Abstract:A core challenge for the brain is to process information across various timescales. This could be achieved by a hierarchical organization of temporal processing through intrinsic mechanisms (e.g., recurrent coupling or adaptation), but recent evidence from spike recordings of the rodent visual system seems to conflict with this hypothesis. Here, we used an optimized information-theoretic and classical autocorrelation analysis to show that information- and correlation timescales of spiking activity increase along the anatomical hierarchy of the mouse visual system under visual stimulation, while information-theoretic predictability decreases. Moreover, intrinsic timescales for spontaneous activity displayed a similar hierarchy, whereas the hierarchy of predictability was stimulus-dependent. We could reproduce these observations in a basic recurrent network model with correlated sensory input. Our findings suggest that the rodent visual system employs intrinsic mechanisms to achieve longer integration for higher cortical areas, while simultaneously reducing predictability for an efficient neural code. How the brain integrates information across different timescales is a fundamental question in neuroscience. Results from primates suggest that higher areas in cortex are specialized on integrating information on longer timescales through stronger network recurrence. However, due to anatomical differences and conflicting empirical evidence, it remains open whether this is a property that is shared among species as a general feature of temporal processing. For example, in rodents, higher cortical areas show an increase in adaptation, which suggests stronger redundancy reduction that might oppose an enhanced temporal integration. Here, we combined an information theoretic analysis with an analysis of correlation timescales and found an increase in information and correlation timescales across the anatomical hierarchy of the mouse visual system. Notably, this upward trend is accompanied by a simultaneous reduction in the predictability of single-neuron spiking, suggesting a decrease in temporal redundancy. We could reproduce these findings in recurrent network models, which demonstrates that enhanced temporal integration through an increase in recurrence does not necessarily oppose a reduction of redundancy for individual neurons. Together with our empirical findings, this suggests that mouse visual cortex might exploit both, enhanced temporal integration and stronger adaptation, to tune hierarchical temporal processing.

Flexible neural population dynamics govern the speed and stability of sensory encoding in mouse visual cortex

Locomotion Enhances Neural Encoding of Visual Stimuli in Mouse V1

Temporal information of subsecond sensory stimuli in primary visual cortex is encoded via high dimensional population vectors

Spatiotemporal Dynamics of Neuronal Population Response in the Primary Visual Cortex.

Long-term stability of cortical population dynamics underlying consistent behavior

Deciphering neuronal variability across states reveals dynamic sensory encoding

Recurrent cortical networks encode natural sensory statistics via sequence filtering

Latent dynamics of primary sensory cortical population activity is structured by fluctuations in the local field potential

Dynamic gating of perceptual flexibility by non-classically responsive cortical neurons

Multiple distinct timescales of rapid sensory adaptation in the thalamocortical circuit

Coordinated Response Modulations Enable Flexible Use of Visual Information

Mesoscale cortical dynamics reflect the interaction of sensory evidence and temporal expectation during perceptual decision-making

Cortical reactivations predict future sensory responses

State-dependent pupil dilation rapidly shifts visual feature selectivity

The neural and computational architecture of feedback dynamics in mouse cortex during stimulus report

Task-dependent recurrent dynamics in visual cortex

State-dependent regulation of cortical processing speed via gain modulation

Experience-dependent predictions of feedforward and contextual information in mouse visual cortex

Signatures of hierarchical temporal processing in the mouse visual system

Distinct functional classes of excitatory neurons in mouse V1 are differentially modulated by learning and task engagement

Learned response dynamics reflect stimulus timing and encode temporal expectation violations in superficial layers of mouse V1