Predicting the biological variability of environmental rhythms: Weak or strong anticipation for sensorimotor synchronization?

Highlights

• Sensorimotor synchronization has generally been studied using regular metronomes.

• Yet natural environmental rhythms exhibit typical fractal variability.

• We model synchronization processes with various types of (fractal) variable rhythms.

• Processes are same for all variable rhythms but different than for regular rhythms.

• The generalizability of results from regular-metronome paradigms is questioned.

Abstract

The internal processes involved in synchronizing our movements with environmental stimuli have traditionally been addressed using regular metronomic sequences. Regarding real-life environments, however, biological rhythms are known to have intrinsic variability, ubiquitously characterized as fractal long-range correlations. In our research we thus investigate to what extent the synchronization processes drawn from regular metronome paradigms can be generalized to other (biologically) variable rhythms. Participants performed synchronized finger tapping under five conditions of long-range and/or short-range correlated, randomly variable, and regular auditory sequences. Combining experimental data analysis and numerical simulation, we found that synchronizing with biologically variable rhythms involves the same internal processes as with other variable rhythms (whether totally random or comprising lawful regularities), but different from those involved with a regular metronome. This challenges both the generalizability of conclusions drawn from regular-metronome paradigms, and recent research assuming that biologically variable rhythms may trigger specific strong anticipatory processes to achieve synchronization.

Introduction

Synchronizing our movements with environmental rhythms such as a partner’s walking cadence is part of our everyday sensorimotor behavior. Among the most striking examples may be our spontaneous tendency to move on the beat of music, observed since the earliest months of life (Zentner & Eerola, 2010). The sensorimotor synchronization paradigm has traditionally investigated the neuropsychological processes involved in synchronizing simple rhythmic movements, such as finger tapping, with basically regular (isochronous) auditory metronomes (Repp, 2005). Despite notable pushes towards studying synchronization with non-isochronous sequences including musical pieces (e.g., Rankin et al., 2009, Repp, 2002), local tempo changes (e.g., Schulze, Cordes, & Vorberg, 2005), and virtual adaptive partner’s movements (e.g., Repp & Keller, 2008), the main focus on regular metronomic sequences however contrasts with the typically variable rhythms contained in environmental stimuli. Especially, it is widely acknowledged that rhythms generated by natural bio-behavioral systems exhibit a characteristic structure of fluctuations over time, namely long-range correlation, or fractal fluctuations (e.g., Bullmore et al., 2009, Gilden, 2009, Kello et al., 2010, Markowitz et al., 2013, West and Schlesinger, 1989). Research has shown that the human perceptual-motor system is sensitive to the temporal structure of biologically variable stimuli (Wark, Lundstrom, & Fairhall, 2007), and synchronizing with fractal signals has recently been assumed to involve a very specific synchronization process, namely Strong Anticipation (Marmelat and Delignières, 2012, Stephen et al., 2008, Stepp and Turvey, 2010).

The strong anticipation hypothesis has been formulated in contrast to more classical conceptions of weak anticipatory processes (Dubois, 2003), involving internal models for short-term prediction and correction of the periods and/or asynchronies produced to achieve synchronization (e.g., Madison and Delignières, 2009, Mates, 1994a, Mates, 1994b, Semjen et al., 2000, Thaut et al., 1998, Vorberg and Schulze, 2002, Vorberg and Wing, 1996; for a review see Repp, 2005, Repp and Su, 2013). In contrast to such local adaptations, strong anticipation is assumed to involve a global tuning of behavior to the statistical properties of environmental fluctuations. This attunement relies on the existence of lawful regularities in environmental fluctuations, hence the role of long-range correlations in the stimuli presented. In this way, the temporal structure of environmental stimuli determines the structure of behavior, leading to the matching of environmental and behavioral fluctuation structures (Stephen et al., 2008).

The Strong Anticipation framework thus questions the generalizability of results from classical synchronization paradigms to ecological, i.e. fractal – or long-range correlated – conditions. In particular it raises the following issues: First, are synchronization processes to be differentiated between long-range correlated stimuli and other variable/regular stimuli, or between any kind of variable (including long-range correlated) stimuli and regular stimuli? Second, to what extent can alternative models, based on local short-term adaptations, account for the matching of fluctuation structures attributed to strong anticipation? These issues have potentially important implications for optimizing human/artificial-environment interaction, like the use of rhythmic auditory stimulation for gait rehabilitation in Parkinson’s patients (e.g., Hove, Suzuki, Uchitomi, Orimo, & Miyake, 2012).

To answer these issues we used a synchronized finger tapping paradigm under five different conditions of variable (long-range and/or short-range correlated, uncorrelated) and regular metronomes. In view of the above mentioned elements we formulated the following working hypotheses:

• Strong anticipation should only be involved in conditions where inter-stimuli intervals comprise long-range correlations, and be unaffected by the presence of other forms of lawful regularities.

• The matching of behavioral and environmental structures of fluctuations should be observed only in conditions where inter-stimuli intervals comprise long-range correlations.

• The structure matching should not be accounted for by simple models implementing local short-term synchronization processes.