Unpredictable Mixed-Valence Outcomes Induce a Chronic and Reversible Generalized Anxiety-like Phenotype in Male Mice

Background: Clinical anxiety is a generalized state characterized by feelings of apprehensive expectation and is distinct from momentary responses such as fear or stress. In contrast, most laboratory tests of anxiety focus on acute responses to momentary stressors.

Methods: Apprehensive expectation was induced by subjecting mice (for 18 days) to manipulations in which a running response (experiment 1) or a conditioned stimulus (experiment 2) were unpredictably paired with reward (food) or punishment (footshock). Before this treatment, the mice were tested in an open field and light/dark box to assess momentary responses that are asserted to reflect state anxiety. After treatment, the mice were assessed for state anxiety in an elevated plus maze, social interaction test, startle response test, intrusive object burying test, and stress-induced corticosterone elevations. In experiment 3, we treated mice similarly to experiment 1, but after mixed-valence training, some mice received either no additional training, additional mixed-valence training, or were shifted to consistent (predictable) reinforcement with food.

Results: We consistently observed an increase in anxiety-like behaviors after the experience with mixed-valence unpredictable reinforcement. This generalized anxiety persisted for at least 4 weeks after the mixed-valence training and could be reversed if the mixed-valence training was followed by predictable reinforcement with food.

Conclusions: Results indicate that experience with unpredictable reward/punishment can induce a chronic state analogous to generalized anxiety that can be mitigated by exposure to stable, predictable conditions. This learned apprehension protocol provides a conceptually valid model for the study of the etiology and treatment of anxiety in laboratory animals.