Summary: New research shows auditory and speech processing happen in parallel across the human auditory cortex, challenging the long-standing idea that sound is processed first and then converted into linguistic representations in a strictly serial fashion.
Source: Cell Press
Neuroscientists have identified a parallel processing pathway in the human brain that handles the sounds of language.
Published August 18 in the journal Cell, the study provides evidence that the brain processes acoustic and linguistic information simultaneously rather than in a simple stepwise sequence. These results challenge the prevailing model that auditory processing moves in a strict serial pipeline from low-level sound analysis to higher-level speech representation.
When spoken sounds reach the ear, the cochlea converts them into electrical signals that travel to the auditory cortex in the temporal lobe. For decades researchers have conceived of speech perception as a hierarchical process: the primary auditory cortex initially analyzes basic acoustic features such as frequency and amplitude, and nearby nonprimary regions—like the superior temporal gyrus (STG)—extract speech-relevant features (consonants, vowels) to form meaningful linguistic units.
Directly testing that serial model requires dense, high-resolution neurophysiological recordings across the entire auditory cortex, including deep areas of primary cortex that are difficult to access. The primary auditory cortex lies buried in a fold between the frontal and temporal lobes, making comprehensive surface recordings a technical challenge.
“We began hoping to demonstrate the classic transformation from low-level acoustic representation to higher-level linguistic representation,” says neuroscientist and neurosurgeon Edward Chang of the University of California, San Francisco.
Over seven years, Chang and colleagues recorded neural activity from nine patients who underwent neurosurgical procedures for clinical reasons—tumor removal or seizure localization. As part of their care, electrode arrays were placed to cover the full extent of the auditory cortex to guide surgical mapping. With patient consent, those clinical recordings were also analyzed to study how speech sounds are encoded and transformed in the human brain.
This dataset is unique: for the first time researchers could record simultaneously across primary and nonprimary auditory cortex directly from the brain surface, providing the spatial and temporal resolution needed to examine signal flow across these areas. Previous studies typically sampled only a few localized spots using penetrating electrodes, which limited the ability to observe distributed activity patterns.
Participants listened to short phrases and sentences while the team recorded cortical responses. If the classic serial model were correct, activity should appear first in primary auditory cortex and then propagate to adjacent regions like the STG. Instead, the researchers found that certain zones of the STG activated as rapidly as primary auditory cortex, indicating parallel onset of acoustic processing across distinct auditory regions.
To complement recordings, clinicians performed electrocortical stimulation during language mapping. Stimulating primary auditory cortex produced auditory hallucinations or non-speech noise but did not prevent participants from clearly hearing and repeating spoken words. In contrast, stimulation of specific nonprimary areas in the STG disrupted speech perception: patients reported hearing speech but being unable to make out words or reported that syllables sounded swapped.
These results argue against a simple hierarchical chain in which primary auditory cortex hands off progressively processed signals to neighboring regions. Instead, the evidence supports a distributed, parallel organization: primary and nonprimary auditory cortices process different aspects of speech simultaneously and can play independent roles in perception.
The discovery that the superior temporal gyrus has a distinct and essential role in decoding speech sounds has clinical implications. It suggests alternative strategies for diagnosing and treating disorders of speech perception, such as dyslexia, where children struggle to identify and manipulate speech sounds. Understanding parallel processing pathways may point to new therapeutic targets or rehabilitation approaches.
“This is an important step forward, but our understanding of the parallel auditory system remains incomplete,” Chang notes. “The findings indicate that the routing of sound information in the brain may be very different than previously imagined and raise many new questions for future research.”
Funding: This work was supported by grants from the NIH, The New York Stem Cell Foundation, The McKnight Foundation, The Shurl and Kay Curci Foundation, and The William K. Bowes Foundation.
About this auditory neuroscience research news
Author: Carly Britton
Source: Cell Press
Contact: Carly Britton – Cell Press
Image: The image is in the public domain
Original Research: Closed access. “Parallel and distributed encoding of speech across human auditory cortex” by Edward Chang et al. DOI: 10.1016/j.cell.2021.07.019
Abstract
Parallel and distributed encoding of speech across human auditory cortex
Highlights
- • Intracranial recordings were obtained from human primary and nonprimary auditory cortex
- • A superior temporal gyrus onset zone activates in parallel with primary auditory areas
- • Stimulation of superior temporal gyrus impairs speech perception
- • Stimulation of primary auditory cortex evokes auditory sensations without disrupting speech perception
Summary
Speech perception has been widely modeled as a feedforward series of transformations from acoustic to linguistic representations. Using extensive intracranial recordings across the entire human auditory cortex, along with electrocortical stimulation and surgical ablation, the authors show that cortical processing does not conform to a simple serial hierarchy. Response latency and receptive field analyses reveal parallel and distinct information processing in primary and nonprimary auditory regions. Functional dissociation is evident: stimulation of primary auditory cortex produces auditory hallucinations without distorting speech perception, whereas stimulation of nonprimary cortex in the STG disrupts comprehension. Ablation of primary auditory cortex does not eliminate speech perception. These findings establish a distributed organization of parallel processing across human auditory cortex and underscore an independent, essential role for nonprimary auditory cortex in speech processing.