Fig. 6: Analysis of data from each individual mouse in the IBL dataset. | Nature Communications

Fig. 6: Analysis of data from each individual mouse in the IBL dataset.

From: Identifying the factors governing internal state switches during nonstationary sensory decision-making

Fig. 6

a Observation weights for the Bernoulli GLM corresponding to distinct states within the four-state model across all IBL animals, highlighting the similarity and trends in state-specific observation parameters. b Transition weights for the Multinomial GLM of the same model across all animals, showcasing the dynamics and consistency of transitioning between different states for all animals. c Test log-likelihood (LL) variation for each mouse in the population, illustrated against the number of states. Each trace represents an individual mouse, with the solid red line indicating the average across all individual fits and the black solid line showing the test log-likelihood for the global fit. a–c Unit of study: mouse. n = 37 mice (biological replicates; independent animals from the IBL dataset). Each point/curve corresponds to one mouse. Panel c shows held-out test log-likelihood per mouse across model sizes (fivefold cross-validation; folds at the session level). Source data are provided as a Source Data file. Stim. stimulus, Obs. observation, Tran. transition, indiv. individual.

Back to article page