Ecological Suboptimality in Naturalistic Foraging: Amplified Deviation from Optimality in a Mouse Model of Alzheimer Disease

Authors:
Preprint
Publisher website
Google Scholar
Abstract
Foraging requires animals to integrate spatial, temporal, and reward-related costs to guide adaptive action selection, yet how neurodegenerative pathology alters this process remains poorly understood. We compared foraging behavior in control (C57BL/6J) and APP NL-G-F transgenic mice, a knock-in mouse model of Alzheimer disease. The foraging task was structured such that distance to home, the size of food-pellet rewards, and food texture jointly shaped choices. These task variables implicitly imposed time costs related to travel, handling, and inter trial intervals. While control mice adjusted foraging behavior in relation to task demands across conditions, APP NL-G-F mice showed systematic deviations from control patterns. Their choices were characterized by increased withdrawal and reduced flexibility as spatial and temporal costs increased. Using a normative rate-maximization framework grounded in classical foraging theory, we quantified optimality and found marked suboptimality in both groups: control mice achieved 53% of the optimal reward rate, whereas APP NL-G-F mice achieved 45%. This suboptimality indicates that naturalistic decision making operates under context-dependent constraints rather than adhering to strict theoretical optimality. Within this constrained landscape, APP NL-G-F mice deviated more strongly from optimality than control mice, particularly under conditions requiring fine-grained integration of time-based costs. Group differences were most pronounced when optimal policies were highly sensitive to small errors in the estimation of temporal and spatial trade-offs. These findings suggest that Alzheimer-related pathology selectively amplified normative decision deficits rather than producing qualitatively distinct foraging strategies.
Funding Information
  • Natural Sciences and Engineering Research Council of Canada (RGPIN-2021-04149)
  • Alzheimer's Association (23-1152151)
  • Compute Canada Resource Allocation
  • Alberta Innovates

This publication has 50 references indexed in Scilit: