 For most of us, the memories we make and the facts we learn every day don't seem to change who we are in meaningful ways. But in reality, learning and memory continuously alter our brain structure in ways that can actually be measured. How do these changes occur? Scientists at the Max Planck Florida Institute for Neuroscience and Duke University developed specialized sensors that show how key proteins work together to control tiny structural changes in our brains. Memories are encoded in the trillions of connections made between neurons. These connections, or synapses, are found along the branching dendrites of neurons where information is transferred. Many synapses occur on even more specialized structures called spines, which look like tiny bumps on dendrites. Sustained changes in spine structure are one way the brain stores and retrieves memories. These dynamic changes, collectively known as structural plasticity, often occur in clusters along dendrites, suggesting that spatial rules govern where structural plasticity occurs. One possibility is that biochemical information shared between clustered synapses sets these spatial rules by influencing actin, the protein that gives dendritic spines their shape. The ROGTPA's family proteins, RAC1, ROA, and CDC42, are particularly attractive candidates for carrying such signals because they regulate actin. But how do these proteins work together to control spine structure and thus learning and memory? To answer this question, the researchers first monitored the fluorescent signals emitted by special sensor molecules planted in live brain slices from rodents. These signals showed where and for how long the GTPase proteins were active within neurons. Once the sensors were in place, the scientists coaxed single spines into changing shape by releasing the excitatory neurotransmitter glutamate onto individual spines using a laser. This stimulation activated all three GTPases simultaneously, but their activation patterns differed in ways that could explain key features of stimulation-induced changes in spine structure. How far activity spread was one of the most notable differences. While the stimulation-induced activation of RAC1 and ROA spread beyond the stimulated spine to the surrounding dendrite and nearby spines, CDC42 activity was restricted to the stimulated spine. Without spreading of both ROA and RAC1 activation into nearby spines, stimulating one spine could no longer facilitate similar changes in nearby spines. Thus, the partitioning of CDC42 signaling may explain how changes in spine structure can be specific to a single active synapse, while the spreading of ROA and RAC1 activity may explain how stimulation can influence nearby unstimulated synapses. The patterns of GTPase activation may also explain the influence of other molecules on plasticity. For example, while the protein BDNF is known to play an important role in plasticity and learning, the way it can promote but not induce plasticity has remained unclear. The researchers found that BDNF was critical for activating RAC1 and CDC42 but not ROA, or some but not all of the components required for plasticity. However, partial GTPase activation due to BDNF or signal spreading likely makes activation of the remaining components easier. This finding highlights how BDNF-dependent and independent pathways, as well as diffuse and restricted signals, shape stimulation-induced changes in spine structure. Taken together, differences in how the proteins that regulate actin are activated reveal much about the changes in spine structure that allow the brain to store and retrieve memories. Better understanding the signaling pathways that link synaptic activity to structural changes in the brain through proteins like GTPases may aid in the development of therapies for treating memory disorders.