An investigation of the implicit learning of metrical and non-metrical temporal patterns

Abstract

Implicit learning (IL) occurs unintentionally and without awareness. The majority of the literature on IL relates to the investigation of ordinal patterns, that is, patterns where the properties of stimuli systematically vary along one or more categorical dimensions (e.g. a repeating sequence of spatial locations). The IL of temporal patterns, that is, patterns where the temporal intervals between consecutive stimuli vary systematically, has received considerably less attention. Moreover, results of previous studies are mixed regarding whether temporal patterns can be learned independently of an ordinal pattern, or when an ordinal sequence is unpredictable. To examine the IL of temporal patterns, five experiments and a model""based analysis are presented in this thesis. The present program of research investigated the IL of temporal patterns and examined the conditions under which the IL of temporal patterns is observed. Based on suggestions that learning of temporal patterns is underestimated in the serial reaction-time task when the upcoming stimulus is unpredictable (Ullén and Bengtsson, 2003), Experiment Sets 2 and 3 investigated the conditions under which temporal learning is observed. According to probabilistic uncertainty, a response cannot be prepared to an unknown stimulus even if the timing of the upcoming stimulus is predictable. Thus, it is hypothesized that temporal learning is observed in a single response serial reaction-time task and an immediate recall task, but that learning is underestimated in a multiple response serial reaction-time task when stimulus identities are unpredictable. Generation and recognition tasks based on the process dissociation procedure (Jacoby, 1991) were used to assess the degree to which learning was implicit. Furthermore, a model for separating conscious and unconscious processes, which is based on a model proposed by Buchner, Steffens, Erdfelder, and Rothkegel (1997), was used to analyze recognition data. Hypotheses regarding how temporal patterns are learned were derived from theories of rhythm and time perception. Rhythm is the systematic sequencing of events in relation to timing, accent, and grouping (Patel, 2008). Metrical patterns are patterns from which an underlying isochronous (evenly spaced) pulse can be abstracted and where event-onsets periodically occur according to equal groupings of pulses. Non-metrical patterns are patterns that cannot be conceived in terms of an isochronous pulse, or equal grouping of pulses (Essens and Povel, 1985). The dynamic attending theory and metric binding hypothesis (Jones, 2009) posit that attentional oscillators guide attention to periodic points in time at one or more periodicities. Based on the dynamic attending theory and the metric binding hypothesis, it was hypothesized that metrical and non-metrical patterns can be implicitly learned and that metrical patterns are learned more readily than non-metrical patterns.

Date of Award	2012
Original language	English