Memory consolidation in Hermissenda crassicornis.

Abstract

Experiments with shell-less molluscs (Aplysia and Hermissenda) have revealed a number of processes that underlie by these organisms and also by some vertebrates. Hermissenda, for example, shows significant Pavlovian conditioning capabilities (1). Follow-up investigations on these molluscs dealt with the sensory stimuli needed for short-term (STM) and long-term (LTM) (2). The relationship of the two memories with in vitro changes in excitatory post-synaptic potentials (EPSPs) have also been investigated in the neural networks of these organisms (3). Many studies of the molecular aspects of these two different regimes have led to quite detailed descriptions of the events (4, 5, 6). Both Aplysia and Hermissenda have been tested for their recall of induced behavioral modifications after one, two, or many conditioning events (CEs). In Aplysia, the EPSP component of produced by 1 CE was compared to that produced by 5 CEs (3). In Hermissenda, the comparison was made between 2 CEs and 9 CEs (2). Five to ten minutes after finishing one or two conditioning events, both animals exhibited significant behavioral recall (i.e. STM); but there was no recall after an hour or more (i.e. no LTM). The larger numbers of CEs, however, did induce LTM in both species. Since STM and LTM are clearly responding to a different set of conditions, we focused first on what might inhibit or block STM. This problem was partially anticipated in 1900, according to McGaugh (7) who cited Muller and Pilzecker as having found that memory of newly learned information was disrupted by the of other information shortly after the original learning (8). This concept led us to test, in Hermissenda, whether STM recall (at 5 min) might be blocked simply by the input of additional information (i.e. extraneous sensory stimuli) if the latter were applied within the first 5 min after conditioning. The initial results of the blocking experiments, which showed that the simple sensory inputs blocking STM also blocked LTM, then led to the hypothesis that temporal consolidation of LTM could be detected by measuring when the blocking sensory input was no longer effective. Hermissenda (Sea Life Supply, Sand City, CA) were tested with 2 and 9 paired CEs for induction of STM and LTM. Conditioning events consisted of exposing the animals to 6 s of bright, white light (CS) explicitly paired with 4 s of strong orbital agitation (US) following a 2-s onset delay with an inter-trial interval of 1 min. Recall of the behavioral modification induced by associative conditioning was assessed by recording the animal's change in foot length when presented with 6 s of light alone. The conditioned response (CR) was foot contraction, the unconditioned response (UR) was foot elongation (9). Two paired conditioning events initiated behavioral recall after 5 min but not after 90 min; the LTM input of 9 pairings was recalled at both 5 and 90 min (Fig. 1A). The small and non-overlapping S.E.Ms for each point indicate statistical significance (P = <0.01, t = 3.18). After giving the animals the paired CS and US stimuli leading to STM, we tested two simple paradigms of blocking sensory stimuli. The first was a modification of the conditioning stimuli: dim orange light and very slow orbital rotation. The second blocking stimulus tested consisted of rotating the tray containing the animals upside down and, after 5 s, rotating it upright again (rotational block). Both experimental paradigms blocked STM and LTM (Fig. 1B). To determine the temporal specificity of LTM in Hermissenda, the following experiments were done. Animals were trained with 9 CEs, and the CR was measured at the usual 90 min. However, at selected time intervals (2, 25, 50, 55, 60, 65 min) post-conditioning, the animals were rotationally blocked. Control animals received only the 9 paired CEs. When the animals' behavior was plotted, a clear and decisive LTM consolidation interval in Hermissenda appeared; consolidation occurred between 55 and 60 min (Fig. 1C). Presentations of rotational blocking prior to 55 min totally blocked consolidation. However, the stimulus given after 60 had no blocking effects, and the animals demonstrated the CR. The consistency of and surprisingly little variability in the response among the majority of the animals indicated the robustness of the paradigm. When the data were analyzed with t-test and F-test statistics, they were found to be highly significant, whether compared between data points or to zero (P = <0.001, t = 15.24; F-value, inf). 182