Expert and Nonexpert String Players’ Movement Durations as They Prepare to Play

Abstract

We analyzed the movement durations of 14 expert and 18 nonexpert string players as they prepared to play their instruments and in three other physical tasks unrelated to music making. We hypothesized that expert musicians would take more time to prepare their playing than nonexperts, but we found this not to be the case. There were no significant differences in movement duration means between experts and nonexperts in any of the four tasks including the Instrument Task. Surprisingly and somewhat inexplicably, we found that both expert and nonexpert musicians who participated in sports activities (n = 17) took significantly more time to prepare their playing than did the other participants. Further inspection of the video recordings revealed important differences in how experts’ and nonexperts’ movements unfolded in the Instrument Task. Nonexperts’ movements tended to be uneven and disjunct, whereas experts’ movements were fluid and even from the start, suggesting that experts’ conceptions of “starting a note” begin prior to the onset of movement and not when the bow is in close proximity to the string.

Music making involves a constellation of coordinated processes—perceptual, cognitive, motor, emotional—all directed toward the accomplishment of musical goals. “Music performance is both a natural human activity, present in all societies, and one of the most complex and demanding cognitive challenges that the human mind can undertake” (Zatorre et al., 2007, p. 547). In fact, active music making is thought to engage more components of the brain than any other human activity (Vuust et al., 2022).

Expert string playing is an exemplary model of skilled motor behavior. There are innumerable pedagogical prescriptions about the physical components of string playing, and across centuries of pedagogy there are remarkable consistencies about how to best position the instrument in relation to the body, how to shape the bow arm and fingers, and how to facilitate the movements of the left hand (e.g., Applebaum & Lindsay, 1986; Flesch, 2000; Galamian, 1962; Green, 2010; Mozart, 1985; Rolland & Mutschler, 2007).

These pedagogical prescriptions focus on the creation of beautiful, in-tune, expressive playing, often with less attention devoted to how musicians develop and regulate their movements to reach this level of expertise. A notable exception is the string pedagogy of Paul Rolland, who devotes considerable attention to the core principles of “total body action,” in which achieving balanced, free movement is the foundation for successfully acquiring string performance skills. Rolland emphasized that the movements preceding tone production should unfold with fluidity and in a manner consistent with the character of the music that follows (Rolland & Mutschler, 2007).

String players’ movements have been studied in a limited number of contexts with the following findings: string playing elicits significant changes in neural motor representations following many years of left-hand development (Elbert et al., 1995); experts’ bow position, force, and acceleration are remarkably consistent between multiple iterations of a given bowing task (Ancillao et al., 2017; Schoonderwaldt & Demoucron, 2009; Turner-Stokes & Reid, 1999); and a variety of motor approaches to string instruments, despite smoothness or replicability of said motions, are susceptible to musculoskeletal disorders brought on by overuse injuries from playing (Ackerman & Adams, 2004; Mizrahi, 2020; Rensing et al., 2018; Rickert et al., 2012; Yang et al., 2021). However, these investigations have all focused on the behavior of highly-skilled, expert string musicians and often in contexts removed from naturalistic musical goals. Nonexpert or developing string players receive no attention on these measures. In addition, none of these studies measure the anticipatory or preparatory movements as string players prepare to play their instruments.

Background

All motile organisms, including humans, move to accomplish goals (Balleine & Dickinson, 1998; Dickinson & Balleine, 1994). Neural activations in the central nervous system (CNS) prompt relevant effectors in pursuit of desired outcomes and are informed by the state of the environment and by previously learned associations (Wolpert & Ghahramani, 2000). Much of this processing occurs below conscious awareness and is immediately paired with an action’s temporally relevant sensory consequences. The binding together of these processes in the CNS creates a conscious impression of agency in action and the perception that accomplishing movement goals occurs in a fluent series of events (Haggard et al., 2002).

Voluntary movements are typically characterized as occurring in three phases: a motor preparation (or planning) phase, a “trigger phase” cooccurring with an imposed or naturally-present cue to begin, and a movement phase (Ames et al., 2019; Wise, 1985). Motor preparatory processes appear to facilitate the execution of skilled movements and are systematically organized in relation to the accomplishment of movement goals, such as encoding the direction or velocity of a reach (Gallivan et al., 2015; Tanji & Evarts, 1976) or producing more accurate attempts with faster reaction times (Afshar et al., 2011; Churchland et al., 2006; Gallivan et al., 2015; Tanji & Evarts, 1976). The organization of these preparatory processes occurs during the interval of time between an individual’s intent to achieve a given outcome and the presentation of the “go” cue just prior to the earliest muscle activations. During the preparation phase, an individual may assume a more formal state of readiness, like the ready position in this study, when awaiting a cue to begin the task.

Vyas and colleagues (2020) demonstrated a causal relationship between the motor preparation period and the trial-to-trial update processes that take place in motor learning. Significant deficits in trial-to-trial improvements occurred when pre-motor cortex processing was disrupted at the end of the motor preparation period (directly before the motor command is executed), which suggests an important interaction between the motor preparation period and movement onset that informs the subsequent update processes (Braquet et al., 2020; Brown et al., 2011). In other words, the time just prior to the presentation of a go signal involves internal planning processes that, if disrupted, diminish performance.

After a motor plan is generated, the sensorimotor system translates motor commands into actual movement. The CNS creates an internal model of the “state” of the system that contains time-relevant information regarding the environment, the organism, and the task. This model is sometimes referred to as an efference copy (Robinson, 1973), and it represents a prediction of expected sensory consequences, including continuous predictions regarding movement positions, velocities, and torques, as well as discrete information about the environment throughout the duration of the movement (Kawato, 1999; Wolpert & Ghahramani, 2000). These predictions allow for movement to unfold effectively when sensory feedback, which is inherently slow and noisy, is initially unavailable.

Athletes, like musicians, are tasked with developing movement-based expertise in dynamic environments in pursuit of goals (Braun Janzen et al., 2014). The types of goals certainly vary within and between different sports, and success in athletic performance is dependent upon an individual’s ability to efficiently assess their environment and prepare their body to execute the most beneficial action in moment-to-moment play. Coordination dynamics, or the study of how coordinated movements are learned and change in natural contexts, is prevalent in sports-movement literature that characterizes patterns of movement, such as identifying preparatory states or preparatory motions, that lead to generally positive outcomes (Balague et al., 2013; Jaitner et al., 2001; Jantzen et al., 2008).

Given the demand for precisely timed movements in both sports and music performance, comparisons between the movements of musicians and athletes have sparked an interest in cross-domain research. The rhythmic and expressive nature of music allows some individuals to better synchronize their movements with a regulated pulse during exercise or other sports routines (Chiat & Ying, 2013; Karageorghis et al., 2010; Karageorghis & Priest, 2012) not unlike the synchronizations required of musicians to a steady beat and when performing with other musicians (Large, 2000; Repp, 2005). Bianco and colleagues (2017) shared an interesting finding in which drummers and athletes demonstrated similar margins of improvement in a cognitive task when compared with their nondrummer, nonathlete counterparts. Drummers and athletes showed enhanced motor preparation measures and attentional control, leading to faster reaction times. The reasons for these connections across sports and music, however, remain unclear and require further investigation.

String instrument playing, in particular, perhaps best illustrates the motor processes involved in music making because, unlike wind-instrument playing and singing, nearly all of the physical components of string instrument sound production are visible to observers. Teachers of string instruments are advantaged in that they can monitor most of the relevant components of tone production moment-to-moment throughout the course of instruction.

An important moment in performance on every instrument, including the voice, is the performer’s preparation to begin a tone. This includes the conscious and unconscious organization of preparatory movements that lead to producing sounds that fulfill the musical goals of the performer. The ways that string players prepare to play consequentially affect the transition from silence to sound and “set up” the movements that follow as music commences. We designed a study to examine this component of string playing. After a comprehensive search, we found no published research to date that describes string musicians’ motor behavior in terms of early movement timing as a function of expertise.

In our own informal observations of instrumental musicians, movements associated with the initiation of playing are sometimes punctuated by brief moments of quiet or stillness that seem to affect the success of the musical attempt. For example, young string musicians who set their bow on the string, pause, then play, seem to achieve successful outcomes more frequently than do other young musicians who begin playing abruptly with no pause. The timing of these moments of stillness and when they occur seem to vary with age and the experience of the performer.

One way to investigate how individual musicians of varying expertise regulate their movements is to measure the time it takes to prepare the production of a target pitch; that is, the time between a signal to move and the initiation of a target pitch. The timing and kinematics in this interval of a movement sequence seem particularly relevant to understanding how musicians organize and regulate their movements in pursuit of musical goals. To test movement duration in preparing to play a string instrument systematically, we designed a study in which nonexpert and expert musicians performed one task with their instrument and, for comparison, three nonmusical motor tasks. We included the nonmusical tasks to test the extent to which musicians’ approaches to movement goals generalize to tasks outside the domain of music.

In Figure 1, the dashed line bracket depicts the part of each participant’s movement sequence that we examined. Movement duration for each task iteration was measured in milliseconds (ms) starting with the proctor’s cue, “You can start [timing of the first consonant of the word ‘start’ defined the beginning of each measured interval],” to the completion of the task. The measured duration includes the reaction time after the presentation of the cue to start in addition to the movement time needed to complete the task.

Figure 1.

Diagram of Movement Sequences in Four Tasks.

Evaluating movement duration may reveal important differences between expert and nonexpert string players in their self-paced movements from a cue to the completion of a musical goal. This study is intended to provide an entry point in these types of analyses. Specifically, we asked whether experts’ movement timing differed from nonexperts’ between a signal to begin playing and the start of a note.

Method

We recruited 32 expert and nonexpert string players whose primary instruments were violin (n = 11), viola (n = 7), cello (n = 7), double bass (n = 5), and guitar (n = 2). The 14 expert participants (seven female) were 24 to 48 years of age (M = 35.4, SD = 7.5) with between 12 and 45 years of playing experience on their primary instruments (M = 26.3, SD = 9.9), and were recruited through the Butler School of Music at The University of Texas at Austin and through the local pool of professional string performers and teachers. The 18 nonexpert participants (13 female) were 6 to 13 years of age (M = 9.2, SD = 1.9) with 1 to 7 years of experience on their primary instruments (M = 3.4, SD = 1.8), and were recruited through The University of Texas at Austin String Project. Nonexpert participants had studied their primary instruments in private and group instruction; the oldest of these participants was in middle school.

All participants completed a demographic survey from which we collected information about participants’ age, gender, primary instrument, years of experience with primary instrument, and level of sports involvement. All participants provided informed consent, as did the parents of participants under the age of 18, who provided assent. Participants received no compensation for their participation. All procedures and recruitment materials were approved by the Institutional Review Board of The University of Texas at Austin.

All data were collected during the COVID pandemic, and we arranged for participants to meet over Zoom video conferencing software. The first author recorded and proctored these video sessions with proper permission from all participants. Each meeting lasted approximately 10 to 15 minutes and took place during the day at the participant’s convenience. The proctor read from a prepared script to ensure consistency across sessions (see Appendix). Eleven volunteers initially piloted the script to test for clarity in the written instructions. The test proctor read the following instructions before participants began each of the following tasks:

1. Fingertip Task: Start with your hands about a foot apart with your palms facing each other like this [proctor demonstrated the ready position]. This is the ready position. When I say you can start, move your hands together so that all ten of your fingertips touch at exactly the same time;

2. Pencil Task: Turn your body sideways in your chair and extend your nondominant arm, making a thumbs-up fist with your hand. Pick up the pencil in your dominant hand, holding it at the end opposite the point. Hold the pencil up next to your ear on the side of your dominant hand with the point pointing toward your nondominant thumbnail like this [proctor demonstrated the ready position]. This is the ready position. When I say you can start, move the pencil toward your nondominant thumb, touching the tip of the pencil to the exact middle of your thumbnail;

3. Toss Task: Position yourself so that the camera of your device is about an arm’s length away from you. For now, please leave your hands on the table or in your lap while I describe the task. Hold the crumpled ball of paper in your dominant hand as if you are ready to throw it like this [proctor demonstrated the ready position]. This is the ready position. When I say you can start, toss the wad of paper at your device’s camera. Your goal is to hit the exact center of the camera lens with the wad of paper; and

4. Instrument Task: Position yourself and the camera so that I can see both your left and right hand on your instrument when you are in your usual playing position:

Violin/Viola/Cello/Bass. Begin with your instrument in playing position [shoulder for violin/viola] and your bow arm down at your side [resting on the right knee for cellos]. This is the ready position. When I say you can start, play a beautiful down-bow sustained E (first finger) on the D string starting at the frog. Your goal is to play the most beautiful note you possibly can.

Guitar. Begin with your instrument in playing position, and your right hand resting on the top of the lower bout of the instrument. This is the ready position. When I say you can start, play a beautifully plucked D followed by a plucked second fret E on the D string. Your goal is to play the most beautiful notes you possibly can.

The nonmusical tasks were generated as examples of other actions participants could demonstrate with clear observable goals. Both the Fingertip Task and the Pencil Task are demonstrations of bimanual controlled movements whereas the Toss Task is more representative of a ballistic movement in which the completion of the goal is not evaluated until after the movement has already been completed. The purpose of including these three tasks in conjunction with the Instrument Task was to examine possible relationships between movement timing in a familiar musical task and in tasks that are unrelated to participants’ musical skills.

All participants performed the four tasks in the order listed above. We wanted the instrument task to be the last task performed, and we thought the likelihood of an order effect was very small, given the varied nature of the four tasks. Our primary interest was in examining possible differences between experts and nonexperts and not among tasks.

The test proctor modeled the ready position for each task, and instructed the participants to “take all the time you want” to perform each task. Participants were asked to repeat the instructions to the test proctor prior to performing each task to check their understanding. Each task was performed twice, returning to the ready position prior to the second task iteration, and initiated by the proctor’s verbal cue, “You can start.” The proctor noticed no apparent delays in timing due to lag or connection issues across all participant sessions in this study.

Participant videos were analyzed with Final Cut Pro X video editing software to obtain an accurate measure of movement duration for each participant’s task performances. We defined movement duration (in ms) as the time elapsed between the proctor’s cue, “You can start [timing of the first consonant of the word ‘start’ defined the beginning of each measured interval],” and the completion of the task (the touching of the fingertips, the contact of the pencil with the fingernail, the release of the paper ball, and the start of the note).

Results

We first considered the consistency in participants’ movement durations between their first and second performances of each task. Across all participants, correlations between the two performances were statistically significant: Fingertip Task, r = .88, p < .001; Pencil Task, r = .92, p < .001; Toss Task, r = .50, p < .004; Instrument Task, r = .73, p < .001. In light of this result, we calculated a mean value for each task for each participant and used these values in all subsequent analyses. The mean movement durations by task and expertise participant group are presented in Figure 2.

Figure 2.

Mean Movement Durations in the Four Tasks for Expert and Novice Participants.

We compared the mean movement durations for each task using a two-way (Group × Task) repeated-measures analysis of variance (ANOVA). ¹ We found no significant interaction between Group and Task, F(2.0, 60.2) < 1, p = .40; no significant effect of Group, F(1, 30) < 1, p = .92; and a significant effect of Task, F(2.0, 60.2) = 40.12, p < .001, η² = .42. Post hoc analyses using Tukey’s HSD test indicated that the mean for the Fingertip Task was not significantly different from the means for the Pencil Task or the Instrument Task. All other pairwise comparisons were statistically significant, p < .01.

In addition, we considered how the movement durations in the Instrument Task related to the other movement tasks in the study. Across all participants, a series of bivariate comparisons revealed that the timing of participant movements for the Instrument Task were not significantly correlated to the Fingertip Task, p = .36; the Pencil Task, p = .54; or the Toss Task, p = .12.

We found no meaningful differences attributable to any of the demographic variables with the exception of sports participation. We grouped participants by those who indicated they regularly participated in sport activities (n =17) and those who did not (n = 15) regardless of years of experience on their instrument. These data are presented in Figure 3. Reported sports activities included ballet, tennis, soccer, running, biking, hiking, and weightlifting. We then compared these groups in each of the four tasks with a series of independent-samples t-tests. Sport and nonsport groups differed significantly in movement duration in the Instrument Task, p < .001, η² = .32, but were not significantly different on the Fingertip Task, the Pencil Task, and the Toss Task, p >. 51. This finding suggests that, across years of experience, participants who regularly engage in sports activities take significantly longer (M = 4,981 ms) to prepare their playing than do participants who do not engage in sports (M = 3,716 ms). ²

Figure 3.

Mean Instrument Task Durations for Sport Participation Groups.

The Instrument Task was of primary interest in the current investigation, and we had expected to find differences in movement timing between experts and nonexperts. The absence of systematic differences in this variable led us to inspect the video recordings of participants’ movements. Although expert and nonexpert participants did not differ significantly in movement duration for the Instrument Task, we observed that there were distinct features that characterized experts’ preparatory movements that were not typical among the nonexpert participants. As a way of organizing our observations, we matched video recordings of one expert and one nonexpert whose Instrument Task timings were within 100 ms of one another, resulting in 11 unique expert–nonexpert pairs. We edited the videos so that the two recordings in each pair could be viewed simultaneously (side by side) on a video monitor.

We observed important differences in how experts’ and nonexperts’ preparatory movements unfolded. Typical expert movements were characterized by a brief pause before any motion began, followed by fluid motion and a distinct “flourish” just prior to the start of the sound. Importantly, there was nearly no pause from the moment experts set their bow on the string to when the sound of their note began. We interpreted this as an indication that, for experts, “starting a note” begins with the first movements raising the bow toward the string. ³

In contrast, nonexpert participants’ movements tended to be more disjunct and unpredictable, with no consistent patterns in movement from participant to participant. Nonexperts, compared with the experts, often contacted the string with their bow, then adjusted their setup, and then began the note. Put another way, the start of the note for nonexperts appeared a separate task from getting the bow in position to play, whereas experts seemed to move as though the note began at the first movement of the right hand and bow.

Discussion

To examine string musicians’ preparatory motor behavior in relation to expertise, we measured movement durations as musicians prepared to initiate a tone on their instruments. We also analyzed the movement durations of the same participants as they completed three other physical tasks unrelated to music making. We anticipated differences between experts and nonexperts regarding their movement durations for the Instrument Task based on our own informal observations of nonexpert and expert string players.

We found no significant differences in movement duration between expert and nonexpert groups on any of the four tasks. Participants’ age, gender, and primary instrument were unrelated to movement timing, but we found significant differences between those who participated in sports activities and those who did not with regard to the Instrument Task. Possible reasons for this effect are unclear. There may be cross-domain motor schemas that lead to longer preparation times for participants involved in multiple physically driven activities with complex motor demands, but there is as yet no evidence to support this conjecture.

Given that we observed no differences in movement duration in the music task between experts and nonexperts, we looked more closely at the participants’ movements in the video recordings and observed notable differences in how expert string players actually move when preparing to play that contradicted our previous assumptions. Experts’ movements in the Instrument Task appear to anticipate the initiation of sound and are fluid throughout the entire preparation of the note, often with a second distinct movement “flourish” just prior to the setting of their bow on the string. It may be that experts conceive that the “start” of a note begins with the start of their movement rather than the initiation of sound. This finding aligns with the principles of “total body action” that serves as the basis for the Paul Rolland methodology, in which he instructs string players that “the preparatory movement should have the same character as the music which follows” (Rolland & Mutschler, 2007. p. 41). This finding is also consistent with the suggestion that experts initially approach tasks in terms of the broader principles required by the task (e.g., the character of the piece or an emotion to communicate) and work on finer details later as needed (Chaffin et al., 2003). Finally, there was no delay between the moment that experts’ contacted the string and the start of the note. Resting the bow on the string for a prolonged period, as is often recommended to young string players, would likely interrupt the flow of experts’ movements and musical thinking.

Further investigation is needed into how these movement patterns develop between the first years of instruction and the point at which string players develop expertise. It is certainly possible that imposing a stillness just prior to pulling the bow helps some young string players in organizing extraneous movements when accomplishing goals in their playing, but this may occur at the risk of delaying movement patterns that are a part of expert playing. It may be that nonexperts would benefit from early instruction that directs them to prepare their playing with balanced fluid movements and with an emphasis on synchronicity of movements with musical intent. Teachers who conceive of students’ preparatory movements and gestures in ensembles and private lessons as integral components of sound production can shape students’ playing to more closely resemble the playing of experts. Further research in determining transferable approximations to lead students to more advantageous preparation is needed.

This study is not without limitations. All participants performed the four study tasks in the same order, and it is possible that the nonmusical tasks, despite being outside of the musical domain of the Instrument Task, could have biased the timing of participants getting ready to play. Also, given the number of participants in our sample and the difference in average age between the expert and nonexpert participant groups, generalization of our findings will require further research. In addition, this study was conducted over Zoom out of necessity to observe safety protocols when collecting data during the Covid-19 pandemic. Conducting this type of motor analysis over video conferencing software may have introduced slight timing variations due to Wi-Fi connection speeds that we could not control for.

The significant difference between sport participation groups on the Instrument Task prompted many questions. Across expertise, gender, and age, those who reported participating in sports regularly took over 1 second longer, on average, to prepare to play than did those who did not regularly participate in sports. String playing involves the coordination of multiple effectors moving in ways that include both fine adjustments (e.g., finger flexions) and larger spans of movement (e.g., bow arm control) to accomplish goals. Sports like tennis and dance (sports that several of the participants reported involvement in) also involve these fine and large motor adjustments and motor plans to achieve goals in practice and in performance.

This study is the first investigation to examine movement timing among string players of varied levels of experience and expertise. That the overall mean movement durations were not different between experts and nonexperts was somewhat surprising, but the visual inspection of the movement dynamics alerted us to important differences in the preparatory movements of experts and nonexperts. It would seem beneficial to further document the motor behavior of skilled performers in an effort to better prepare aspiring young musicians.