Rational designing of oscillatory rhythmicity for memory rescue in plasticity-impaired learning networks

In the brain, oscillatory strength embedded in network rhythmicity is important for processing experiences, and this process is disrupted in certain psychiatric disorders. The use of rhythmic network stimuli can change these oscillations and has shown promise in terms of improving cognitive function, although the underlying mechanisms are poorly understood. Here, we combine a two-layer learning model, with experiments involving genetically modified mice, that provides precise control of experience-driven oscillations by manipulating long-term potentiation of excitatory synapses onto inhibitory interneurons (LTPE→I). We find that, in the absence of LTPE→I, impaired network dynamics and memory are rescued by activating inhibitory neurons to augment the power in theta and gamma frequencies, which prevents network overexcitation with less inhibitory rebound. In contrast, increasing either theta or gamma power alone was less effective. Thus, inducing network changes at dual frequencies is involved in memory encoding, indicating a potentially feasible strategy for optimizing network-stimulating therapies. [Display omitted] •E→I synaptic plasticity regulates oscillatory strength during learning and memory•LTPE→I mediated by γCaMKII drives dual oscillations at theta and gamma frequencies•Theta and gamma rhythmicity balances inhibition and inhibitory rebound for memory•Network stimulating at dual frequencies might be useful for cognitive therapies To achieve higher cognitive functions, rhythmic network oscillations are important and often get disrupted in psychiatric disorders. Li et al. find that learning-driven dual oscillations at theta and gamma frequencies require E→I synaptic plasticity mediated by γCaMKII, which plays an essential role in preventing network overexcitation during learning and memory.