Mathematical modelling of the reaction latency. Part III: A model of avoidance reaction latency and avoidance learning

Area of Science:

• Behavioral neuroscience
• Computational neuroscience

Background:

• Reaction latency is crucial in understanding animal behavior and learning.
• Existing models may not fully capture the complexities of avoidance learning dynamics.

Purpose of the Study:

• To present a novel linear-dynamic-stochastic model for analyzing reaction latency in avoidance experiments.
• To classify different types of reactions based on the model and analyze their latency distributions.
• To investigate the learning processes and the relationship between avoidance and escape reaction thresholds.

Main Methods:

• Development and application of a linear-dynamic-stochastic model.
• Classification of reactions into escape, avoidance, late avoidance, inter-trial, and no reaction.
• Fitting the model to experimental latency data from avoidance conditioning.
• Estimation of the correlation coefficient between avoidance and escape reaction thresholds.

Main Results:

• The model successfully splits experimental latency distributions into components for escape, avoidance, and late avoidance responses.
• Two learning processes influencing latency were identified: one decreasing latency and another increasing it, with the former dominating.
• A negative correlation between avoidance and escape reaction thresholds was observed, tending towards -1 during learning, suggesting a functional dependency.

Conclusions:

• The model offers new insights into latency distributions and provides a tool for analyzing latency-based experiments.
• The identified learning processes explain latency changes and the decay of inter-trial responses.
• The functional dependency of avoidance and escape thresholds during learning may serve as an objective index of the learning state.