A team of neuroscientists at the Massachusetts Institute of Technology has identified, for the first time, a population of neurons in the auditory cortex of the human brain that responds only to the singing voice, and not to the spoken voice or instrumental music.
For the first time, neuroscientists at the Massachusetts Institute of Technology in the United States ( MIT ) have identified a population of neurons in the human brain that lights up when we hear singing, but not other types of music.
These neurons, found in the auditory cortex , appear to respond to the specific combination of voice and music (or melodious singing voice), but not to normal speech or instrumental music . Researchers say that what exactly they do is unknown and that more work will have to be done to find out.
“This study reveals some of the dominant response dimensions that organize the human auditory cortex, which is very important for understanding and modeling the neural mechanisms of auditory processing,” the first signatory of the study, Samuel Norman- Haignere , former postdoc at MIT and now assistant professor of neuroscience at the University of Rochester Medical Center (USA).
The work also suggests that music representations are fractionated into distinct neuronal populations that respond selectively to different types of music. Samuel Norman-Haignere, lead author
“The work also suggests that music representations are fractionated into distinct neuronal populations that respond selectively to different types of music. Understanding how and why this organization develops is an important question for future research,” adds the expert.
The study, published today in the journal Current Biology , builds on 2015 work in which the same research team used functional magnetic resonance imaging (fMRI) to identify a population of neurons in the brain’s auditory cortex that responds specifically to music. In the new research, the scientists used measurements of electrical activity taken on the surface of the brain, which gave them much more precise information than fMRI.
“There is a population of neurons that responds to singing, and very close there is another that responds broadly to a large amount of music. On the fMRI scale, they are so close that they cannot be separated, but with intracranial measurements we get additional resolution, and that is what we think allowed us to distinguish them,” says Norman-Haignere.
Combination of techniques
In the 2015 study, neuroscientists used fMRI to scan participants’ brains while they listened to a collection of 165 sounds, including different types of speech and music, as well as everyday sounds like tapping a finger or tapping. barking of a dog Thanks to this research, the scientists identified six neural populations with different response patterns, including a population selective for music and another population that responds selectively to speech.
Now, in their new work, the researchers used a technique known as electrocorticography (ECoG), in order to obtain higher resolution data. This technique allows electrical activity to be recorded using electrodes placed directly on the surface of the brain, providing a much more accurate picture of electrical activity in the brain compared to fMRI, which measures blood flow in the brain as an indicator of neural activity.
One of the methodological innovations of the study was to develop a technique to combine the precision of electrocorticography with the dense spatial coverage of fMRI.
“With most human cognitive neuroscience methods, you can’t see neural representations,” explains Nancy Kanwisher , professor of cognitive neuroscience and member of the McGovern Institute for Brain Research and the Center for Brains, Minds, and Machines ( CBMM ). from MIT. “Most of the data we can collect tells us that here is a piece of brain that does something, but that is quite limited. “We want to know what is represented there.”
Electrocorticography cannot normally be performed on humans because it is an invasive procedure , but it is often used to monitor epilepsy patients who are about to undergo surgery to treat their seizures. People are monitored for several days so doctors can determine where the attacks originate before operating. During that time, if patients agree, they can participate in studies that measure their brain activity while they perform certain tasks. For this work, the MIT team was able to gather data from 15 participants over several years.
“ECoG provides a much more precise measure of the neural response, but the opportunities to record responses are scarce and the coverage is much worse than fMRI, since the electrodes are implanted for clinical reasons (mainly to localize the foci of epileptic seizures). ),” Norman-Haignere explains to SINC. “One of the methodological innovations of the study was to develop a technique to combine the precision of ECoG with the dense spatial coverage of fMRI responses,” he adds.
Discovery of a new neuronal population
For these participants, the researchers played the same set of 165 sounds that they used in the previous study. The placement of each patient’s electrodes was determined by their surgeons, so some didn’t pick up any response to auditory input, but many did. Through a novel statistical analysis they developed, the experts were able to infer the types of neuronal populations that produced the data recorded by each electrode.
When analyzing the data, the scientists observed a neural response pattern that only responded to singing. Therefore, this population was distinct from the music- and speech-selective populations identified in 2015.
“When we applied this method to the data set, this neural response pattern emerged that only responded to singing .” It was a finding that we really did not expect, so it largely justifies the objective of the method, which is to reveal potentially novel things,” says Norman-Haignere.
That population of neurons had very weak responses to speech or instrumental music, and is therefore distinct from the music- and speech-selective populations identified in the 2015 research.
“In our previous study, we discovered six different response patterns, two of which responded selectively to speech and music, and four of which had responses that correlated with standard acoustic measures (e.g., sound frequency). Here, in addition, we discovered a neural response that responded selectively to singing,” highlights the expert.
Music in the brain
In the second part of their study, the researchers devised a mathematical method to combine the data from the intracranial scans with the fMRI data from their 2015 study. Since the MRI can cover a much larger portion of the brain, this allowed them to determine with greater precision the locations of the neuronal populations that respond to song.
“This way of combining ECoG and fMRI is an important methodological advance,” says Josh McDermott , associate professor in the department of Brain and Cognitive Sciences at MIT. “A lot of people have been doing ECoG in the last 10 or 15 years, but they have always been limited by this problem of the scarcity of recordings. “Samuel Norman-Haignere is really the first person who figured out how to combine the improved resolution of electrode recordings with fMRI data to get better localization of global responses.”
Their location suggests that these neurons can respond to features such as word interaction or perceived tone, before sending information to other parts of the brain for further processing.
This population of song-selective neurons is located in the upper part of the temporal lobe , near regions that are selective for language and music . Their location suggests that these neurons can respond to features such as perceived pitch, or the interaction between words and perceived pitch, before sending the information to other parts of the brain for further processing, the scientists explain.
Researchers hope to learn more about what aspects of singing drive the responses of these neurons. Additionally, together with the laboratory of MIT professor Rebecca Saxe, they intend to study whether babies have selective areas for music, in the hope of learning more about when and how these brain regions develop.
Reference:
Samuel Norman-Haignere et al. (2022) “A neural population selective for song in human auditory cortex”. Current Biology . DOI: 10.1016/j.cub.2022.01.069
