The development of expertise in music is complex and multifactorial, influenced by both genes and the environment (Ullén et al., Reference Ullén, Hambrick and Mosing2016). The salience of practice in the development of musical expertise is controversial. While the deliberate practice theory claims that the accumulation of hours of practice is the sole determinant of expert performance (Ericsson et al., Reference Ericsson, Krampe and Tesch-Roemer1993), others have shown practice alone to be insufficient to explain individual differences in ability (Hambrick et al., Reference Hambrick, Oswald, Altmann, Meinz, Gobet and Campitelli2014; Hambrick et al., Reference Hambrick, Macnamara, Campitelli, Ullén and Mosing2016; Kragness et al., Reference Kragness, Swaminathan, Cirelli and Schellenberg2020; Macnamara et al., Reference Macnamara, Hambrick and Oswald2014; Macnamara & Maitra, Reference Macnamara and Maitra2019; Mosing, Madison et al., Reference Mosing, Madison, Pedersen, Kuja-Halkola and Ullén2014). Studies highlight the importance of psychological factors such as personality traits (Butkovic et al., Reference Butkovic, Ullén and Mosing2015; Corrigall et al., Reference Corrigall, Schellenberg and Misura2013; Corrigall & Schellenberg, Reference Corrigall and Schellenberg2015) and cognitive ability (Corrigall et al., Reference Corrigall, Schellenberg and Misura2013; Lynn et al., Reference Lynn, Graham Wilson and Gault1989; Meinz & Hambrick, Reference Meinz and Hambrick2010; Mosing et al., Reference Mosing, Hambrick and Ullén2019; Mosing, Pedersen et al., Reference Mosing, Pedersen, Madison and Ullén2014; Schellenberg & Weiss, Reference Schellenberg, Weiss and Deutsch2013; Sergeant & Vhatcher, Reference Sergeant and Vhatcher1974; Swaminathan & Schellenberg, Reference Swaminathan and Schellenberg2018; Swaminathan & Schellenberg, Reference Swaminathan and Schellenberg2020), as well as the role of genetic factors (Hambrick & Tucker-Drob, Reference Hambrick and Tucker-Drob2015; Kragness et al., Reference Kragness, Swaminathan, Cirelli and Schellenberg2020; Mosing, Madison et al., Reference Mosing, Madison, Pedersen, Kuja-Halkola and Ullén2014; Ullén et al., Reference Ullén, Hambrick and Mosing2016; Ullén et al., Reference Ullén, Mosing, Holm, Eriksson and Madison2014) in skill acquisition.
One aspect of music ability in which the role of practice is still an open question is pitch naming. Absolute pitch (AP) is the uncommon ability to name isolated musical pitches without the use of a reference tone. AP has largely been regarded separately to other questions around the development of expertise, as it is relatively rare and appears to be less reliant on deliberate practice. AP possessors often report ‘discovering’ their ability rather than acquiring it through practice (West Marvin et al., Reference West Marvin, VanderStel and Siu2020), and a heritable component to the skill is likely (see Tan et al., Reference Tan, McPherson, Peretz, Berkovic and Wilson2014, for a review). While practice-related factors such as pedagogical methods with a focus on pitch naming have been shown to be positively associated with AP, AP itself is not explicitly taught (Gregersen et al., Reference Gregersen, Kowalsky, Kohn and Marvin2001; Miyazaki et al., Reference Miyazaki, Rakowski, Makomaska, Jiang, Tsuzaki, Oxenham, Ellis and Lipscomb2018; Wilson et al., Reference Wilson, Lusher, Martin, Rayner and McLachlan2012).
Although AP is present in relatively few musicians, the ability to name some pitches is more widespread. A relatively understudied phenomenon, variously known as partial- or quasi-absolute pitch (QAP), refers to those who perform at above chance level in pitch-naming tasks, but below the near-ceiling performance generally accepted as the standard to qualify as possessing genuine AP (Miyazaki, Reference Miyazaki1988, Reference Miyazaki1989; Wilson et al., Reference Wilson, Lusher, Martin, Rayner and McLachlan2012; Wilson et al., Reference Wilson, Lusher, Wan, Dudgeon and Reutens2009). This is in contrast to relative pitch (RP), the ability to use relationships between pitches to label musical tones. RP is the typical form of pitch identification used in music, and people with RP alone perform at chance level in pitch-naming tasks. The prevalence of QAP is not known, and the factors predicting its development remain under investigation. Preliminary studies suggest that QAP is related to AP insofar as the same factors predict its acquisition, although to a lesser extent (Wilson et al., Reference Wilson, Lusher, Martin, Rayner and McLachlan2012). If AP is the extreme, the existence of QAP suggests that there is a broader continuum of pitch-naming ability; however, more research is needed to determine if there are distinct phenotypes.
One factor that consistently emerges in pitch-naming research is the notion of a critical or sensitive period for acquisition (Baharloo et al., Reference Baharloo, Johnston, Service, Gitschier and Freimer1998; Deutsch et al., Reference Deutsch, Dooley, Henthorn and Head2009; Levitin & Zatorre, Reference Levitin and Zatorre2003; Vitouch, Reference Vitouch2003; Wilson et al., Reference Wilson, Lusher, Martin, Rayner and McLachlan2012). Analogous to the concept in language development, it refers to a developmental window in which children who may be predisposed to AP should be exposed to musical training to maximize the probability that they will acquire AP. The majority of AP possessors begin musical activities in early childhood, between the ages of 3 and 7, and the proportion of AP possessors who start at an older age decreases steadily, consistent with a gamma distribution (Levitin & Zatorre, Reference Levitin and Zatorre2003). Young age of onset of musical training is also associated with QAP ability (Wilson et al., Reference Wilson, Lusher, Martin, Rayner and McLachlan2012).
It has also been suggested that current musical practice is important in pitch-naming ability. Some AP musicians can ‘shift’ their pitch templates through practice on transposing instruments (i.e. instruments for which written notation differs from the produced pitch; West Marvin et al., Reference West Marvin, VanderStel and Siu2020), and engagement in musical practice has been associated with maintenance of pitch-naming ability (Dohn et al., Reference Dohn, Garza-Villarreal, Ribe, Wallentin and Vuust2014; Wilson et al., Reference Wilson, Lusher, Martin, Rayner and McLachlan2012). Reduced engagement in practice can lead to a self-reported decline in pitch-naming ability (West Marvin et al., Reference West Marvin, VanderStel and Siu2020) and exposure to mistuned stimuli can alter musicians’ pitch templates (Hedger et al., Reference Hedger, Heald and Nusbaum2013; van Hedger et al., Reference Van Hedger, Heald, Uddin and Nusbaum2018). These findings suggest that engagement in music is beneficial for the maintenance and adaptation of pitch-naming skills.
While the role of current practice in maintaining pitch naming is acknowledged, the influence of practice over the course of development has not yet been examined. Practice has largely been considered as present or absent, with little consideration of the role of the amount or intensity of practice. As pitch-naming in adulthood is somewhat malleable when based on practice factors, it stands to reason that practice-related variables during childhood may shape pitch-naming ability as it emerges.
With this in mind, we used data from a large sample of professional musicians to explore whether practice throughout development predicts pitch-naming ability in adulthood, and how this relates to the age of onset of playing. Since previous research has shown that instrument choice also relates to pitch-naming ability (Miyazaki et al., Reference Miyazaki, Rakowski, Makomaska, Jiang, Tsuzaki, Oxenham, Ellis and Lipscomb2018; Wilson et al., Reference Wilson, Lusher, Martin, Rayner and McLachlan2012), we examined the role of the main instrument that musicians played. To determine whether our findings generalized beyond expert musicians, we replicated these analyses in a large population twin sample who played music during childhood and currently. For a subsample of complete monozygotic (MZ) and dizygotic (DZ) twin pairs, we also investigated the importance of genetic and specific environmental factors on pitch naming ability, including a review of the practice histories of high-performing twin pairs.
Methods
Participants
Participants were drawn from two existing datasets consisting of (1) professional musicians and (2) twins. Data were collected between 2012 and 2014, using matched web-based surveys that collected a range of music-related information.
Musician Sample
Professional musicians were recruited through advertisements in professional music publications, Swedish orchestras and music schools (for details, see Wesseldijk et al., Reference Wesseldijk, Mosing and Ullén2021). Sample size was not predetermined, so all participants who responded to the advertisement and fulfilled the following criteria were included. Criteria for inclusion in the present study included participants who (1) currently played music professionally or as tertiary music students, (2) received music lessons during childhood and (3) responded to at least four trials of the pitch-naming task (see task details below). A web-based survey, including music experience-related items (Mosing, Madison et al., Reference Mosing, Madison, Pedersen, Kuja-Halkola and Ullén2014), was completed by 582 musicians in 2013–2014, of which 290 (M age = 48.71, SD = 12.06) met the inclusion criteria.
Twin Sample
Participants were drawn from the Study of Twin Adults: Genes and Environment (STAGE) cohort, which consists of approximately 32,000 twins from the Swedish Twin Registry, born between 1959 and 1985 (Lichtenstein et al., Reference Lichtenstein, De Faire, Floderus, Svartengren, Svedberg and Pedersen2002; Lichtenstein et al., Reference Lichtenstein, Sullivan, Cnattingius, Gatz, Johansson, Carlström, Björk, Svartengren, Wolk, Klareskog, de Faire, Schalling, Palmgren and Pedersen2006). Of this cohort, 11,525 individuals completed the web-based survey in 2012–2013. Sample size was not predetermined as all members of the cohort were invited to participate. Applying the inclusion criteria as detailed for the musicians, this resulted in a sample of 1447 twins (M age = 40.63, SD = 7.99) including 367 complete twin pairs (734 individuals; 193 MZ pairs, 174 DZ pairs; M age = 40.27, SD = 8.11).
The musician and twin studies were approved by the Regional Ethics Review Board in Stockholm (Dnr 2011/570-31/5; Dnr 2013/1777-32) and the Office of Research Ethics and Integrity at the University of Melbourne (2021-14,445-14,697-3). All participants provided informed consent.
Measures
The web-based surveys included questions relating to the participants’ music experience (for further details, see Mosing, Madison et al., Reference Mosing, Madison, Pedersen, Kuja-Halkola and Ullén2014). Relevant items for the current study included: age of onset of musical training, age at which music playing ceased (where appropriate), main instrument and self-reported hours of music practice during different periods of life.
Age of onset of playing
Participants reported the age at which they commenced playing an instrument.
Hours of music practice
Since retrospective self-report has been shown to be a reasonably reliable estimate of practice hours (de Bruin et al., Reference de Bruin, Smits, Rikers and Schmidt2008; Ericsson et al., Reference Ericsson, Krampe and Tesch-Roemer1993; Howard, Reference Howard2011), participants were asked to estimate the number of hours they practiced music per week, based on 10 response options, ranging from 0 to >40 h per week, for each of three age intervals: 0–5, 6–11 and 12–17 years. Based on these responses and the number of years within each age bracket, taking account of age of onset, the total amount of practice within each age bracket was calculated for each participant.
Current weekly practice
Participants estimated the amount they currently practiced music per week. Response options were again divided into 10 categories, ranging from 0 to >40 h per week.
Main instrument
Since previous research has shown that both early and current practice on a ‘fixed-do’ instrument predicts AP (Wilson et al., Reference Wilson, Lusher, Martin, Rayner and McLachlan2012), participants were asked to name their main instrument (or voice). For analysis, responses were categorized into either fixed- or moveable-do instruments. Fixed-do instruments are those with fixed tunings, such as the piano, compared with instruments with pitch that varies based on the musician’s skill (e.g. string and woodwind instruments).
Pitch-naming task
Pitch-naming ability was assessed using a brief online task. Participants were asked to identify the pitch class (chroma) of 24 single-tone auditory stimuli by clicking the note name corresponding to the pitch they heard. Note names were presented on the screen in a circle, with C in the 12 o’clock position. Twelve stimuli were piano tones sampled from a Bösendorfer grand piano in Kontakt software 4 (Native Instruments), and 12 were sine tones generated using Audacity 2.1.2 (Audacity Team, www.audacityteam.org). Stimuli were drawn from the central range (C4-B5), with each chroma presented once as a piano tone and once as a sine tone. Each stimulus was 1 s in duration, and participants were given a maximum of 3 s to respond. A six-trial practice block comprising three piano and three sine tones preceded the task. No feedback was provided throughout the task.
The task was scored and analyzed as the mean absolute deviation (MAD) in cents from the participant’s response to the target chroma across all 24 trials. Larger cent values refer to larger errors, with an error of one semitone (e.g. responding G# when the correct answer is G) corresponding to a deviation of 100 cents. We treated pitch naming as a continuum, rather than applying a strict threshold for AP qualification, to gain an understanding of the full spectrum of the trait and its relation to practice variables. The total number of tones correctly identified was also calculated for descriptive purposes.
Statistical Analyses
Statistical analyses were performed using RStudio (Version 1.3.959), including R packages tidyverse, version 1.3.0 (Wickham et al., Reference Wickham, Averick, Bryan, Chang, D’Agostino McGowan, François, Grolemund, Hayes, Henry, Hester, Kuhn, Lin Pedersen, Miller, Milton Bache, Müller, Ooms, Robinson, Seidel, Spinu, Takahashi and Yutani2019), psych, version 2.1.9 (Revelle, Reference Revelle2021), broom, version 0.7.9 (Robinson et al., Reference Robinson, Hayes and Couch2021), sandwich, version 3.0–1 (Zeileis, Reference Zeileis2004; Zeileis et al., Reference Zeileis, Köll and Graham2020), lmtest, version 0.9-38 (Zeileis & Hothorn, Reference Zeileis and Hothorn2002) and ggridges, version 0.5.3 (Wilke, Reference Wilke2021) and IBM SPSS Statistics 26.
Factors predicting pitch-naming ability
We used hierarchical multiple regressions to investigate the factors associated with pitch-naming ability. To correct for skew, a Box-Cox transformation was applied to the MAD variable (λ = 2.30 for musicians, λ = 18 for twins). Three regression models were fitted to each dataset, first predicting MAD from age, sex, age of onset of playing and main instrument (base model 1). Second, the three accumulated practice variables (i.e. total hours of childhood practice during 0–5, 6–11 and 12–17 years) were added to the base model (Model 2), followed by current weekly practice (Model 3). To adjust for relatedness in the twin sample, the robust standard error estimator for clustered observations was used.
Additional twin analyses
Twin correlations for MAD scores were computed separately for MZ and DZ twin pairs. As the familial environment is assumed to be shared across both members of a twin pair, trait differences between twins may be attributable to genetic and unique environment influences (for DZ twins) or unique environment alone (for MZ twins). To further explore environmental influences on pitch-naming ability, within-pair differences in the practice histories of twins with at least one co-twin performing with a MAD of <200 cents (one tone) were explored in greater detail.
Results
Pitch-Naming Accuracy
Detailed participant information for both samples is shown in Table 1, with additional music experience and pitch-naming task details in Supplementary Table 1. Figure 1 shows the distribution of MAD for both samples. Although musicians outperformed the general population twin sample as expected, pitch-naming accuracy was generally low in both samples, falling along a continuum from chance level to highly accurate performance. Supplementary Figure 1 shows the distribution of correctly identified tones against a binomial (chance) distribution (p = 1/12 [2/24]) for each sample, confirming that most participants (particularly in the population twin sample) performed around the chance level and lacked the ability to reliably name pitches despite being musically active.
Table 1. Demographic and music experience of the participants
a Means reflect spans across categorical response options rather than point estimates.
b MAD = mean absolute deviation, measured in cents (1 semitone = 100 cents).
Fig. 1. Mean absolute deviation (MAD; cents) distributions for (a) musicians and (b) the general population twin sample
Note: Dashed lines indicate the mean MAD for each sample.
Factors Predicting Pitch-Naming Ability in Musicians
Hierarchical multiple regressions showed that the base model predicted the transformed MAD (F[4, 28] = 3.17, p = .014, adjusted R 2 = .03; see coefficients in Table 2). Sex significantly predicted pitch-naming ability, with females showing better performance (lower MAD). Age of onset of playing showed a trend, suggesting that younger age of onset resulted in lower MAD. When adding the three childhood practice variables (Model 2: F[7, 282] = 2.95, p = .005, adjusted R 2 = .05; see coefficients in Table 2), sex showed a trend and age of onset was nonsignificant, while accumulated hours of practice between the ages of 6 and 11 years were associated with lower MAD. In Model 3 (F[8, 281] = 2.71, p = .007, adjusted R 2 = .05), accumulated practice between the ages of 6 and 11 years remained significant (p = .040), with no significant contribution of current practice (p = .309). Regression coefficients for Model 3 and correlations between predictors are shown in Supplementary Tables 2–4.
Table 2. Models predicting pitch-naming ability in musicians
Note: *p < .05, **p < .01, ***p < .001.
Factors Predicting Pitch-Naming Ability in the Population Twin Sample
Regression analyses were replicated in the general population twin sample. Results from Model 1 are shown in Table 3 (F[−4, 1446] = 6.17, p < .001, adjusted R 2 = .01). Both a younger age of onset and playing a fixed-do main instrument predicted better pitch-naming ability (MAD).
Table 3. Models predicting pitch-naming ability in the twin sample
Note: *p < .05, **p < .01, ***p < .001.
Adding the practice variables in Model 2 resulted in age of onset no longer contributing significantly to the model (see Table 3). In this sample, however, accumulated hours of practice between 6 and 11 years did not approach significance, while playing a fixed-do instrument was associated with better pitch-naming ability (MAD). Adding current practice to Model 3 showed that playing a fixed-do main instrument remained significant (p = .002), while current practice did not contribute to the model (p = .422). Regression coefficients for Model 3 and correlations between predictors are shown in Supplementary Tables 5–7.
Twin Correlations for Pitch-Naming Performance in MZ Versus DZ Twins
Twin correlations for MAD in MZ and DZ pairs are shown in Figure 2. MAD was significantly correlated for MZ twin pairs, r = .27, p < .001, indicating that twins within MZ pairs were likely to have similar pitch-naming scores. No such correlation was seen in DZ pairs, r = −.04, p = .63. It should be noted that the MZ correlation was heavily influenced by two high-performing MZ twin pairs, in which both members had very low MADs. There were no similarly outstanding performances by both members of DZ twin pairs. Since it is unlikely that both twins performed so well due to chance and pitch naming is treated as a continuous ability here, we decided to retain these high-performing pairs in the analysis.
Figure 2. Scatterplot showing pitch-naming performance (MAD) for MZ and DZ twin pairs
Practice Profiles of High-Performing Twins and Their Co-Twins
Practice profiles for twin pairs in which at least one individual had a MAD < 200 cents on the pitch-naming task (n = 14 pairs, 28 individuals) are shown in Figure 3.
Figure 3. Practice profiles and pitch-naming performance of high-performing twins and their co-twins.
Note: Panels are to be read vertically, such that aligned columns show data for the same participant. Paired columns show data for a single twin pair, with the highest-performing individual on the pitch-naming task always shown on the left. Panels grouped to the left show monozygotic (MZ) twin pairs, and the right panels show dizygotic (DZ) twin pairs. In the top row, the columns refer to the raw scores on the pitch-naming task (higher values = better performance), while the lollipops refer to the mean absolute deviations (MADs; lower values = better performance). Round tops on the lollipops denote female participants, while diamonds denote males. The practice panels show the accumulated hours of practice for each individual across the three age periods of interest. In these panels, column color indicates main instrument choice: black refers to movable-do instruments, and gray refers to fixed-do instruments.
The majority of MZ and DZ twin pairs played instruments within the same class (both movable-do or both fixed-do) and often within the same instrument family (e.g. both members of Pair 1 play a string instrument and both members of Pair 12 are singers). Where instrument choice differed within a pair, the individual with the higher pitch-naming score was usually the one who played a fixed-do instrument, with the exception of Pair 4. Few participants accumulated any practice hours between the ages of 0 and 5 years, indicating that they did not start playing until after this time. With the exception of one participant (the female member of Pair 8), all twins were musically active between ages 6 and 11. Within MZ pairs, but not DZ pairs, the twin who accumulated more hours of practice in childhood consistently outperformed the twin with fewer hours of practice.
Discussion
In this study, we explored how the age of onset and accumulated music practice throughout childhood relate to pitch-naming ability in adulthood, in both professional musicians and a sample of twins. We found that high pitch-naming accuracy is relatively rare even in a sample with a significant history of music playing. As expected, professional musicians performed somewhat better than amateur musicians from the general population, with pitch-naming accuracy lying along a continuum of ability for both groups. Overall, pitch-naming ability in musicians was significantly predicted by hours of accumulated practice between ages 6 and 11, and an analysis of twin pairs suggested a heritable component to pitch-naming ability.
Factors Predicting Pitch Naming
Age of onset of musical training initially predicted pitch-naming in both samples, although not quite reaching significance in musicians. When accumulated practice during childhood was considered, however, age of onset no longer predicted pitch-naming performance in either group. In musicians, accumulated practice between ages 6 and 11 predicted pitch-naming, whereas the earlier and later practice measures were nonsignificant. Although this finding did not approach significance in the twin sample, the overall direction of results mirrored that of the musicians — an initial contribution of age of onset, disappearing in favor of practice in the 6-to-11 age period. The contribution of practice is consistent with previous findings by Wesseldijk and colleagues (Reference Wesseldijk, Mosing and Ullén2021), who similarly showed that age of onset associations with musical and sporting achievements was no longer significant after adjusting for lifetime accumulated practice. Overall, the finding of a critical window for accumulated practice is consistent with the notion that an early start is relevant. The age bracket of 6–11 years broadly adheres with models of AP, showing that younger age of onset is associated with greater AP prevalence (Baharloo et al., Reference Baharloo, Johnston, Service, Gitschier and Freimer1998; Deutsch et al., Reference Deutsch, Dooley, Henthorn and Head2009; Levitin & Zatorre, Reference Levitin and Zatorre2003; Vitouch, Reference Vitouch2003; Wilson et al., Reference Wilson, Lusher, Martin, Rayner and McLachlan2012).
While the association between early accumulated practice and pitch-naming may be causal, equally it could suggest that a predisposition for good pitch-naming may lead to an increased likelihood of intense early practice (i.e. a reverse causality or gene–environment correlation), or that a generalized predisposition to musicality leads to both intense early practice and good pitch naming (aligning with findings suggesting that a tendency towards practice is heritable; Kragness et al., Reference Kragness, Swaminathan, Cirelli and Schellenberg2020; Mosing, Madison et al., Reference Mosing, Madison, Pedersen, Kuja-Halkola and Ullén2014). Further research is required to explore each of these possibilities.
In the twin sample, choice of main instrument was a significant predictor of pitch-naming accuracy, with playing a fixed-do instrument predicting better performance. This aligns with previous AP research showing that pitch-naming is more accurate in musicians who play fixed-do instruments such as piano (Miyazaki et al., Reference Miyazaki, Rakowski, Makomaska, Jiang, Tsuzaki, Oxenham, Ellis and Lipscomb2018; Wilson et al., Reference Wilson, Lusher, Martin, Rayner and McLachlan2012). In the more expert sample of musicians, however, instrument choice was not a significant predictor, suggesting that the salience of this factor may reduce with greater musical expertise.
Pitch Naming in Twin Pairs
Among twins, there was a significant within-pair correlation in pitch-naming ability for MZ but not DZ pairs, pointing to a role for genetic factors in pitch-naming ability. This is consistent with research suggesting genetic contributions to AP (Baharloo et al., Reference Baharloo, Johnston, Service, Gitschier and Freimer1998; Baharloo et al., Reference Baharloo, Service, Risch, Gitschier and Freimer2000; Gregersen et al., Reference Gregersen, Kowalsky, Kohn and Marvin2001; Gregersen et al., Reference Gregersen, Kowalsky, Lee, Baron-Cohen, Fisher, Asher, Ballard, Freudenberg and Li2013; Profita & Bidder, Reference Profita and Bidder1988; Theusch et al., Reference Theusch, Basu and Gitschier2009; Theusch & Gitschier, Reference Theusch and Gitschier2011); however, this finding should be interpreted with caution due to the small number of high-performing pairs in the sample. It is noteworthy that the pairs in which both members performed well were MZ, and that no similar patterns of performance were seen among DZ pairs.
Although there was much variation in the practice profiles of high-performing twin pairs and too few pairs to draw any firm conclusions, there were a few notable features. Differences in practice hours appeared to be consistently associated with pitch-naming differences among MZ pairs, but not DZ pairs. Although this is a small sample, it suggests that when genetic contributions are held constant (as in MZ pairs), differences in practice intensity may partly explain pitch-naming differences, while additional elements (e.g. playing a fixed- vs. movable-do instrument) may be at play when genetic factors vary (as in DZ pairs). These findings are consistent with an underlying gene–environment interplay. Aligning with the overall predictive model for the general population sample, the higher-performing individuals were generally those who played fixed-do instruments. As in previous work, these high-performing pairs tended to play instruments from the same class (Mosing & Ullén, Reference Mosing and Ullén2018), though there were too few pairs to make comparisons between MZ and DZ pairs.
Limitations and Future Directions
A notable feature of both samples is the rarity of good pitch-naming performance. While this is consistent with an estimated AP prevalence of 1/10,000 (Bachem, Reference Bachem1955), it is not reflective of more recent estimates, particularly when taking QAP-level performance into account. AP has been reported in between <1% and up to 65% of tertiary music students (Deutsch et al., Reference Deutsch, Henthorn, Marvin and Xu2006; Leite et al., Reference Leite, Mota-Rolim and Queiroz2016; Miyazaki et al., Reference Miyazaki, Makomaska and Rakowski2012; Miyazaki et al., Reference Miyazaki, Rakowski, Makomaska, Jiang, Tsuzaki, Oxenham, Ellis and Lipscomb2018), a wide estimate reflective of different ethnicities, populations and accuracy thresholds for AP.
The task used in this study is comparable to other lab-based and web-based pitch-naming studies in terms of stimulus length and response interval (Athos et al., Reference Athos, Levinson, Kistler, Zemansky, Bostrom, Freimer and Gitschier2007; Bermudez & Zatorre, Reference Bermudez and Zatorre2009; Chavarria-Soley, Reference Chavarria-Soley2016; Deutsch et al., Reference Deutsch, Henthorn, Marvin and Xu2006; Miyazaki et al., Reference Miyazaki, Makomaska and Rakowski2012; Oechslin et al., Reference Oechslin, Imfeld, Loenneker, Meyer and Jäncke2010; Wilson et al., Reference Wilson, Lusher, Wan, Dudgeon and Reutens2009), so increased task difficulty is unlikely to account for the low overall performance. More nuanced profiles of intermediate performance could have been gained with an increased number of trials, as increased accuracy for some notes over others would have been evident, but highly accurate AP can be detected with a small number of trials (e.g. 20 trials, Hutka & Alain, Reference Hutka and Alain2015; 24 trials, Aruffo et al., Reference Aruffo, Goldstone and Earn2014). Generally, larger estimates of AP come from studies conducted in East Asian countries, with lower prevalence in Europe and the United States (Miyazaki et al., Reference Miyazaki, Rakowski, Makomaska, Jiang, Tsuzaki, Oxenham, Ellis and Lipscomb2018). Musicians with East Asian ethnicity or language background are more likely to have AP than those of other ethnicities (Deutsch et al., Reference Deutsch, Henthorn, Marvin and Xu2006; Gregersen et al., Reference Gregersen, Kowalsky, Kohn and Marvin2001; Henthorn & Deutsch, Reference Henthorn and Deutsch2007; Miyazaki et al., Reference Miyazaki, Rakowski, Makomaska, Jiang, Tsuzaki, Oxenham, Ellis and Lipscomb2018). Since our sample is drawn from Sweden, this may have contributed to the relatively low prevalence of AP. As this study drew from existing datasets and did not recruit for participants with AP, this may have impacted the prevalence of good pitch-naming in this sample. It is notable that we still found consistent practice-related results in such a low-prevalence sample, which might have been expected to weaken our findings.
Although our results show that pitch-naming ability is likely to be influenced by childhood practice, the extent to which this applies to higher-accuracy AP musicians is yet to be determined. Intermediate pitch-naming ability (QAP), which is largely seen in this sample, may be influenced by practice-related factors to a greater extent than AP. For example, QAP musicians often use ‘reference notes’ to aid their pitch-naming, which tend to be pitches common in their practice experience, such as notes used for tuning (Bachem, Reference Bachem1937; Miyazaki, Reference Miyazaki1988, Reference Miyazaki1989, Reference Miyazaki1990; Wilson et al., Reference Wilson, Lusher, Wan, Dudgeon and Reutens2009). Increased practice-related exposure to these notes leads to their increased recognition in pitch-naming tasks, potentially accounting for the influence of practice in the current sample. Future work with AP musicians should investigate the possibility that different mechanisms may be at play for AP as opposed to QAP musicians.
Conclusion
Our study is the first to show that in expert musicians, the amount of accumulated music practice in early childhood relates to pitch-naming ability in adulthood over and above the age of practice onset. Increased practice during a sensitive period may interact with other predisposing factors to influence pitch-naming aptitude, or a predisposition for pitch-naming ability may encourage additional practice. In a general population sample, playing a fixed-do instrument is associated with improved pitch-naming ability, though this is not evident in expert musicians. The basis of an underlying genetic predisposition to pitch-naming requires additional exploration, though this endeavor is challenged by the rarity of good pitch-naming ability, even in our large genetically informative sample. Our preliminary exploration shows that pitch naming is more similar in MZ than DZ twin pairs, supporting a heritable component to this skill.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1017/thg.2022.29.
Author Contributions
JB analyzed the data, interpreted the results and drafted the manuscript. FU and MM designed the study and collected the data. MM and SW interpreted the results. All authors provided critical revisions and contributed to the manuscript. JB is responsible for the integrity of the data analyzed.
Financial Support
The present work was supported by the Bank of Sweden Tercentenary Foundation (M11-0451:1), the Sven and Dagmar Salén Foundation and the Marcus and Amalia Wallenberg Foundation (MAW 2018.0017). The Swedish Twin Registry is managed by Karolinska Institutet and receives funding through the Swedish Research Council under the grant no 2017-00641.
Conflict of Interest
None.
Open access
