Discerning Mouse Trajectory Features With the Drift Diffusion Model.

Area of Science:

• Cognitive Psychology
• Behavioral Neuroscience
• Computational Modeling

Background:

• Decision-making theories increasingly incorporate dynamic, action-based measures.
• Mouse tracking offers a novel behavioral measure, suggesting decision strategies are stable across task designs.
• Mouse trajectory features may map to specific decision-making segments.

Purpose of the Study:

• To test the stability of decision strategies across different task designs.
• To validate mouse-tracking features against established decision-making models.
• To compare decision-making parameters derived from mouse-tracking versus keypress tasks.

Main Methods:

• Applied hierarchical drift diffusion and Bayesian delay discounting models.
• Compared delay discounting task (DDT) and stop-signal task (SST) in keypress and mouse-tracking formats.
• Analyzed mouse-motion features (e.g., maximum velocity, AUC) and their correlation with model parameters.

Main Results:

• High agreement in delay discounting rates between keypress and mouse-tracking DDT (ρ = 0.90).
• Convergent rates of evidence accumulation in both DDT (ρ = .86) and SST (ρ = .55).
• Mouse-motion features correlated significantly with nondecision time (ρ = -.42) and boundary separation (ρ = .44).

Conclusions:

• Mouse-tracking and keypress decision-making tasks show fundamental convergence.
• Mouse-tracking features like AUC and maximum velocity reflect decision conflict and impulsivity.
• Mouse tracking is a valid and reliable method for studying dynamic decision-making processes.