Creation and Recollection of Artificial Memories

Post by Deborah Joye

What's the science?

As animals move through the world, they learn to associate certain environmental cues with events which help them survive. These memory associations have been localized to specific brain regions and underlying circuits. Specifically, the way that mice create odor associations is well-characterized and occurs similarly across individuals. If our detailed understanding of odor memory in mice is correct, it leads to an interesting question: can a false memory be created through direct stimulation of the brain? And can this memory be recalled by something in the real world, as if it actually happened? This week in Nature Neuroscience, Vetere and colleagues use association training along with direct cell-type- and region-specific brain stimulation to demonstrate that both good and bad artificial memory associations can be created in mice and recalled by a real-world cue.

How did they do it?

The authors first trained mice to form a real odor association by pairing a specific odor (acetophenone) with a mild foot shock. Since memory associations depend on presentation of the odor and the foot shock right after one another, the authors included conditions where the odor or the foot shock were presented independently or were presented 24 hours apart from one another (too far apart for memory associations to form). In the memory test, mice were put into a box with two chambers – one with the trained odor and the other with a new odor. The idea is that if a memory of the odor-foot shock pairing has been formed, the mouse will avoid the compartment containing the foot-shock-paired odor.

The authors then repeated this experimental structure multiple times. First, to test whether the odor association could be formed by direct brain stimulation, the authors genetically altered acetophenone-specific olfactory cells to be activated by a laser (called optogenetic stimulation). Mice were then exposed to the same two-chamber box containing either acetophenone or another smell. Second, to test that both the odor and the foot-shock could be created using only direct brain stimulation, the authors used optogenetics to activate both olfactory cells and cells in the brain associated with positive or negative experience (laterodorsal tegmental or lateral habenula inputs to the ventral tegmental area, respectively). Finally, since cells in the basolateral amygdala are important for memory associations, the authors tested whether this region is likewise necessary for artificial memory associations by chemogenetically silencing it. They virally expressed an inhibitory DREADD (designer receptor exclusively activated by designer drug) to turn off basolateral amygdala cells in the presence of a specific chemical. The authors then repeated the memory association experiment, using the inhibitory DREADD to block basolateral amygdala activity a subset of mice.

What did they find?

The authors found that mice formed memory associations between an odor and a foot-shock, as expected. Direct optogenetic stimulation of acetophenone-sensitive olfactory cells paired with foot shock also produced a memory association. This memory association could be recalled by mice, as evidenced by an avoidance of the compartment containing the actual acetophenone odor. Optogenetic stimulation of olfactory cells paired with stimulation of negative experience cells produced a behavioral aversion to the acetophenone odor, like mice that had been exposed to a real foot shock. In contrast, optogenetic stimulation of olfactory cells paired with stimulation of positive experience cells produced a behavioral attraction to the acetophenone odor. In all instances, artificial memory associations were only created if stimulation of the brain regions occurred close together in time, as with real memory associations. The authors also found that real and artificial memory associations engaged similar neural circuits, as shown by markers of cellular activation. Finally, when researchers chemogenetically blocked basolateral amygdala activity with DREADDs, expressions of both real and artificial memory associations were lost.


What's the impact?

This study successfully creates a fully artificial memory in mice through direct brain stimulation. The characteristics of this artificial memory were like a natural memory: time dependent, similar brain circuits, behavioral responses were specific to the trained cue, and memory expression depended on the basolateral amygdala. This is the first study to show that artificially created memories can be recalled by a real-world cue. This study presents a valuable window into how memory associations are created and integrated with real-world experience.


Vetere et al., Memory formation in the absence of experience, Nature Neuroscience (2019).Access the original scientific publication here.