Jumping to predictions: Auditory and visual prediction in hurdling

Area of Science:

• Neuroscience
• Cognitive Science
• Motor Control

Background:

• The brain integrates multisensory information (auditory, visual) to create a coherent perception.
• Internal predictive models, shaped by experience, are crucial for maintaining perception during sensory deprivation.
• Complex movements like hurdling require sophisticated sensorimotor integration and prediction.

Purpose of the Study:

• To investigate how the brain compensates for temporary auditory and visual masking during a movement prediction task.
• To examine the neural mechanisms underlying prediction accuracy and reliance on internal models.
• To assess the impact of sensorimotor training (hurdling) on predictive capabilities and neural activation patterns.

Main Methods:

• Functional magnetic resonance imaging (fMRI) was used to monitor brain activity in participants watching masked hurdling videos.
• Participants performed a movement prediction task under conditions of visual masking, auditory masking, or combined masking.
• A six-week hurdling training intervention was implemented between fMRI sessions to enhance sensorimotor models.

Main Results:

• Prediction accuracy decreased significantly when auditory or visual input was masked.
• Under visual masking, participants showed increased reliance on auditory input and engaged frontal, motor, and visual areas, suggesting top-down visual prediction.
• Hurdling training improved prediction accuracy, and associated neural changes were observed in frontal control and visuomotor regions, indicating enhanced internal models for visual prediction.

Conclusions:

• The brain flexibly recruits modality-specific networks based on available sensory input, utilizing top-down predictions when visual information is absent.
• Sensorimotor training strengthens internal models, leading to more efficient predictive processing, particularly for visual information.
• These findings highlight the dynamic interplay between sensory input, internal models, and motor learning in perceptual prediction.