Brain function depends on neural communication in circuits, which in turn depends on neurons forming connections. New memories and other responses to experience are therefore often reflected in physical changes in these connections, called synapses. By developing technologies with MIT Mechanical Engineering Professor Peter So to observe changes in individual synapses in live animals as often as daily and for weeks at a time, Elly Nedivi has discovered crucial patterns of when and how synapses change.

By 2012, researchers had begun tracking the ongoing remodeling of excitatory synapses, which reside on spines that protrude from the long, spiny, root-like branches that neurons extend to receive incoming messages from other neurons. But they lacked a way to do the same for inhibitory ones, which provide important balance and nuance in neural “conversations” and often reside on the shafts of dendrites. Nedivi’s lab debuted a method of imaging dendrites and inhibitory synapses in different colors. This novel capability allowed the lab to discover that inhibitory synapses were uniformly arrayed along dendrite shafts but that about 30 percent of inhibitory synapses inhabited spines. Of those, they were much more likely to be out near the tips of dendrites than in closer to the cell body. When they examined how inhibitory synapse patterns changed with experience (e.g. when an animal’s eye was closed for a few days), they found that even though spines didn’t show net changes, inhibitory synapses along dendrite shafts disappeared at twice the normal rate. They also found that inhibitory synapse changes closely accompanied individual spine changes. These observations provided new evidence of close orchestration of inhibitory and excitatory changes in response to experience.

Above: A neuron with dendrites labeled with white to denote inhibitory synapses and yellow to denote excitatory synapses.

In 2016 the lab followed up with another key set of discoveries. This time they added a third color to additionally track excitatory synapses and instead of looking every few days over the course of weeks, they looked every day for nine straight days. They found that most spines just housed an excitatory synapse and that most of those stayed stable. Spines that didn’t have an excitatory synapse appeared and disappeared often. The most interesting spines, though, had both an excitatory and an inhibitory synapse side by side. These spines and their excitatory residents were very stable but further observations showed that the inhibitory synapses on those spines were absolutely not. Instead, on a daily basis they might appear, disappear and then often reappear at the same spots. Moreover, the pattern of their appearance and disappearance varied with visual experience (e.g. if the animal’s eye was closed for a few days). The evidence strongly suggested that neurons use the rapid addition and removal of inhibitory synapses to dynamically moderate more stable excitatory input – not only within individual spines but possibly across the whole cell.