Neural activity flows through cortical subnetworks during speech production

Area of Science:

• Neuroscience
• Cognitive Science
• Speech Science

Background:

• The precise neural mechanisms underlying speech production, including the mapping of linguistic and articulatory information, remain incompletely understood.
• It is unclear whether the complex computations involved in speech production manifest as discrete spatiotemporal patterns of neural activity.

Purpose of the Study:

• To investigate the spatiotemporal dynamics of neural activity during speech production.
• To determine if distinct neural substrates and activity patterns underlie different stages of speech processing, such as perception, planning, and motor execution.

Main Methods:

• Utilized electrocorticography (ECoG) to directly record brain activity from neurosurgical participants.
• Employed an interactive speech paradigm to elicit naturalistic speech production.
• Applied clustering analyses and unsupervised dimensionality reduction to characterize neural activity patterns.

Main Results:

• Observed diverse cortical modulation profiles, clustering into distinct classes related to sensory perception, planning, motor execution, and task suppression.
• Demonstrated that these distinct activity classes are localized to separate neural substrates, suggesting specialized networks.
• Identified sequentially active subnetworks within planning and motor networks during speech preparation and articulation.

Conclusions:

• The findings support a localizationist model of speech production, where distinct neural pathways are specialized for different processing steps.
• Neural activity in speech production flows dynamically within and across these identified discrete pathways.
• This research provides a more detailed understanding of the neural organization underlying human speech.