Relating a Spiking Neural Network Model and the Diffusion Model of Decision-Making.

Area of Science:

• Computational Neuroscience
• Cognitive Psychology
• Decision Making Models

Background:

• Decision-making models often assume evidence accumulation to a threshold.
• Spiking neural networks offer a biophysically plausible neural implementation.
• The diffusion model is a widely used cognitive-level account of decision making.

Purpose of the Study:

• To investigate the relationship between parameters of a neural-level spiking model and a cognitive-level diffusion model.
• To understand how neural parameters map onto diffusion model parameters like drift rate, threshold, and non-decision time.

Main Methods:

• Simulated experiments using a spiking neural network model.
• Factorial manipulation of choice difficulty and spiking model parameters.
• Fitting the diffusion model to simulated spiking network data.

Main Results:

• Spiking model parameters related to input sensitivity, threshold, and processing time mapped to drift rate, threshold, and non-decision time, respectively.
• Spiking model parameters without direct diffusion model analogues (e.g., background input, recurrent excitation) mapped to the diffusion model's threshold.
• Demonstrated a clear mapping between neural and cognitive model parameters.

Conclusions:

• The study provides a bridge between neural and cognitive models of decision making.
• Results inform the interpretation of diffusion model fits to behavioral data by linking them to underlying neural mechanisms.
• Highlights the utility of computational modeling in understanding decision processes.