Learning and memory in the Port Jackson shark, Heterodontus portusjacksoni

Abstract

Basic understanding of the fundamental principles and mechanisms involved in learning is lacking for elasmobranch fishes. Our aim in this study was to experimentally investigate the learning and memory capacity of juvenile Port Jackson sharks, Heterodontus portusjacksoni. Sharks (N=30) were conditioned over a 19-day period to associate an underwater LED light or stream of air-bubbles [conditioned stimulus (CS)] with a food reward [unconditioned stimulus (US)], using three procedures (delay, trace and control). During experiments, the CS signalled at a random time between 180 and 300s for 30s (six times per day). For the delay the US overlapped in time with the CS, for the trace the US delivered 10s after the CS and for our control the US was delivered at random time between 180 and 300s after the CS. H. portusjacksoni sharks trained in all procedures improved consistently in their time to obtain food, indicative of Pavlovian learning. Importantly, the number of sharks in the feeding area 5s prior to CS onset did not change over time for any procedures. However, significantly more sharks were present 5safter CS onset for delay for both air-bubble and light CS. Sharks trained in the delay and trace procedures using air-bubbles as the CS also displayed significantly more anticipatory behaviours, such as turning towards the CS and biting. Sharks trained with the light CS did not exhibit such behaviours; however, trace procedural sharks did show a significant improvement in moving towards the CS at its onset. At 20 and 40days after the end of the conditioning experiments, some sharks were presented the CS without reward. Two sharks trained in the delay procedure using air-bubbles as the CS exhibited biting behaviours: one at 20 and the other at 40days. This study demonstrates that H. portusjacksoni have the capacity to learn a classical conditioning procedure relatively quickly (30 trials during 5days) and associate two time-separated events and retention of learnt associations for at least 24h and possibly up to 40days.