Inferring neural dynamics of memory during naturalistic social communication

Area of Science:

• Neuroscience
• Computational Biology
• Animal Behavior

Background:

• Complex behaviors, such as social communication, rely on memory systems that continuously update with new information.
• The neural underpinnings of how the brain forms and maintains these evolving memories are not well understood.
• Investigating memory formation in naturalistic social behaviors, like fly courtship, offers a unique window into these mechanisms.

Approach:

• Developed 'Natural Continuation' (NC), a simulation-based model comparison framework for analyzing neural codes of complex stimuli using naturalistic behavioral data.
• Applied NC to study female fly memory for male courtship song history in unrestrained interactions.
• Utilized computational modeling to evaluate candidate neural mechanisms for representing extended auditory information.

Key Points:

• Evidence suggests an adaptive population coding mechanism in female flies' auditory system for processing long song histories.
• Heterogeneous nonlinear adaptation dynamics transform song temporal patterns, integrating them into persistent neural activity.
• This neural dynamics model effectively retains continuously unfolding information over extended periods, predicting female courtship behavior accurately.

Conclusions:

• The identified neural population mechanism provides a generic, scalable format for memory formation, capable of encoding extended input signals.
• This process generates multi-dimensional, advection-diffusion-like neural responses that differentiate songs across various timescales.
• Naturalistic behavioral data can directly inform neural population coding models, revealing novel memory formation processes.