Associative learning in the cnidarian Nematostella vectensis

Abstract

The ability to learn and form memories allows animals to adapt their behavior based on previous experiences. Associative learning, the process through which organisms learn about the relationship between two distinct events, has been extensively studied in various animal taxa. However, the existence of associative learning, prior to the emergence of centralized nervous systems in bilaterian animals, remains unclear. Cnidarians such as sea anemones or jellyfish possess a nerve net, which lacks centralization. As the sister group to bilaterians, they are particularly well suited for studying the evolution of nervous system functions. Here, we probe the capacity of the starlet sea anemone Nematostella vectensis to form associative memories by using a classical conditioning approach. We developed a protocol combining light as the conditioned stimulus with an electric shock as the aversive unconditioned stimulus. After repetitive training, animals exhibited a conditioned response to light alone-indicating that they learned the association. In contrast, all control conditions did not form associative memories. Besides shedding light on an aspect of cnidarian behavior, these results root associative learning before the emergence of NS centralization in the metazoan lineage and raise fundamental questions about the origin and evolution of cognition in brainless animals.

Keywords: associative learning; cnidaria; cognition; learning; nervous system evolution.

Fig. 1.

(A) Experimental setup for aversive classical conditioning experiments. A white light (CS, measured illuminance ≃ 300 lux) is paired with an electric shock (ES, voltage of ≃6 V) as US. When the ES is applied, the animal retracts immediately (screenshots at 0 s and 3 s) and extends to its initial state in less than 3 min (screenshot at 180 s). (Scale bar: 2 mm.) (B) Timeline of conditioning experiment protocol. A 50-min training session consists of 20 rounds of training (2.5 min each). (C) Stimuli presentation during training for each condition: CS and US presented simultaneously (paired condition, green), CS and US explicitely unpaired (unpaired, blue), CS alone (light only, pink), and US alone (shock only, grey). (D) Manual scoring of animals retracting before (pretest, yellow) and after conditioning for each condition. Error bars: 95% CI. Chi square test, ${\chi }^2=10.897,$ P = 0.001, n_subj=36 animals/condition from n = 2 independent experiments.

Fig. 2.

(A) Screenshot of a video after conditioning, showing three points per individual. Body length as sum of the mouth–midpoint and midpoint–foot segments’ lengths (see SI Appendix for details). The lower panels show screenshots of individual 2 before (t1), during (t2), and after (t3) CS application. (Scale bars: 5 mm.) (B) Plots of body length variation during video recordings after conditioning. Curves from individual animals (thin grey lines), mean (thick line), and SE of the mean (shaded area around the mean) for the paired and unpaired conditions. The right panel (paired condition) corresponds to the video depicted in (A), and the body length variation of the individual 2 is highlighted [thin dark green line, time points corresponding to the screenshots in the lower panel in (A)]. (C) Proportion of animals retracting more than 10% of their initial body length during CS application. Error bars: 95% CI. Chi-square test, ${\chi }^2= 9.6204$ , P = 0.002, animals from n = 5 independent experiments. (D) Maximum retraction measured during CS application in unpaired (blue) and paired (green) conditions. Mann–Whitney U test, W = 2,620, P = 0.008, n = 5 independent experiments.