Multimodal mismatch responses in mouse auditory cortex

Area of Science:

• Neuroscience
• Computational Neuroscience
• Sensory Processing

Background:

• Predictive coding models suggest that sensory prediction errors are computed in primary sensory areas based on deviations between expected and actual sensory input.
• Sensorimotor prediction errors arise from the mismatch between self-generated movements and their sensory consequences.
• The influence of prediction errors in one sensory modality on the computation of prediction errors in another remains largely unknown.

Purpose of the Study:

• To investigate multimodal prediction errors within the mouse auditory cortex.
• To understand how auditory prediction errors are modulated by visual feedback and vice versa.
• To explore the nature of cross-modal interactions in shaping prediction error signals.

Main Methods:

• Utilized a virtual reality environment to couple running behavior with auditory and visual feedback in mice.
• Employed two-photon microscopy to characterize neuronal responses in layer 2/3 (L2/3) of the auditory cortex.
• Experimentally induced audiomotor (AM) and visuomotor (VM) mismatches to probe prediction error computations.

Main Results:

• Layer 2/3 neurons in the auditory cortex responded to auditory, visual, and running stimuli.
• Audiomotor (AM) mismatch responses were found to be similar to visuomotor (VM) mismatch responses observed in the visual cortex.
• AM mismatch responses were significantly amplified when concurrently presented with VM mismatches, indicating cross-modal influence.

Conclusions:

• Prediction error computations in the auditory cortex are subject to multimodal influences.
• These interactions appear to be non-hierarchical, suggesting parallel processing between sensory modalities.
• The findings highlight the integrated nature of sensory processing in cortical circuits for prediction error signaling.