Decision-making and Weber's law: a neurophysiological model

Area of Science:

• Computational Neuroscience
• Decision Neuroscience
• Neuroscience

Background:

• Ventral premotor cortex (VPC) neurons exhibit decision-related activity during sensory comparison tasks.
• Previous work has explored neural mechanisms underlying decision-making but lacked a direct link to psychophysical laws.

Purpose of the Study:

• To develop an integrate-and-fire attractor model for VPC neural activity in a vibrotactile frequency comparison task.
• To investigate the neurophysiological basis of Weber's Law using computational modeling.

Main Methods:

• Constructed a biased competition attractor network model.
• Incorporated firing rates of VPC neurons representing vibrotactile frequencies as bias inputs.
• Included finite size noise effects and divisive feedback inhibition via interneurons.

Main Results:

• The model's attractor states reflect the sign of frequency difference (Deltaf), not absolute frequencies.
• Transition probabilities to attractor states depend on scaled frequency difference, influenced by noise and inhibition.
• The model reproduces the probabilistic nature of decision-making observed in psychophysical experiments.

Conclusions:

• The study links neural network dynamics, specifically statistical fluctuations and divisive inhibition, to Weber's Law.
• Attractor network models can explain the neurophysiological basis of psychophysical phenomena in decision-making.
• Finite size noise and network architecture are crucial for understanding probabilistic decision outcomes.