Memory decay and susceptibility to amnesia dissociate punishment--from relief-learning

Abstract

Painful events shape future behaviour in two ways: stimuli associated with pain onset subsequently support learned avoidance (i.e. punishment-learning) because they signal future, upcoming pain. Stimuli associated with pain offset in turn signal relief and later on support learned approach (i.e. relief-learning). The relative strengths of such punishment- and relief-learning can be crucial for the adaptive organization of behaviour in the aftermath of painful events. Using Drosophila, we compare punishment- and relief-memories in terms of their temporal decay and sensitivity to retrograde amnesia. During the first 75 min following training, relief-memory is stable, whereas punishment-memory decays to half of the initial score. By 24 h after training, however, relief-memory is lost, whereas a third of punishment-memory scores still remain. In accordance with such rapid temporal decay from 75 min on, retrograde amnesia erases relief-memory but leaves a half of punishment-memory scores intact. These findings suggest differential mechanistic bases for punishment- and relief-memory, thus offering possibilities for separately interfering with either of them.

Keywords: Drosophila melanogaster; memory stability; punishment-learning; reinforcement; relief-learning; retrograde amnesia.

Figure 1.

Different decay of punishment- and relief-memory. (a) Schematic of experimental design. (b) Punishment- and relief-memory assayed after retention periods of either 25, 50, 75, 240 min, or 24 h. Light shading indicates a difference from the scores at the earliest retention period (25 min), lack of fill indicates lack of significance from chance (i.e. from zero). For the 240 min retention period, relief-memory scores were neither different from the scores obtained at 25 min, nor from zero (hatched fill). Box-whisker plots (see §2 for details) show learning indices, positive values indicating conditioned approach to the trained odour (i.e. relief-memory) and negative values conditioned avoidance (i.e. punishment-memory). *p < 0.05 in KW-tests.

Figure 2.

Punishment- and relief-memory differ in sensitivity for cold-amnesia. (a) Schematic of experimental design. (b) Punishment- and relief-memory assayed after 120 min retention period. For each kind of memory, one group underwent cold-amnesia treatment 60 min after training, while the other did not (i.e. control). Differences in shading indicate p < 0.05 in the respective MWU test, and lack of shading indicates lack of significance from chance performance (i.e. from zero).