Memory Engrams in a Synapse-Specific manner

 

Science  15 Jun 2018: Vol. 360, Issue 6394, pp. 1182-1183

Crystallizing a memory

Steve Ramirez

Department of Psychological and Brain Sciences, Boston University, Boston, MA, USA.

[paraphrase]

What is the physical basis of memory? What does it take to retrieve a memory in the brain? What would it take to activate or erase memories? In the early 20th century, the German zoologist Richard Semon coined the term “engram” to denote the physical manifestation of a memory in the brain. Two decades later, Canadian psychologist Donald Hebb posited a physiological correlate for learning and recollection: The process of learning strengthens the connections, or synapses, between neurons, which leads to the development of brain-wide cell assemblies that undergo changes in their structural and functional connectivity. The coordinated activity of these assemblies—called ensembles, traces, or engrams—that occurs during learning (memory formation) is thought to be reengaged during recall and thereby forms a stable neuronal correlate of memory. As subsequent memories are formed, the dynamics of these assemblies evolve and provide preexisting scaffolds to influence how the brain processes the variety of memories an organism forms. Studies by Abdou et al. and by Choi et al. develop new technologies to visualize discrete engrams in the brain and modulate them in a synapse-specific manner to understand memory strength and memory restoration from an amnestic state. This improved understanding could eventually be translated to modulate memories to alleviate maladaptive memory states.

Hebb's conceptualization of memory in the brain became an oft-quoted creed in brain science: Neurons that fire together    wire together. In the spirit of Semon, cells that are active during learning, that undergo enduring learning-induced changes, and that facilitate recollection are referred to as engram cells. A physical manifestation of Hebb's principle, of engram cells communicating and linking with one another during learning, has been recently demonstrated in mice. A discrete ensemble of hippocampal cells that were simultaneously active during learning preferentially strengthened their structural and functional connectivity relative to quiescent cells. It was a remarkable demonstration of hippocampal engram cells firing together at the time of learning and physically interlinking together to facilitate memory retrieval.

Choi et al. developed an activity-dependent strategy to tag and visualize not just active engram cells but also active synapses between engram cells and non-engram cells. They used a clever trick: They engineered a system in which presynaptic and postsynaptic membranes of a neuron had complementary green fluorescent protein (GFP) fragments that reconstituted a functional GFP on synapse formation. Excitingly, by using fluorescent proteins of different colors, the researchers were able to visualize two different presynaptic neurons that projected to a single postsynaptic cell. This allowed the authors to measure how learning modulates connectivity    between engram cells, engram to non-engram cells, between non-engram cells, and non-engram to engram cells. Their results are striking: Learning induced preferential increases in synaptic connectivity specifically between engram cells and not between non-engram cells. They also found that, although weak and strong fear memories activated a similar proportion of cells in the hippocampus, a stronger fear memory elicited stronger connectivity (that is, a higher density of synapses and potentiation) specifically between engram cells.

A cell ensemble can process multiple memories, but how the same population of cells can encode separate memories has remained unclear. Abdou et al. combined cutting-edge techniques to visualize and directly modulate discrete memories in a synapse-specific manner. They used an auditory fear conditioning task in which a tone is paired with a foot shock such that mice subsequently display fear responses to hearing the tone without foot shock. Of the myriad of neural circuits involved in a memory, the auditory cortex (AC), the medial geniculate nucleus (MGN), and the lateral amygdala (LA) are key nodes involved in processing an auditory fear memory.

In an attempt to induce complete amnesia, the researchers both blocked protein synthesis and induced autophagy—a degradation process in which the cytosolic constituents of a cell are recycled.

The researchers demonstrated that the acquisition of two fear memories engaged two different sets of synapses from the AC and MGN, which interacted with the overlapping neural ensembles in the LA—as such, each memory could be individually, and lastingly, restored or suppressed by modulating the activity levels of each of their respective synapses in the LA.

These data provide a tantalizing demonstration of linking and controlling an individual memory amid the ocean of experience that a mouse remembers. They provide a glimpse into the microstructure of recollection. Embedded within and across subregional activity are discrete circuits with unique histories that sculpt the morphological and physiological properties of a neuron and, by extension, of a memory. Engrams are not localized to a single X-Y-Z coordinate in the brain, but rather appear to be distributed with key nodes in the brain being necessary, sufficient, or both, to regulate individual components of memory. Semon's engram recruits Hebb's assemblies in a brainwide manner.

[end of paraphrase]

 

Return to — Memory

Return to — Associative nature of Memory