Age-related changes in noise correlations in a decision-making task

Age-related declines in sensory and cognitive performance may arise from changes in neural variability, including how variability is shared across neural populations. Here, we examined age-related changes in noise correlations, which quantify shared trial-to-trial variability between pairs of neurons, in mice performing a visual decision-making task. We analysed large-scale extracellular recordings from 149 mice (male and female) spanning 3 to 20 months of age, comprising 348,937 simultaneously recorded neuron pairs within 17 cortical and subcortical brain regions. Across the dataset, noise correlations showed expected relationships with firing rate, anatomical distance, and signal correlations, consistent with canonical observations of correlated variability. We then evaluated age effects in two steps. First, we replicated previous findings of increased noise correlations with ageing in the primary visual cortex. Second, we extended these analyses across recorded cortical and subcortical regions. Region-resolved analyses revealed substantial heterogeneity, showing that previously reported sensory-cortical increases represent one component of a broader, region-specific pattern of age-related changes in shared variability. Age effects on pre- and post-stimulus noise correlations differed in direction across regions, whereas stimulus-induced modulation of noise correlations was more consistently attenuated with age. Together, these results show that ageing does not simply shift shared variability globally, but reshapes the regional profile and stimulus-dependent regulation of shared variability during perceptual decision-making.