Switches in brain states in memory consolidation: a computational approach

Brain information processing is shaped by fluctuations in population rhythmic activities, each defining distinctive brain states. Switches in brain states, driven by neuromodulation, plays a key role in the sleep-wake cycle. Such transitions are known to contribute to memory; a property called sleep-dependent memory consolidation; whose mechanisms are still poorly understood. Current computational models have often focused on the role of connectivity shift to reproduce sleep-wake pattern. Such models are not appropriate to study the impact of network oscillations on synaptic plasticity, since the rhythmic switch itself relies on a disruption of the connectivity established through learning.
Previous works have highlighted the role of the slow activation of T-type calcium channel in the generation of brain state switches that are robust to cellular heterogeneity, independent from network connectivity, and thus compatible with synaptic plasticity.
Here, we construct a simple thalamocortical network using conductance-based models embedding this slow channel activation. This network allows cellular-induced switches in brain activity resulting in the generation of sleep-wake cycles without alteration of synaptic weights. Then, synaptic plasticity is modeled by combining; (i) spike timing- dependent plasticity (for learning during wakefulness), (ii) long-term potentiation and (iii) long-term depression (for memory selection and consolidation during sleep). Learning is quantified through an associative memory task, i.e. the ability of neurons coding for concomitant events to strongly connect together. The proposed model makes it possible to compare the evolution of synaptic wiring during di!erent tasks and explore how sleep-wake cycles can help both consolidate relevant connections and weaken irrelevant connections.