Dendritic spines are small membranous protrusions that receive most excitatory synaptic input. During LTP, spine volume and surface area increase, correlating with synaptic strengthening. These morphological changes are driven by actin polymerization and myosin motors in response to Ca2+ influx, allowing the physical structure of neural circuits to be refined by experience.
Image dendritic spines using two-photon microscopy before and after LTP induction. Compare spine density and morphology across different developmental stages or after environmental enrichment.
You already know that NMDA receptors act as coincidence detectors, opening only when the postsynaptic membrane is depolarized while glutamate is bound, and that the resulting calcium influx triggers long-term potentiation. But LTP is not just an electrical or chemical change — it has a physical, structural counterpart. The tiny protrusions on dendrites where most excitatory synapses sit, called dendritic spines, actually grow larger and change shape when a synapse is potentiated. This structural plasticity is what converts a transient electrical event into a lasting architectural modification of the brain.
Dendritic spines are remarkably small — typically less than one micrometer in length — yet they are packed with sophisticated molecular machinery. Each spine is essentially a biochemical compartment, partially isolated from the parent dendrite by its narrow neck. This compartmentalization means that calcium signals and activated signaling molecules remain concentrated within the stimulated spine rather than flooding neighboring synapses. When Ca²⁺ enters through NMDA receptors during LTP induction, it activates CaMKII (calcium/calmodulin-dependent protein kinase II), which in turn triggers a cascade that reorganizes the spine's internal skeleton. The key structural element is actin — spines contain almost no microtubules but are densely packed with actin filaments. Rapid actin polymerization physically pushes the spine membrane outward, increasing the spine's volume and surface area within minutes of stimulation.
Spines come in several morphological categories that reflect their functional state. Thin spines have a narrow neck and small head — they are highly motile, appear and disappear frequently, and are thought to represent "learning spines" that sample potential synaptic partners. Mushroom spines have a large, bulbous head and are more stable — they represent mature, strengthened synapses and are sometimes called "memory spines." Stubby spines lack a clear neck and are common during early development. When LTP occurs at a thin spine, it tends to enlarge into a mushroom shape: the head expands, more AMPA receptors are inserted into the postsynaptic membrane, and scaffolding proteins accumulate to stabilize the new configuration. This enlargement is not merely cosmetic — the increased surface area accommodates more receptors, the wider spine head has lower electrical resistance, and the expanded postsynaptic density provides more docking sites for signaling molecules.
The bidirectional nature of structural plasticity is equally important. Just as LTP enlarges spines, long-term depression (LTD) shrinks them and can cause them to retract entirely. This pruning is essential for circuit refinement — without it, the brain would accumulate synapses without bound, losing the signal-to-noise ratio that makes neural computation meaningful. During development, spine density peaks in early childhood and then declines through adolescence as experience-dependent pruning eliminates weak connections and stabilizes strong ones. Disruptions in this process are associated with neurodevelopmental disorders: excess spine density is found in autism spectrum disorder, while excessive pruning has been linked to schizophrenia. The physical structure of dendritic spines is therefore not a passive scaffold but an active participant in learning, memory, and brain development.