ detectors, increasing their activity whenever auditory feedback is different from what is expected.9 The exact circuitry that encodes the internal model and compares it with auditory feedback to aid learning is still yet to be fully elucidated, but the Mooney lab have been able to show that ablating cells in a premotor area of the songbird brain, the HVC, that project to the auditory cortex mean that developing zebra finches have difficulty in copying another bird’s song. However, if these cells are removed in adulthood, the zebra finches’ ability to sing is unaffected, suggesting a strong role of these cells specifically in learning the song.
And what about in humans? Suppression of (some of) the auditory cortex has been found during speech9 and other self-generated sound10, although one imaging study has in fact found the converse during music playing: when participants were asked to play notes on a keyboard, the authors found an enhancement of activity in the auditory cortex in comparison to passive listening11.
Whilst the Mooney lab and others have made a lot of progress in terms of understanding the relationship between the auditory and motor cortices in terms of the corollary discharge signal and its circuitry, there of course still remain some unanswered questions. The anticipatory nature of this signal suggests a role in error-based learning, but specifically how this is encoded is yet unknown. Additionally, how specific is the corollary discharge signal in making these predictions – does it respond differently depending on the movement or context? Does this signal change over time?
Finally, whilst the basic finding of a corollary discharge signal in the auditory cortex seems to translate to humans, there is less evidence to suggest that this same signal contributes to speech learning specifically. However, the songbird, rodent, and non-human primate literatures suggest that it may have an important role to play in the refinement of sound-generating movements – and so whilst more translational research is needed, it may have helped the violinist playing at the Sheldonian Theatre learn how to play so beautifully.
1. Felleman, D. J., & Van Essen, D. C., (1991). Distributed hierarchical processing in the primate cerebral cortex. Cerebral Cortex, 1, 1-47.
2. Schneider, D. M., Nelson, A., & Mooney, R. (2014). A synaptic and circuit basis for corollary discharge in the auditory cortex. Nature, 513, 189-194.
3. Eliades, S. J., & Wang, X. (2003). Sensory-motor interaction in the primate auditory cortex during self-initiated vocalisations. Journal of Neurophysiology, 89, 2194-2207.
3. Sommer, M. A., & Wurtz, R. H. (2008). Visual perception and corollary discharge. Perception, 37(3), 408-418.
4. Schneider, D. M., Nelson, A., & Mooney, R. (2014). A synaptic and circuit basis for corollary discharge in the auditory cortex. Nature, 513, 189-194.
5. Nelson, A., & Mooney, R. (2016). The basal forebrain and motor cortex provide convergent yet distinct movement-related inputs to the auditory cortex. Neuron, 90(3), 635-648.
6. Nelson, A., Schneider, D. M., Takatoh, J., Sakurai, K., Wang, F., & Mooney, R. (2013). A circuit for motor cortical modulation of auditory cortical activity. Journal of Neuroscience, 33(36), 14342-14353.
7. Schneider, D. M., & Mooney, R. (2015). Motor-related signals in the auditory system for listening and learning. Current Opinion in Neurobiology, 33, 78-84.
8. Keller, G. B., & Hahnloser, R. H. R. (2009). Neural processing of auditory feedback during vocal practice in a songbird. Nature, 457, 187 – 190.
9. Flinker, A. , Chang, E. F., Kirsch, H. E., Barbaro, N. M., Crone, N. E., & Knight, R. T. (2010). Single-trail speech suppression of auditory cortex activity in humans. Journal of Neuroscience, 30(49), 16643-16650.
10. Martikainen, M. H., Kaneko, K., & Hari, R. (2005). Suppressed responses to self-triggered sounds in the human auditory cortex. Cerebral Cortex, 15, 299-302.
11. Reznik, D., Henkin, Y., Schadel, N., & Mukamel, R. (2013). Lateralized enhancement of auditory cortex activity and increased sensitivity to self-generated sounds. Nature Communications, 5, 4059.