Introduction

The ability to keep a beat lies at the heart of music-making. A typically-developing child can maintain a steady beat as young as four years old[1], and even infants can “feel the beat”, demonstrating sensitivity to metrical structure well before they are able to synchronize to it themselves[2,3]. Though the emergence and mastery of rhythm skills can continue through to adulthood, rhythmic competence can depend greatly on the extent of engagement with music [4-7]. While a child who struggles to speak his or her native language may be assessed for a language deficit, individuals who fail to develop basic musical skills (even those engaged in musical training) are often simply considered “unmusical”. However, performance on simple tapping tasks relates with a variety of critical abilities outside the musical domain, from cognitive and linguistic skills[8,9] to handwriting[10]. Moreover, abnormal rhythmic performance is associated with language deficits such as dyslexia[11-13] as well as attention-related deficits[14]. The potential for music classes to strengthen fundamental temporal processing mechanisms therefore holds great promise for educators and clinicians.

Despite significant advances in the understanding of sensorimotor synchronization skills across the lifespan, including the effects of musical expertise[15,16], no studies to our knowledge have investigated the relationship between classroom music instruction and beat-keeping abilities in young children, at an age when both rhythm and literacy skills are emerging. While the ability to attend to rhythmic patterns and maintain a steady beat increases between the ages of 5 and 7 [1,5], it has been suggested that rhythm abilities reach adultlike “nonmusician” levels by ages 7 to 8, and may not develop further without additional training [5]. Given the evidence for links between rhythm skills and reading ability, it is of great interest to educators and clinicians to determine whether musical training can support the development of reading skills by strengthening underlying temporal processing mechanisms.

The present study was carried out in collaboration with Harmony Project, a non-profit organization providing free music education to underserved children in gang-reduction zones of Los Angeles. We used continuation and paced tapping paradigms to assess beat-keeping skills in a group of elementary school children who had received one year of musical training, compared with a matched group of untrained controls. We hypothesized that musical training strengthens basic timekeeping skills, resulting in greater accuracy in keeping a beat in the musically-trained children.

Methods

Results

A series of repeated measures ANOVAs were run on each continuation and paced tapping measure to determine whether the rhythmic abilities of the musically-trained children differed from those of the control children. Tempo (2 Hz versus 1.5 Hz) was included as the within-subjects variable and group (musically trained versus control) was included as the between-subjects variable. For unsigned accuracy in the continuation condition there was a main effect of group (F(1,56) = 6.45, p = 0.0139): musicians performed more accurately (median composite error = 3.14%) compared to nonmusicians (median composite error = 4.81%). (Shapiro-Wilk’s test for non-normality revealed that all data points except for lag 1 autocorrelation were non-normally distributed at p < 0.05; as a result we report medians and use Wilcoxon rank sum tests rather than t-tests for post-hoc analyses.) (Nonmusician performance, at 4.81% composite error, was less than the 10.6% error reported for nine-year-old typically developing children by Corriveau and Goswami [11]. The worse performance reported by Corriveau and Goswami may be due to the fact that they used a longer continuation interval, giving subjects more time to drift away from the beat.) There was also a main effect of group for continuation signed accuracy (F(1,56) = 7.08, p = 0.0102): musicians sped up less relative to the target tempo (median composite error = 1.74%) compared with nonmusicians (median composite error = 4.81%). There were no main effects for paced unsigned accuracy, paced signed accuracy, continuation variability, paced variability, paced lag 1 autocorrelation, and continuation lag 1 autocorrelation. There was an interaction between group and condition for paced variability (F(1,56) = 5.60, p = 0.0214), but post-hoc Wilcoxon rank sum tests revealed no significant group differences (p > 0.05).

To determine whether the enhanced continuation tapping performance in the trained children was due to increased encoding ability or better resilience to memory deterioration, tapping data were divided into first and second halves and analyzed separately. A series of repeated measures ANOVAs were run for each tapping measure, comparing composite performance (averaged across tapping rates) in the first versus the second half as the within-subjects variable and group as the between-subjects variable. For continuation unsigned accuracy, there was a main effect of time (F(1,56) = 4.05, p = 0.0491), a main effect of group (F(1,56) = 5.73, p = 0.0201), and an interaction between time and group (F(1,56) = 5.50, p = 0.0226). Post-hoc Wilcoxon rank sum tests revealed no significant group difference in the first half of the tapping data (p > 0.1). There was, however, a significant group difference in the second half of the tapping data (p = 0.0205, zval = -2.32, ranksum = 706), such that trained children (median error = 3.14%) were more accurate than controls (5.34%) (Figure 1). See Figure 2 for a histogram showing the distribution of unsigned accuracy scores for the two groups. For continuation signed accuracy, there was a main effect of group (F(1,56) = 7.02, p = 0.0105), such that trained children were more accurate than controls, but there was no main effect of time. Finally, for paced variability, there was a main effect of time (F(1,56) = 9.03, p = 0.0040), such that performance in the first half of the test was more variable (median variability 3.95%) than performance in the second half of the test (median variability 5.02%).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 1. Comparison of musically-trained children (red) and untrained controls (black) showing no group difference in the first half of the continuation tapping period (Wilcoxon rank sum test: p > 0.1) but significantly reduced error in the trained group in the second half (p = 0.0205, zval = -2.32, ranksum = 706).

A smaller mean error reflects better ability to keep the beat. (Error bars indicate +/-1SE of the mean.)

https://doi.org/10.1371/journal.pone.0077250.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 2. Histograms showing the distribution of tapping accuracy scores in the musically trained and untrained control groups.

https://doi.org/10.1371/journal.pone.0077250.g002

Discussion

We herein assess basic timekeeping abilities in elementary school children using a finger-tapping paradigm. Superior tapping skills have been demonstrated in trained adult musicians [16], but there has been little work to assess the impact of music training on the development of these skills during childhood years. The present findings add to previous work by demonstrating that children who received one year of classroom-based music instruction are significantly more accurate in maintaining a steady beat than their untrained peers.

Both the trained and untrained groups showed, on average, a negative signed error, meaning that after the pacing stimulus was removed both groups tended to increase the speed of their tapping relative to the target tempo. The trained group, however, sped up to a lesser extent than did the untrained group. The mechanism underlying this increase in tempo is unclear, but it may be related to the tendency for subjects to anticipate the beat when tapping to a pacing stimulus, resulting in a negative asynchrony. Musically-trained individuals show smaller mean negative asynchrony than untrained individuals[19], suggesting that in the general population there is a tendency to overestimate the amount of time that has elapsed since the last beat, and that musical training lessens this tendency. Alternatively, musical training could lessen the tendency to speed up by slowing the preferred rate, as musical training has been linked to slower spontaneous tapping rates[20].

The continuation tapping paradigm used in this study requires that the beat be maintained in memory after the pacing stimuli have stopped, and the greater difference in accuracy between the trained and control groups during the second half of the continuation period suggests that memory may indeed be a factor in how training influences beat-keeping abilities. The initial encoding of the beat period, on the other hand, may be less affected by musical training, as the trained and untrained group performance did not differ during the first half of the continuation period, and synchronization accuracy did not differ in the paced condition, in which the stimulus was present throughout the experiment. Previous research indicates that auditory working memory may also be strengthened by musical training in young children[21], but given the multi-faceted nature of memory it is unknown whether these same capacities underlie any training-related improvement in the ability to maintain a beat over time.

In both the paced and continuation conditions we found no relationship between training and tapping variability, a measure that has previously been associated with musical experience [22,23]. Our participants, however, had much less musical experience (one year) than participants in prior studies examining synchronization ability and musical experience, which have tested adult professional or semiprofessional musicians. Decreased variability, then, may require either more extensive training or training that is more focused on instrumental performance (in the present study, much of the subjects’ training was dedicated to fostering musicianship skills; participants only began instrumental training after several months of instruction.) On the other hand, the ability to hold a tempo in memory is enhanced after only a year of training that focused in large part on listening skills.

Locking onto temporal patterns in a complex acoustic signal helps the brain to anticipate when meaningful information is likely to occur and streamline allocation of neural resources accordingly, as well as facilitating coordination and planning of associated motor actions. Oscillatory activity may play a role in the dynamic modulation of attention over time[24,25] and in facilitating communication between brain areas such as auditory and motor cortices[26]. Recent research suggests that oscillatory activity associated with cognitive functions such as attention and memory is greater in adult musicians than non-musicians, and that musical training during childhood promotes the development of these underlying oscillatory mechanisms[27] Similarly, musical training may strengthen the mechanisms underlying the brain’s “internal timekeeper,” which is essential to the accurate perception and production of temporal patterns[28].

The ability to perceive and predict temporal patterns is also critical to social interaction and communication[29] and there is evidence for some overlap in the neural pathways involved in processing temporal structure in both music and language[30-32], with neural oscillatory activity playing an important role in both rhythm perception and speech processing[33-38]. It is proposed that language difficulties such as dyslexia may result from deficiencies in underlying oscillatory mechanisms and that these deficiencies also result in impaired rhythm perception, which may help to explain the observed relationships between rhythm abilities and reading[9,11-13,39,40].

There is evidence to suggest that the perception of meter is malleable with experience, with infants as young as 12 months old demonstrating culture-specific patterns of response to musical rhythms, whereas at 6 months old no cultural specialization is evident[41,42]. Research also indicates that the greater awareness of metrical structure developed through musical training can influence how sequential motor actions are planned and executed, providing a hierarchical framework which facilitates coordination of motor movements over longer timespans[43]. It is possible that one mechanism by which musical training may improve tapping abilities is by providing or reinforcing hierarchical frameworks within which an ongoing beat can more easily be conceived and maintained. Musical activities, particularly ensemble playing, also provide experience in the simultaneous planning and execution of sequential movements over time, which is necessary to maintain a steady beat.

While the musicianship training provided to the trained group (curriculum is outlined in Table 2) includes tasks relating to rhythm such as call and response rhythmic patterns, these involve hand clapping and have no direct similarities with the tapping measures used in our assessments as far as the equipment (tapping pad) or specific instructions involved. It therefore seems likely that any transfer of ability from the musical training to the finger-tapping task involved generalization of underlying beat perception and beat generation abilities to a new context, rather than direct familiarity with the task per se. It is possible that the children engaged in musical training have developed a greater interest in music-related activities and therefore were more motivated to perform well on a rhythm-related task. However, the fact that the trained and control groups do not differ in tapping variability also suggests that the training effect is not due to a general familiarity or motivation-related effect but is instead the result of a specific enhancement in the ability to maintain a beat after the pacing stimulus is removed.

Our findings contribute to the growing body of research showing selective enhancements in both neural and cognitive function resulting from short-term instrumental music training[27,44-46]. There is significant interest in the role of music education in the development of literacy skills[47-51], and the ability to shape rhythm-related abilities with short-term training is of particular relevance given increasing evidence of links between rhythmic abilities and reading skills[9,11-13].

There is much work still to be done in understanding the mechanisms by which training can strengthen basic temporal processing abilities during critical developmental years, and how this, in turn, may influence emerging literacy skills. Previous work from our lab has demonstrated that even a few years of musical training in childhood can influence the neural processing of sound later in life, years after training has ceased[52], and further research is needed to determine whether the benefits of early musical training for temporal processing are maintained into adulthood.

Conclusions

Understanding how classroom music instruction can shape underlying temporal processing abilities has broad educational implications, given converging evidence for the importance of temporal processing for language and literacy skills[9,11-13,53,54]. Our findings indicate that wider benefits of music training can be made accessible to at-risk populations through programs such as Harmony Project, which not only support academic and personal development, but also play a unique and invaluable role in community-building and social cohesion.

Acknowledgments

The authors wish to thank the families who participated in this study, and the Harmony Project staff for their support. We also wish to thank Dana Strait, Samantha O’Connell and Shivani Bhatia for significant contributions to project coordination and testing, as well as Trent Nicol, Samira Anderson, Erika Skoe and Karen Chan who provided comments on an earlier version of this manuscript.