A fluctuation-driven mechanism for slow decision processes in reverberant networks

Area of Science:

• Neuroscience
• Computational Neuroscience
• Cognitive Science

Background:

• Cellular activity in cortical areas correlates with reaction times in decision tasks.
• Winner-take-all network models explain choice formation via excitation and inhibition.
• Traditional models induce decisions by increasing external input, destabilizing the network.

Purpose of the Study:

• To explore an alternative decision-making mechanism based on noise-induced escape.
• To investigate how finite-size noise can trigger transitions from a spontaneous state.
• To provide a dynamical basis for experimentally observed decision time variability.

Main Methods:

• Theoretical modeling of biologically plausible winner-take-all networks.
• Analysis of transitions from spontaneous to decision states.
• Numerical simulations to study noise effects on decision times.

Main Results:

• A novel mechanism of decision formation via noise-induced escape from the spontaneous state is proposed.
• This mechanism naturally occurs at low stimulus strengths.
• Decision times are exponentially distributed for small noise levels.
• Mean decision times show exponential dependence on noise amplitude.

Conclusions:

• The noise-induced escape mechanism offers a dynamical explanation for decision formation.
• This model accounts for the wide range and variability of decision times observed in behavioral experiments.
• The findings provide new insights into the neural basis of decision-making under uncertainty.