Decoding Articulatory Features from fMRI Responses in Dorsal Speech Regions

Area of Science:

• Neuroscience
• Cognitive Science
• Speech Communication

Background:

• The overlap between brain regions for speech perception and production is crucial for development and behavior.
• The role of sensorimotor integration in speech perception is debated, with challenges in separating acoustic and articulatory neural responses.
• Understanding how the brain decodes speech involves investigating the interplay between auditory signals and motor commands.

Purpose of the Study:

• To investigate the generalization of articulatory speech features (place, manner, voicing) beyond their acoustic form during passive listening.
• To disentangle neural responses to acoustic versus articulatory features of speech using advanced fMRI decoding techniques.
• To explore the representation of sensorimotor information within the auditory pathway during speech perception.

Main Methods:

• Utilized a passive listening paradigm with spoken syllables.
• Employed multivariate decoding of single-trial functional Magnetic Resonance Imaging (fMRI) responses.
• Trained classifiers to discriminate articulatory features and tested generalization across different syllable types.

Main Results:

• Revealed generalization of place and manner of articulation across cortical levels in the dorsal auditory pathway, including auditory, sensorimotor, motor, and somatosensory regions.
• Demonstrated generalization of voicing information to the right anterior superior temporal sulcus and bilateral somatosensory regions.
• Indicated the availability of articulatory codes during passive speech perception, suggesting sensorimotor information representation.

Conclusions:

• Sensorimotor integration is fundamental to verbal communication, linking speech perception and production.
• The brain actively uses articulatory representations during passive speech perception, not just acoustic/phonetic ones.
• Findings confirm articulatory-specific brain responses at multiple cortical levels, supporting the role of sensorimotor information in speech processing.