Modeling attention-driven plasticity in auditory cortical receptive fields.

Area of Science:

• Neuroscience
• Auditory Perception
• Computational Neuroscience

Background:

• Listeners selectively attend to relevant sounds in complex acoustic environments.
• Top-down selective attention reshapes neural representations, particularly in the auditory cortex.
• Spectro-temporal receptive fields (STRFs) adapt to enhance foreground sounds and suppress background noise.

Purpose of the Study:

• Propose a novel computational framework to model attention-driven plasticity of STRFs in the auditory cortex.
• Investigate how cortical plasticity optimizes neural representations for improved sound segregation.
• Explore both feature-based and object-based attention mechanisms.

Main Methods:

• Developed a discriminative modeling framework for attention-driven plasticity.
• The model optimizes STRFs to maximize foreground-distractor discriminability while maintaining representational stability.
• Instantiated the model for feature-based attention and extended it for object-based attention.

Main Results:

• The feature-based attention model replicated neurophysiologically observed STRF adaptation patterns, acting as a contrast matched filter.
• The object-based attention model provides testable predictions for auditory cortex function.
• The model aligns with known anatomical circuits involved in active attention.

Conclusions:

• Attention-driven plasticity in sensory cortex can be interpreted as a discriminative adaptation strategy.
• The proposed framework offers insights into the neural mechanisms of selective auditory attention.
• This work bridges computational modeling with neurophysiological findings in auditory perception.