Normalization accounts for temporal dynamics in human somatosensory cortex.

Area of Science:

• Neuroscience
• Sensory Processing
• Computational Neuroscience

Background:

• Sensory processing is shaped by stimulation history, with adaptation in the visual cortex well-modeled by divisive normalization.
• Nonlinear temporal integration is documented in visual cortex but less explored in the human somatosensory system.

Purpose of the Study:

• To investigate if computational principles governing visual adaptation, specifically divisive normalization, apply to human somatosensory processing.
• To characterize the temporal dynamics of neural responses to vibrotactile stimuli in the human somatosensory cortex.

Main Methods:

• Utilized functional magnetic resonance imaging (fMRI) and intracranial electroencephalography (iEEG) in human participants.
• Measured brain responses to time-varying vibrotactile stimuli with varying durations and interstimulus intervals.
• Employed computational modeling, comparing linear integration with divisive normalization models to explain observed neural dynamics.

Main Results:

• Observed significant sub-additive temporal summation in somatosensory responses, indicating responses were lower than predicted by linear summation.
• Divisive normalization models demonstrated superior cross-validated accuracy in predicting neural responses compared to linear models.
• Somatosensory temporal dynamics were found to closely mirror those observed in the visual system.

Conclusions:

• Divisive normalization accurately models nonlinear temporal integration in the human somatosensory system.
• Suggests that divisive normalization is a canonical neural computation shared across sensory modalities for processing temporal information.
• Provides a unified framework for understanding how the brain encodes dynamic sensory stimuli across different senses.