Phase-encoded fMRI tracks down brainstorms of natural language processing with sub-second precision.

Area of Science:

• Neuroscience
• Cognitive Science
• Computational Linguistics

Background:

• The precise neural mechanisms underlying natural language processing, including the interplay between cognitive and sensorimotor regions, remain incompletely understood.
• Current noninvasive neuroimaging methods, often relying on subtraction-based techniques, lack the necessary spatial and temporal resolution to capture dynamic information flow across the entire brain during language tasks.

Approach:

• Developed novel phase-encoded designs for functional magnetic resonance imaging (fMRI) to maximize temporal information extraction from brain activity.
• Overcame technical challenges such as scanner noise and head motion artifacts inherent in overt language tasks.
• Visualized neural information flow as coherent traveling waves across the cortical surface during listening, reciting, and cross-language interpreting.

Key Points:

• Captured dynamic neural information flows as traveling waves, providing unprecedented temporal and spatial detail.
• Visualized brain activity as 'brainstorms' on brain 'weather' maps, illustrating timing, location, direction, and intensity of neural activity.
• Demonstrated functional and effective connectivity of the brain during active language processing.

Conclusions:

• The developed technique offers a powerful new method for visualizing whole-brain neural information flow in real-time.
• These findings illuminate the functional neuroanatomy of language perception and production.
• Motivates the development of more sophisticated models of human information processing and brain function.