The Plasticitome: Why Every Synapse in the Brain Learns by Its Own Rules

For more than thirty years, the study of learning and memory has been anchored to a remarkably small set of model synapses. The Schaffer collateral connection onto hippocampal CA1 pyramidal neurons, the parallel fiber input onto cerebellar Purkinje cells, and unitary connections between neocortical pyramidal neurons have each yielded Nobel-caliber mechanistic insights, from the role of NMDA receptors in coincidence detection to the dendritic calcium signals that gate long-term potentiation. But a growing body of evidence now argues that these canonical preparations, however productive, cannot tell the whole story. Synaptic plasticity—the cellular substrate widely regarded as underpinning learning, memory, and adaptive circuit function—turns out to differ profoundly depending on the synapse type, the identity of the neurons involved, the architecture of the circuit, the neuromodulatory state of the brain, the developmental stage, and even the presence of disease. In place of a single universal learning rule, the brain appears to operate a rich repertoire of mechanisms, each specialized for the computational demands of its own synapses and circuits.

That realization has given rise to a new organizing concept: the plasticitome. Coined by analogy to the genome, proteome, and connectome, the plasticitome denotes the full landscape of synaptic learning rules across the brain, together with the biological contexts in which each rule is engaged—pre- and postsynaptic cell identity, induction protocol, neuromodulatory milieu, developmental stage, and pathological state. Just as genomics transformed biology by systematically cataloging genes rather than studying them one at a time, proponents of the plasticitome argue that understanding how circuits learn requires systematically charting the diversity of plasticity mechanisms across synapse types and brain regions. A new Editorial in Frontiers in Synaptic Neuroscience, published by Christina You Chien Chou of McGill University, Serena M. Dudek of the National Institute of Environmental Health Sciences, Clive R. Bramham of the University of Bergen, Hyungju Park of the Korea Brain Research Institute, and P. Jesper Sjöström of McGill, lays out this vision and surveys the contributions of a dedicated Research Topic that spans molecules, cell types, circuits, and disease states.

Charting the plasticitome is easier said than done. Traditional electrophysiological approaches to plasticity—pairing presynaptic stimulation with postsynaptic depolarization or spiking and measuring the resulting change in synaptic efficacy—are inherently low throughput, typically examining one connection at a time. This bottleneck has motivated the development of higher-throughput experimental and computational strategies, including robotic automation of targeted whole-cell recordings, all-optical interrogation of neural circuits, and large-scale efforts to quantify connectivity and short-term plasticity across neuronal cell types in mouse and human neocortex. The Editorial argues that an initial plasticitome atlas may emerge by combining standardized measurements across synapse types, brain regions, and physiological states into a common experimental framework, much as recent large-scale connectomic efforts have done for neuronal wiring.

The molecular studies collected in the Research Topic illustrate how differently specialized synapses can implement plasticity. In Drosophila, Acharya and colleagues compared the presynaptic active zones of olfactory sensory neurons detecting alarm odors via the Or56a receptor with those detecting food odors via Or22a. The alarm-odor neurons contained fewer active zones but expressed more Unc13A, a presynaptic release protein, consistent with increased transmitter release, whereas the food-odor neurons carried a molecular signature associated with lower release. Strikingly, only the Or22a neurons compensated for reduced presynaptic release by increasing both active zone number and Unc13A expression—a form of cell-type-specific homeostatic plasticity revealing that even sensory neurons of the same organ deploy distinct molecular strategies for synapse-type-specific adjustment.

Computational modeling is proving essential for probing molecular mechanisms that are difficult to access experimentally. Bartol and colleagues built a spatially realistic model of hippocampal dendritic spines to investigate what prolongs the lifetime of autophosphorylated CaMKII, the calcium- and calmodulin-dependent protein kinase II long implicated in maintaining long-term potentiation. Their simulations showed that calmodulin trapping extends the lifetime of the autophosphorylated enzyme by hindering dephosphorylation rather than by promoting further autophosphorylation—a subtle but consequential distinction for how synaptic memory traces are stabilized at the molecular level. In a complementary line of work, Noriega-Prieto and colleagues demonstrated that insulin-like growth factor I exerts bidirectional control over NMDA receptor-dependent plasticity in mouse barrel cortex: higher concentrations of IGF-I facilitated Hebbian LTP by inducing long-term depression of inhibitory transmission, while lower concentrations prevented LTP by depressing excitatory transmission. Because neuronal activity is known to drive IGF-I transport across the blood-brain barrier into the CNS, physiological fluctuations in local IGF-I levels may dynamically tune cortical plasticity in vivo.

Neuromodulation also sculpts plasticity within central circuits. In the mouse olfactory tubercle, a reward-related structure where odor-guided behaviors are evaluated, reward and punishment induce domain-specific plasticity in the anteromedial and lateral domains. Podder and colleagues found that although both regions expressed long-term potentiation, the neuropeptide orexin selectively lowered the LTP threshold in the anteromedial domain, where orexin receptor expression was higher. The same shared plasticity mechanism is thus tuned by regional differences in neuromodulation, potentially supporting distinct behavioral functions—a vivid example of how biological context, not just synapse identity, defines a learning rule.

Cell-type-specific learning rules emerge with particular clarity in the cortical interneuron studies. Chou and colleagues showed that postsynaptic spiking determines anti-Hebbian long-term depression at excitatory synapses onto visual-cortex basket cells, extending earlier observations at hippocampal pyramidal-to-basket-cell synapses. Synchronous pre-before-post pairing elicited supralinear dendritic calcium signals and LTD in these inhibitory neurons, whereas asynchronous pairing did not. Using high-throughput two-photon optomapping, the same team revealed that an identical pairing protocol potentiates pyramidal-to-pyramidal connections but depresses pyramidal-to-basket-cell inputs—opposing plasticity at neighboring synapse types within the same microcircuit. Guinet and colleagues extended the principle to inhibitory synapses in the subiculum, showing that associative stimulation induced inhibitory LTP in both burst-spiking and regular-spiking pyramidal neurons, but non-associative stimulation elicited iLTP only in regular-spiking cells. Even among closely related principal neurons, inhibitory plasticity rules depend on postsynaptic cell identity.

Short-term plasticity adds another dimension to the diversity. McFarlan and colleagues demonstrated that the short-term plasticity of vasoactive intestinal peptide interneuron outputs in mouse motor cortex is strongly synapse-type specific: VIP outputs onto Martinotti and basket cells consistently exhibited short-term depression, while excitatory inputs to VIP interneurons displayed highly heterogeneous dynamics linked to presynaptic release probability and postsynaptic cell identity. Because heterogeneous short-term plasticity provides a substrate for temporally structured population dynamics, this diversity emerges as a key ingredient of circuit computation. In a related synthesis, Severin and Kirkwood discussed how activity-dependent recruitment of parvalbumin-positive interneurons regulates inhibitory tone across development, aging, and disease, identifying the synaptic protein NPTX2 as a key regulator of PV interneuron excitation and proposing that restoring excitatory drive onto PV cells may offer a therapeutic strategy for limiting the network hyperexcitability seen in neurodegenerative disease. Meanwhile, Nasrallah and colleagues uncovered a previously unrecognized form of heterosynaptic NMDA receptor plasticity in dentate granule cells: burst-timing-dependent potentiation at medial perforant path synapses induced NMDAR LTP not only at the stimulated inputs but also at unstimulated lateral perforant path inputs, where AMPAR-mediated transmission instead underwent heterosynaptic LTD. These dissociable NMDAR and AMPAR rules at convergent entorhinal inputs onto the same neuron reveal distinct learning mechanisms operating side by side.

Physiological and pathological states reshape the plasticitome in equally striking ways. Gall and colleagues argue that many longstanding controversies surrounding hippocampal LTP arise because LTP is not a single uniform phenomenon but comprises multiple forms differing across hippocampal regions, sexes, and patterns of activity—the how and the why of LTP both depend on biological context. Feighan and colleagues reviewed the convergence and divergence of molecular mechanisms underlying Hebbian and homeostatic plasticity, comparing synaptic scaffolding proteins, intracellular signaling pathways, cell-adhesion molecules, and secreted factors in the regulation of AMPA receptors, and identifying both shared and distinct molecular pathways. García-Magro and colleagues showed that diabetes impairs plasticity in the primary somatosensory cortex by converting whisker stimulation-induced LTP into NMDA receptor- and metabotropic glutamate receptor-dependent LTD; local application of IGF-I partially restored LTP, directly linking impaired IGF-I signaling to altered cortical learning rules. And Halim and colleagues demonstrated that the structure of developing inner hair cell ribbon synapses in the auditory system is bidirectionally regulated by the temporal pattern of spontaneous activity before hearing onset, showing that pre-sensory activity instructs structural plasticity during synaptic maturation.

The collective message is unambiguous: the general principles governing learning cannot be inferred from any single model system. Rather than being governed by a universal mechanism, synaptic plasticity comprises a rich repertoire of learning rules specialized for the computational demands of individual synapses and circuits. Just as genomics, proteomics, and connectomics have provided unifying frameworks for understanding genes, proteins, and neuronal connectivity, a comprehensive plasticitome promises a comparable framework for understanding how neural circuits develop, adapt, learn, and malfunction. The Editorial’s authors hope their Research Topic will serve as a foundation for increasingly systematic efforts to chart the plasticitome—an atlas of synaptic learning rules spanning brain regions, synapse types, developmental stages, behavioral states, and disease that could ultimately transform how neuroscience thinks about the brain’s capacity to change.

Subject of Research: Synaptic plasticity diversity and the plasticitome framework

Article Title: Editorial: The diversity of the plasticitome

Article References: Chou, C. Y. C., Dudek, S. M., Bramham, C. R., Park, H., & Sjöström, P. J. (2026). Editorial: The diversity of the plasticitome. Frontiers in Synaptic Neuroscience, 18, Article 1969861. https://doi.org/10.3389/fnsyn.2026.1969861

Image Credits: AI Generated

DOI: 10.3389/fnsyn.2026.1969861

Keywords: plasticitome, synaptic plasticity, long-term potentiation, interneurons, neuromodulation, learning and memory, Editorial, diversity, scientific research