Integrating neuroplasticity into whole-brain modeling: From synapses to neural populations

The brain's ability to learn, adapt, and recover hinges on synaptic plasticity, yet most large-scale neural models overlook this core feature. In this review, we argue that embedding plasticity is essential for the simulation of realistic population-level brain activity. To this end, we explore specific computational attempts at scaling microscopic single-neuron plasticity rules to mesoscopic populations, drawing from mechanistic insights into short-term synaptic dynamics, long-term potentiation, and homeostatic regulation. We also discuss how plasticity mechanisms in neural field and neural mass models can potentially help us understand recovery trajectories, and how dysfunctional plasticity can play a role in them. Finally, we review state-of-the-art multiscale co-simulations methods that bridge spiking neuron networks with neural field and mass models and discuss their ability to link microscopic synaptic mechanisms with macroscopic brain dynamics. By embedding plasticity as a core feature, we can develop mean-field models that are more realistic and powerful tools for understanding brain function and recovery.