Researchers at UC San Francisco reveal how the brain controls the complex movements of speech
A research team at UC San Francisco has identified the neural mechanisms underlying speech motor control — the precise, rapid coordination of tiny brain regions that govern the lips, jaw, tongue and larynx during speaking.
Reported in the journal Nature, this study has important implications for developing brain-computer interfaces for artificial speech, improving treatments for speech disorders, and deepening scientific understanding of a uniquely human ability: fluent spoken language.
“Speaking is central to human identity, and almost everyone learns to speak,” said senior author Edward Chang, MD, a neurosurgeon at the UCSF Epilepsy Center and faculty member in the UCSF Center for Integrative Neuroscience. “Yet it is one of the most complex motor behaviors we perform.”
The complexity arises because producing words depends on tightly coordinated actions of multiple vocal-tract articulators — the lips, tongue, jaw and larynx. Until now, how the brain coordinates these distinct effectors at the millisecond timescale remained poorly understood.
To build this map, Chang and colleagues recorded electrical signals directly from the cortical surface of three people undergoing brain surgery at UCSF. Using electrode arrays placed on the speech-related sensorimotor cortex, the team identified which cortical regions correspond to specific vocal-tract articulators and how those regions engage during speech.
The researchers then applied a sophisticated “state-space” analysis to capture the complex spatial and temporal patterns of neural activity that unfold when a person speaks. This approach reveals how neural population activity evolves over time, providing a dynamic view of how the brain controls rapid articulatory movements.
The results show that the speech sensorimotor cortex is organized in a hierarchical and cyclical fashion, enabling split-second, orchestra-like coordination of the tongue, jaw, larynx and lips. Rather than isolated, static hotspots, activity patterns form temporal sequences and recurrent loops that simplify the control of multiple articulators for fluent speech.
“These organizational principles may be cortical strategies that greatly reduce the complexity of coordinating the many articulators required for fluent speech,” said Kristofer Bouchard, PhD, a postdoctoral fellow in the Chang lab and the paper’s first author.
Just as a symphony requires musicians to time their notes precisely to create music, speaking depends on tightly timed interactions among different cortical areas to shape sounds into meaningful syllables and words.
Brain mapping during epilepsy surgery
The patients in this study were undergoing surgery at UCSF for severe, drug-resistant epilepsy. Surgical treatment can be life-changing for these patients, often stopping seizures when clinicians accurately identify and remove the small regions of brain responsible for seizure onset. Achieving that precision requires advanced intracranial monitoring.
At the UCSF Comprehensive Epilepsy Center, clinicians use surgically implanted electrode arrays placed on the brain’s surface to record neural activity directly. This intracranial monitoring can reveal seizure foci that noninvasive tests miss. In a follow-up procedure, the electrodes are removed and the problematic tissue is surgically resected.
These clinical procedures also create a rare opportunity to examine fundamental questions about human brain function. Speech motor control is difficult to study in animal models, and most noninvasive imaging tools cannot capture the extremely fast articulator movements that occur on the order of hundredths of a second. Intracranial recordings, by contrast, resolve neural changes on the order of milliseconds, making them uniquely suited to study rapid speech dynamics.
Historically, much of what was known about this region came from single-point electrical stimulation experiments dating to the 1940s, which produced muscle twitches in face or throat but never evoked meaningful speech sounds. The UCSF team used a different approach: while implanted arrays recorded ongoing cortical activity, patients read aloud a list of English syllables such as “bah,” “dee,” and “goo.” The researchers then related distinct spatial and temporal brain activity patterns to different consonants and vowels.
“Although our study used English syllables, the core neural patterns we observed echo principles that linguists have identified across languages, suggesting there may be universal neural strategies for speech production,” Chang noted.
Study details and acknowledgments
The paper, “Functional organization of human sensorimotor cortex for speech articulation,” is authored by Kristofer E. Bouchard, Nima Mesgarani, Keith Johnson and Edward F. Chang. It is reported in the February 20, 2012 issue of Nature.
Funding for the work was provided by the National Institutes of Health (grants R00-NS065120, DP2-OD00862 and R01-DC012379) and the Ester A. and Joseph Klingenstein Foundation.
Written by Jason Bardi
Contact: Jason Bardi – UCSF
Source: UCSF press release
Image source: Image adapted from the UCSF press release.
Original research: Abstract for “Functional organization of human sensorimotor cortex for speech articulation” by Kristofer E. Bouchard, Nima Mesgarani, Keith Johnson and Edward F. Chang in Nature. Published online February 20 2013 doi:10.1038/nature11911