A simple three layer excitatory-inhibitory neuronal network for temporal decision-making.

Area of Science:

• Neuroscience
• Computational Neuroscience
• Animal Behavior

Background:

• Temporal bisection is a key psychophysical task for understanding duration-based decision-making in humans and animals.
• Existing models often focus on choice proportions, neglecting crucial response time data.
• Accurate modeling of interval timing and categorization remains a challenge.

Purpose of the Study:

• To propose and validate a novel time-cell neural network model for temporal bisection.
• To simultaneously account for both choice proportions and response times in temporal categorization tasks.
• To investigate the neural mechanisms underlying duration-based decision-making.

Main Methods:

• Development of a time-cell neural network model simulating interval timing and categorization.
• Utilizing lurching wave activity for time interval tracking.
• Training the model to learn reference durations and categorization associations.
• Comparing model predictions against empirical data from rat temporal bisection experiments.

Main Results:

• The time-cell neural network successfully predicted canonical behavioral signatures of temporal bisection.
• The model accurately reproduced the sigmoidal relationship between test duration and long choice probability.
• It accounted for the superposition of choice functions on a relative time scale.
• The model correctly predicted the point of subjective equality and differential response time modulations.

Conclusions:

• The proposed time-cell neural network provides a robust framework for understanding temporal categorization and interval timing.
• This model offers a unified explanation for both choice behavior and response times in temporal bisection.
• It advances our understanding of the neural basis of time perception and decision-making.