Principles of parsimony: fMRI correlates of beat-based versus duration-based sensorimotor synchronization.

Abstract

(ProQuest: ... denotes formula omitted.)The temporal processing of successive intervals requires two different mechanisms: beat-based and duration-based timing (Essens & Povel, 1985; Grahn & McAuley, 2009; Keele, Nicoletti, Ivry, & Pokorny, 1989; Pashler, 2001; Teki, Grube, & Griffiths, 2012), which seem to engage different cerebral networks (Sakai et al., 1999; Teki, Grube, Kumar, & Griffiths, 2011). Most studies comparing beat-based and duration-based timing focused on perceptual tasks (Grube, Cooper, Chinnery, & Griffiths, 2010; Grube, Lee, Griffiths, Barker, & Woodruff, 2010; Teki et al., 2011) or memory tasks (Sakai et al., 1999). Despite the fact that timing perception and production involve different cerebral networks- though with some overlap (Coull, Cheng, & Meck, 2011; Wiener, Turkeltaub, & Coslett, 2010)-no studies reported human neuroimaging results of direct comparison between beat-based and duration-based sensorimotor synchronization (SMS). The reason might ensue from the difficulty to create SMS tasks in an irregular context that can be compared to corresponding tasks in a regular context.Duration-based timing requires encoding and retaining the absolute duration of discrete intervals. Beat-based timing involves an internal pulsation emerging from the perceived underlying regularity (or beat) of a sequence. The beat is a psychological construct: listeners can extract a pulsation from an auditory signal without it being physically present (Honing, 2013; Sethares, 2007) and without effort (Grahn & Rowe, 2013). Beat-based timing uses this underlying pulsation as a reference (Teki et al., 2011). Rhythmic patterns based on small-integer ratios, such as the 2:1 pattern, induce internal pulsation more strongly than those based on more complex ratios (Parncutt, 1994).SMS refers to synchronizing actions with external events and implies anticipation (see Repp, 2005 for a review). Many studies have shown the benefits of metrical beat subdivisions (i.e., based on integer ratios) on synchronization accuracy (e.g., Patel, Iversen, Chen, & Repp, 2005; Repp, 2003; Wohlschlager & Koch, 2000). In SMS tasks, error correction responses occurred for phase perturbations of the stimuli introduced at the level of the subdivision, that is, not interfering with the expected intertap intervals (ITI), even for subdivisions not directly preceding the target tone (Repp, 2008). This suggests hierarchical metrical processing: all temporal information available contributes to creating an internal model of the stimulus. Thus, regular patterns of noninteger ratios are likely to engage error-correction mechanisms. Such patterns tend to be adjusted and produced as integer ratio patterns (Essens, 1986; Franek , Radil, & Indra, 1988; Sakai et al., 1999).The cortical and subcortical structures most cited for their implication in both timing perception and production are the cerebellum, the supplementary motor area (SMA), and the basal ganglia (e.g., Coull et al., 2011; Grahn & Brett, 2009; Ivry & Spencer, 2004; Matell & Meck, 2000; Thaut, 2003; Wiener et al., 2010; Wing, 2002). In a review, Coull and collaborators (2011) reported differential activations for perception versus production tasks. Both types of tasks, though more rarely perception tasks, resulted in activation of the cerebellum, especially of lobule VI. At the cortical level, perceptual timing tasks activated almost exclusively the pre-SMA. Motor timing tasks also activated the SMA proper. At the level of the basal ganglia, motor tasks involved mostly the putamen, whereas perceptual tasks also activated the caudate nucleus and the globus pallidus.The basal ganglia and the pre-SMA/SMA have been mainly associated with beat processing (Grahn, 2009; Grahn & Brett, 2007, 2009; Schwartze, Keller, Patel, & Kotz, 2011; Teki et al., 2011). These structures show strong anatomical connections (Johansen-Berg et al. …