Observational sequence learning

The objective of this thesis was to assess whether human and nonhuman animals can learn about serial order by observation of a conspecific performing a task designed to assess sequence knowledge. With regard to human sequence learning, two questions were asked. The first, whether procedural knowledge of sequences, typically acquired following direct practice, can be gained following observation, the second, how such knowledge is represented and by which mechanisms. Using an adaptation of a serial reaction time task (Nissen and Bullemer 1987) and a free generation task the implicit versus explicit knowledge distinction was examined. Evidence reported in this thesis suggests that humans are capable of observational sequence learning on an implicit task and sequence knowledge is stored in terms of a motor code. However, experiments suggest that observational sequence learning may be limited to situations where the sequences are unique in structure and where there are 16 exposures of the sequence per block of trials. There was no evidence for a dissociation in observer performance between measures designed to assess procedural and declarative knowledge and it was concluded that this was not an appropriate distinction to make when characterising observer performance on the SRT task. The experiments with animals were pursued using a two lever method, and both free operant and fixed trial procedures. Under free operant conditions, there was no evidence of observational sequence learning following observer pretraining on individual sequence items or when pretraining was dispensed with. There was, however, evidence of learning individual sequence items in the form of recency effects. The fixed trial procedure, designed to enhance the salience of the demonstrated sequence, decreased the delay inherent in the free operant procedure between the observation and testing phases and to reduce the effects of olfactory cues, yielded no observational sequence learning. Under these conditions a primacy effect emerged. The fixed trial procedure is a significant innovation to social learning research and recommendations are made as to how this procedure can be improved