In a groundbreaking study conducted by researchers at MIT, a closer look into mouse behavior during reward-based tasks has uncovered surprising insights into their decision-making process. Published in the journal PLOS Computational Biology, the study utilized a novel analysis tool called blockHMM to decode the cognitive patterns of mice. The findings not only shed light on how mice learn and adapt in challenging situations but also have potential implications for understanding neurological conditions such as schizophrenia and autism.

The task assigned to the mice was seemingly simple: turn a wheel left or right to receive a reward and promptly recognize when the reward direction switches. Neurotypical individuals quickly grasp the optimal approach, known as the "win-stay, lose-shift" strategy. However, individuals with schizophrenia often struggle in performing this task. Surprisingly, the mice showcased the ability to learn this strategy but consistently deviated from it, perplexing the researchers.

"The surprising thing is that they don't persist with it. Even in a single block of the game where you know the reward is 100 percent on one side, every so often they will try the other side," noted corresponding author Mriganka Sur, Newton Professor in The Picower Institute for Learning and Memory and MIT's Department of Brain and Cognitive Sciences.

The study postulated that there could be various reasons behind the mice's departure from the optimal strategy. Lead author Nhat Le, a graduate student in the Sur Lab, hinted at the possibility that the mice might not commit to the "win-stay, lose-shift" approach due to a lack of trust in the stability or predictability of their circumstances. In natural settings, stability is not always guaranteed, prompting the mice to deviate from the optimal approach to test for rule changes.

To better understand the decision-making strategies employed by the mice, the research team utilized an analytical framework known as a Hidden Markov Model (HMM), adapted specifically for this study, which they named the blockHMM. This computational tool allowed them to decipher the true hidden states of the mice's behavior during the task. Through simulations, the researchers found that each mouse employed a combination of low, medium, and high-performance behavior modes, showcasing a mixture of strategies rather than a singular approach.

The implications of this research extend beyond understanding mouse behavior in reward-based tasks. By delving deeper into the brain regions and circuits involved, the researchers hope to gain insights that could potentially explain why individuals with schizophrenia exhibit diminished performance in reversal learning tasks. They also aim to investigate how the findings relate to the behavior of individuals with autism spectrum disorders, who tend to persist with unrewarded behaviors longer than neurotypical individuals.

"This reversal learning paradigm fascinates me since I want to use it in my lab with various preclinical models of neurological disorders. The next step for us is to determine the brain mechanisms underlying these differences in behavioral strategies and whether we can manipulate these strategies," shared Mriganka Sur.

The study's revelations on the complex decision-making process of mice provide valuable insights into the field of neuroscience. By recognizing that mice do not adopt a stationary strategy but instead shift between various approaches, researchers can ensure their interpretations of neural activity are accurate. This newfound understanding, facilitated by the blockHMM tool, has the potential to pave the way for innovative studies and treatments in neurological research.

The study, titled "Mixtures of strategies underlie rodent behavior during reversal learning," was authored by Nhat Minh Le, Murat Yildirim, Yizhi Wang, Hiroki Sugihara, Mehrdad Jazayeri, and Mriganka Sur. It was published on September 14, 2023, in PLOS Computational Biology.