Learning not to be scared

October 8, 2018

Fear memories allow animals to predict and avoid possibly dangerous situations. However, if an expected threat does not happen, it is also important to learn to be less fearful. This re-learning process is called memory extinction. Although extinction has been postulated for many years to involve parallel memories, it is unclear how and where extinction memories are formed and how they compete with the original memory to neutralise learned fear.

In a recent study published in Cell Johannes Felsenberg and collaborators of the Waddell lab studied the neural mechanisms of extinction in the small brain of Drosophila. They discovered that extinction results from competition between two memories of opposing valence.

Flies can learn to avoid an odour that they experienced with unpleasant electric shock. However, if they are exposed to the odour again after training without shock, this reduces, or extinguishes, their learned avoidance behaviour.

Strikingly, extinction of aversive memory required dopamine-releasing neurons, some of which code for rewards, such as sugar. This suggests that flies perceive the absence of expected danger as something good.

Previous work from the Waddell lab established that good and bad odour memories are stored as changes in the efficiency of odour-specific input to different mushroom body output neurons. Looking for learning-induced changes at these locations in the fly brain allowed Pedro Jacob to see the co-existence of the original aversive memory, and a new extinction memory.

To understand how the two memories interact Amelia Edmondson-Stait, Markus Pleijzier and Nils Otto with collaborators in Cambridge and Janelia Research Campus traced the fine-structure of both neurons from a recently acquired electron microscope volume of the entire brain of the fly. Detailed reconstruction revealed that the two neurons housing the aversive and extinction memories are directly connected. The specific placement of inhibitory connections allows one neuron to gate information flow in the other. By directing his recordings to the processes downstream of the inhibitory synaptic connections, Pedro Jacob could see the two memories interact.

The full paper can be viewed here.

Johannes Felsenberg, Pedro Jacob and Nils Otto discuss the work in a short film entitled Not Bad is Good