Engram · Grey Matter

An engram is the physical change that an experience leaves in the brain and that later lets the experience be recalled, and in current research it means a sparse set of neurons, active during learning, whose reactivation is enough to bring the memory back.

Engram. An engram is the physical change that an experience leaves in the brain and that later lets the experience be recalled, and in current research it means a sparse set of neurons, active during learning, whose reactivation is enough to bring the memory back.

Richard Semon coined the word in 1904, and for most of the century it named a trace no one could find. Susumu Tonegawa's laboratory made it something to point at. In mice engineered so that neurons switching on the activity gene c-fos during a chosen window also make a light-gated channel (optogenetics), the cells active while a mouse learned to fear one box were labelled in the dentate gyrus of the hippocampus. In 2012, light on those cells alone made the mouse freeze in a different box. In 2016 the same approach was turned on mouse models of early Alzheimer's disease: the mice learned a fear normally and seemed to forget it days later, yet light on the tagged cells brought the fear back, so the memory had been stored and only its retrieval had failed.

It is sparse and spread out. A small fraction of the cells in a region is tagged for one experience, and one memory is held by linked ensembles in the hippocampus, the amygdala and the cortex together.

Its substrate is the synapse. The tagged cells are bound by connections strengthened during learning (long-term potentiation), and in the Alzheimer's mice they carried fewer dendritic spines; repeated stimulation that restored the spines restored recall for up to six days.

Storing and retrieving can come apart. A trace can exist and stay silent, which reframes some amnesia as a lost route back to the memory.

The evidence is from mice. Tagging needs genetic engineering and light through an implanted fibre, so in people engrams are inferred from recordings and lesions.

An engram can survive while the way back to it fails.

The 2016 result separated a memory that is gone from one that is only out of reach, and the second can in principle be reached again.

Questions: How do researchers find the cells that hold one memory? They mark the neurons that are active while an animal learns and then switch those same neurons back on. In the mice used by Tonegawa's laboratory, a neuron that turns on the activity gene c-fos also makes a light-gated channel, but only during a window opened by withdrawing a drug from the diet, so only the cells active during that one experience are tagged. Shining light through an implanted fibre onto the tagged cells of the dentate gyrus made the mice freeze as if they were back in the box where they had learned to fear a shock. Cells tagged in a harmless box did not, which is what shows the tag captured that particular memory. Why could mice with early Alzheimer's disease not recall a fear they had learned? In the 2016 study the mice learned normally and failed only days later, and light on their tagged engram cells brought the fear back, so the memory was stored and the retrieval was broken. The engram cells in the dentate gyrus had fewer dendritic spines, the small protrusions where synapses sit, and the loss grew with age before any amyloid plaques appeared. A single burst of light restored recall for about a day; repeated high-frequency stimulation regrew the spines to normal numbers and recall lasted up to six days. The fault lay in the connections that lead back to the trace.