Parameter-dependent shift from rational to irrational decisions in mice

Abstract

Cost-benefit analysis is a key determinant of decision-making , yet little is known about the underlying neural circuit mechanisms, and investigating this concept using laboratory animals is challenging without quantitative behavioral readouts and theoretical frameworks. In order to tackle this challenge, here we took a theory-based approach and created an experimental design in order for mice to perform a task as a rationale agent and optimize decisions. Our novel behavioral paradigm offers two possibilities of obtaining reward. Mice initially prefer the easier method, but as its cost increases, their preference switches. We quantified these switching decisions using an indifference point where the values of the choices became equivalent. A novel component of our task design is the implementation of a systematic and flexible variation of cost and benefit parameters in two-dimensional parameter space. This allows researchers to choose a wide range of parameter combinations and quantify the shift of the switching decisions. Our results demonstrated that indeed the parametric manipulation successfully influenced the switching decisions relative to the given parameter values, suggesting that mice perform the task as rational agents. A theoretical framework based on the optimization principle further confirmed the switching decisions of mice were optimal. Furthermore, we demonstrated that although the preference shift was influenced by different internal states of motivation, the valuation process of mice was intact regardless of motivational states. Considering the arsenal of tools available for mice , our theory-centered approach provides a quantitative

Enhancement of mediodorsal thalamus rescues aberrant belief dynamics in a mouse model with schizophrenia-associated mutation

“…In fact, numerous studies have shown impaired belief updating in schizophrenia [13][14][15][16][17] , but the underlying neural mechanisms remain unknown. In order to tackle this challenge, we modified a lever-pressing behavioral task 19 in which mice choose between either collecting high-value rewards with progressively increasing cost, or low-value rewards with fixed cost. To maximize the reward-to-cost ratio, mice must keep track of the cost changes and update their strategy accordingly.…”

“…Next, we trained WT and Grin2a Y700X+/mice on a lever-pressing task we adapted 19 to capture the belief updating process (Fig. 1).…”

“…Next, we trained WT and Grin2a Y700X+/− mice on a lever-pressing task we adapted 19 to capture the belief updating process (Fig.1). In this task, a mouse is provided with two levers to choose between on single trials, one that yields high reward (HR lever, 3 drops of milk) and another that yields low reward (LR lever, 1 drop of milk) (Fig.1a).…”