Integration of Parallel Opposing Memories Underlies Memory Extinction

Abstract

SummaryAccurately predicting an outcome requires that animals learn supporting and conflicting evidence from sequential experience. In mammals and invertebrates, learned fear responses can be suppressed by experiencing predictive cues without punishment, a process called memory extinction. Here, we show that extinction of aversive memories in Drosophila requires specific dopaminergic neurons, which indicate that omission of punishment is remembered as a positive experience. Functional imaging revealed co-existence of intracellular calcium traces in different places in the mushroom body output neuron network for both the original aversive memory and a new appetitive extinction memory. Light and ultrastructural anatomy are consistent with parallel competing memories being combined within mushroom body output neurons that direct avoidance. Indeed, extinction-evoked plasticity in a pair of these neurons neutralizes the potentiated odor response imposed in the network by aversive learning. Therefore, flies track the accuracy of learned expectations by accumulating and integrating memories of conflicting events.

Learning with reward prediction errors in a model of the Drosophila mushroom body

“…Appetitive conditioning does indeed potentiate responses in MB-V3/α3 and MVP2/γ1-pedc approach MBONs (11; 31), and depress responses in M4β /β 2mp and M6/γ5β 2a avoidance MBONs (10). Similarly, removal of an expected aversive stimulus, which constitutes a positive RPE, depresses M6/γ5β 2a avoidance MBONs (14). In addition, aversive conditioning depresses responses in MPV2/γ1-pedc MB-V2/α2α 2 approach MBONs (18; 25), and potentiates responses in M4β /β 2mp and M6/γ5β 2a avoidance MBONs (10; 24).…”

“…For example, a reduction in electric shock magnitude, after an initial period of training, would elicit an excitatory (inhibitory) response in appetitive (aversive) DANs. Felsenberg et al (13;14) provide indirect evidence for this. The authors trained flies on a CS+, then re-exposed the fly to the CS+ without the US.…”

“…MBONs provide excitatory feedback to their respective DANs (12)(13)(14). In both the valence-specific (VS) and mixed-valence (MV) models, M ± synapses onto D ∓ with unit synaptic weight.…”

“…To date, there are only indirect lines of evidence for RPE signals in MB DANs. DAN activity is modulated by feedforward reward signals, but some DANs also receive excitatory feedback from MBONs (12)(13)(14)(15), and it is likely this extends to all MBONs whose axons are proximal to DAN dendrites (16). MBON activity is, in turn, modulated by DANs via plasticity at KC→MBON synapses (10; 17; 18).…”

Multilevel feedback architecture for adaptive regulation of learning in the insect brain

“…However, the extent to which learning in a specific compartment is regulated by the output from its own compartment or from other compartments is still unclear. A few direct anatomical connections from MBONs to modulatory neurons have been identified 13,14,[45][46][47][48][49]53,54 , but possible indirect connections via intermediate feedback neurons have not been investigated. In total, despite a good understanding of the structure and function of the core components of the MB 14,47,48,58 , the circuits presynaptic to modulatory neurons that regulate their activity have re-mained largely uncharacterized, in both adult and larval Drosophila.…”