 Memories seem to be created in the blink of an eye. We can recall an event as soon as it is over. But is this actually the case? Scientists know the key signals for memory formation, but have lacked the tools to figure out the details, such as when and for how long particular molecules need to be active. By developing a more precise way of disrupting signaling, researchers have now defined this window for one of the most important molecules responsible for memory. Memories are recorded through changes in synapses or the connections between neurons. Many synapses are located on tiny bumps called spines that are found along the branching dendrites of neurons. When synapses are very active, like they are during learning or a memorable event, molecular signaling cascades turn on. These cascades strengthen synapses and make spines grow. The protein calcium-calmodulin kinase 2, or CAMK2, is required for both of these events. But precisely when and how long CAMK2 needs to act was unclear. To find out, researchers from the Mox Planck Florida Institute for Neuroscience, in collaboration with the National Institute of Physiological Sciences in Japan, created a light-activated version of AIP2, a CAMK2 inhibitor, by attaching a blue light sensor to it. Turning on the light activated the inhibitor and blocked CAMK2 activity almost instantaneously, while turning it off allowed CAMK2 activity to resume in less than one minute. With this tool, the researchers could manipulate the timing of memory-related CAMK2 activity with second-to-minute resolution. They first investigated the timing of CAMK2 activity required for spines to grow and synapses to get stronger. Normally, stimulating a synapse with a particular rhythm strengthens it and makes spines grow. But if the researchers shined the blue light while stimulating, thus turning off CAMK2 activity at the same time, those changes didn't happen. When they turned the light on just one minute after stimulation, the synapse grew and the spine got stronger, as expected. These results indicate that memory-related changes in synapse strength and structure only require a short one-minute burst of CAMK2 activity. Next, the researchers looked at the timing of CAMK2 activity needed for actual memory formation. They placed mice in the bright side of two rooms connected by a hole and trained the mice to avoid the dark room by giving them a tiny electrical shock whenever they entered it. In some mice, they also shined blue light to inhibit CAMK2 in the amygdala, a brain region important for this type of memory formation. Inhibiting CAMK2 activity in the amygdala during training, but not after, blocked learning. These results indicate that CAMK2 activity during training is necessary for learning, but CAMK2 activity after training is not. Using the new light-activated inhibitor, the team was able to define the brief window of CAMK2 activity needed for memory and associated synapse changes with an unprecedented level of precision. Because the same strategy can be used to make light-activated inhibitors of proteins in other signaling pathways, the method may also increase understanding of many biological processes and lead to treatments for the many conditions caused by abnormal signaling, such as intellectual disability or addiction.