Signatures of Hebbian plasticity in the nanoscale morphodynamics of cortical spines

The size, shape, and nanoscale organization of synaptic spines are predictive of physiological synaptic strength and thus key to connectome-based neural circuit models. In the intact brain, spine size and shape undergo continuous morphological remodeling. While these nanoscale morphodynamics are not well understood, it is clear that morphological changes that strongly impact synaptic strength should modify spine head and neck in a coordinated manner and that such coordinated remodeling can be triggered by spike-timing-dependent synaptic plasticity. Here we demonstrate unambiguous signatures of such coordinated spine remodeling from in vivo long-term nanoscopy of cortical spines. We infer data-driven generative models of synaptic spine morphodynamics. These models imply that although coordinated remodeling represents the smallest component of ongoing morphodynamic fluctuations, this component causes massive fluctuations in instantaneous synaptic strength. Despite these fluctuations, synaptic strengths exhibit a synapse-specific long-term persistent component that results from spine fluctuations exploring only a limited subregion of accessible spine morphospace. Our results suggest that nanoscale morphodynamic spine remodeling injects substantial fluctuations of instantaneous synaptic strength into the operation of cortical circuits. Connectome-based neural circuit models thus need to consider features of spine organization beyond spine head size and should be augmented by data-driven generative models of spine morphodynamics.