INTRODUCTION
Music is one of the most complex human abilities and has been the object of numerous investigations from specialists in both the human and natural sciences. In recent years, neuroscience has made a substantial contribution to our understanding of the special characteristics of musical experience. Indeed, by using currently available brain imaging techniques, it is now possible to investigate the changes that music produces in the healthy human brain. From this point of view, a substantial amount of evidence converges to suggest that musical training is a powerful means for reorganization of brain structures (Jäncke, Reference Jäncke2009; Schlaug, Reference Schlaug2015). According to these investigations, it appears that changes linked to musical experiences take place in both the gray matter of specific cerebral regions and the size, volume, and composition of white matter fiber bundles (Pantev et al., Reference Pantev, Oostenveld, Engelien, Ross, Roberts and Hoke1998; Gaser & Schlaug, Reference Gaser and Schlaug2003; Oechslin, Imfeld, Loenneker, Meyer, & Jäncke, Reference Oechslin, Imfeld, Loenneker, Meyer and Jäncke2010; Halwani, Loui, Rüber, & Schlaug, Reference Halwani, Loui, Rüber and Schlaug2011). As a result of this music-dependent rearrangement of neuronal networks, musicians seem to possess a different connectome than non-musicians (Schmithorst & Wyke, Reference Schmithorst and Wilke2002).
In this regard, the corpus callosum, the largest white matter fiber network in the human brain, has attracted a great deal of attention. It is well known that the corpus callosum mediates interactions between the hemispheres, which are essential for many aspects of human behavior (Gazzaniga, Reference Gazzaniga2000, Reference Gazzaniga2005). In particular, its role in mediating the transfer of information between the two hemispheres is strongly supported by studies that have examined the effects of partial or complete callosotomy (Myers & Sperry, Reference Myers and Sperry1985; Funnell, Corballis, & Gazzaniga, Reference Gazzaniga2000).
Many studies have found that interhemispheric transfer is affected by development and aging (Davis, Kragel, Madden, & Cabeza, Reference Davis, Kragel, Madden and Cabeza2012; Jeeves & Moes, Reference Jeeves and Moes1996; Schulte, Sullivan, Müller-Oehring, Adalsteinsson, & Pfefferbaum, Reference Schulte, Sullivan, Müller-Oehring, Adalsteinsson and Pfefferbaum2005). The capacity for information transfer between hemispheres progressively improves until around the age of 10–14 years, when it reaches performance levels typical of adulthood. Starting around the sixth decade of life, there is a progressive decline in performance (Boyson, Reference Boyson2013; Reuter-Lorenz & Stanczak, Reference Reuter-Lorenz and Stanczak2000; Scally, Burke, Bunce, & Delvenne, Reference Scally, Burke, Bunce and Delvenne2018). For example, significant differences have been reported between adults aged 35–45 and 55–60 years (Bellis & Wilber, Reference Bellis and Wilber2001). This age-related decline appears to occur in conjunction with the onset of atrophy of the corpus callosum. Indeed, the majority of neuroimaging studies have found that both the global surface of the callosal area and the integrity of the callosal fibers progressively decrease, following an inverted-U curve across the human lifespan (Hasan et al., Reference Hasan, Kamali, Kramer, Papnicolaou, Fletcher and Ewing-Cobbs2008; Sullivan & Pfefferbaum, Reference Sullivan and Pfefferbaum2006). Furthermore, post-mortem investigations have shown that from the sixth to eighth decade of life, there is a 37.9% reduction in the total number of myelinated fibers in the corpus callosum (Hou & Pakkenberg, Reference Hou and Pakkenberg2012; Riise & Pakkenberg, Reference Riise and Pakkenberg2011; Tang, Nyengaard, Pakkenberg, & Gundersen, Reference Tang, Nyengaard, Pakkenberg and Gundersen1997).
According to the disconnection hypothesis of cognitive aging (Bennett & Madden, Reference Bennett and Madden2014; O’Sullivan et al., Reference O’Sullivan, Jones, Summers, Morris, Williams and Markus2001), these age-related structural and functional alterations of the corpus callosum may account for the onset of cognitive disorders in normal aging. In older populations, the severity of callosal atrophy correlates with the degree of cognitive impairment and there is accumulating evidence that callosal dysfunctions are associated with the impairment of memory and executive dysfunctions (Ryberg et al., Reference Ryberg, Rostrup, Paulson, Barkhof, Scheltens and van Straaten2011; Zahr, Rohlfing, Pfefferbaum, and Sullivan, Reference Zahr, Rohlfing, Pfefferbaum and Sullivan2009). Furthermore, severe callosal atrophy has been recorded in subjects with degenerative and vascular dementia (Hallam et al., Reference Hallam, Brown, Ross, Buckwalter, Bigler, Tschanz and Breitner2008). Conversely, several investigations suggest that integrity of the corpus callosum predicts both better motor and cognitive abilities in old age (Johansen-Berg, Della-Maggiore, Behrens, Smith, & Paus, Reference Johansen-Berg, Della-Maggiore, Behrens, Smith and Paus2007; Kennedy & Raz, Reference Kennedy and Raz2009). Moreover, several investigations have correlated successful aging with maintaining corpus callosum integrity (Madden et al., Reference Madden, Spaniol, Costello, Bucur, White, Cabeza and Huettel2009; Penke et al., Reference Penke, Muñoz Maniega, Houlihan, Murray, Gow, Clayden and Deary2010; Wolf et al., Reference Wolf, Fischer, Fesenbeckh, Yakushev, Lelieveld, Scheurich and Fellgiebel2012). In summary, while white matter atrophy seems to be a global phenomenon in typical aging (O’Sullivan et al., Reference O’Sullivan, Jones, Summers, Morris, Williams and Markus2001; Salat et al., Reference Salat, Tuch, Greve, van der Kouwe, Hevelone, Zaleta and Dale2005; Salthouse, Reference Salthouse2011), there appears to be a direct relationship between brain aging, cognitive functioning, and the function of the corpus callosum. Damage to the corpus callosum seems to be a strong predictor of cognitive and motor deterioration later in life, whereas its functional efficiency appears to protect against cognitive aging and can help explain, at least in part, the mechanisms underlying successful aging (Penke et al., Reference Penke, Muñoz Maniega, Houlihan, Murray, Gow, Clayden and Deary2010; Persson et al., Reference Persson, Nyberg, Lind, Larsson, Nilsson, Ingvar and Buckner2006).
Due to the putative impact of corpus callosum integrity on general cognitive processes, it is important to examine its efficiency. It should be borne in mind that neuroradiological tests can only provide indirect information on connectivity. For example, it has been stressed that even when the size of the corpus callosum is normal, the effectiveness of connections can be reduced; abnormalities on a microscopic level, which cannot be detected by neuroimaging methods, can yet play an important functional role (Sullivan & Pfefferbaum, Reference Sullivan and Pfefferbaum2006). In this regard, neuropsychological and behavioural studies can be of additional value.
To explore the effectiveness of information transfer, different behavioral paradigms have been proposed (Poffenberger, Reference Poffenberger1912; Reuter-Lorenz & Stanczak, Reference Reuter-Lorenz and Stanczak2000; Saron & Davidson, Reference Saron and Davidson1989; Thompson, Narr, Blanton, & Toga, Reference Thompson, Narr, Blanton, Toga, Zaidel and Iacoboni2003) and a number of studies have correlated neuroradiological indexes of callosal integrity with behavioral measures of interhemispheric transfer (Schulte & Müller-Oehring, Reference Schulte and Müller-Oehring2010; Westerhausen et al., Reference Westerhausen, Kreuder, Woerner, Huster, Smit, Schweiger and Wittling2006).
Currently, research shows a close relationship between the organization of the corpus callosum and musical training. There is evidence that individuals with extensive musical training have a larger corpus callosum and show more efficient white matter organization in the corpus callosum (Oztürk, Tasçioglu, Aktekin, Kurtoglu, & Erden, Reference Oztürk, Tasçioglu, Aktekin, Kurtoglu and Erden2002; Schlaug, Jäncke, Huang, Staiger, & Steinmetz, Reference Schlaug, Jäncke, Huang, Staiger and Steinmetz1995) and music can trigger plastic changes in the corpus callosum (Jäncke, Reference Jäncke2009; Wan and Schlaug, Reference Wan and Schlaug2010). Based on these findings, it is likely that professional musicians would perform differently on interhemispheric transfer tests than control subjects. However, it has not yet been clearly determined whether music-induced plastic changes in the corpus callosum correspond to more efficient interhemispheric interaction.
In the present study, we tested the hypothesis that musical training can enhance cross-hemispheric communication by comparing the performance of professional musicians and non-musicians in the fingertip cross-localization test, an experimental paradigm widely used in split-brain subjects to measure the transfer of somesthetic information (Bogen, Reference Bogen, Heilman and Valenstein1979; Fabri et al., Reference Fabri, Del Pesce, Paggi, Polonara, Bartolini, Salvolini and Manzoni2005). We hypothesized that changes in callosal plasticity induced by musical exercise would be associated with greater accuracy in the interhemispheric transfer of information. Our main working hypothesis was that older professional musicians should make significantly fewer errors during the fingertip cross-localization test than age-matched control subjects. This would be an indication that the functional efficiency of the corpus callosum was maintained despite age. Second, we tried to assess whether the musicians’ test performance was correlated with a number of variables linked to individual musical competence.
METHODS
Participants
Sixty-five professional classical musicians (42 male and 23 female) aged >55 years old (M = 68.89, SD = 10.33, age range = 56–90) were recruited for this study at the Academy of Santa Cecilia in Rome, the Casa della Musica Giuseppe Verdi in Milan, and the music conservatories of Trento, Bolzano, Perugia, and L’Aquila. The control group of 65 participants (42 male and 23 female), aged 56–86 years (M = 69.95, SD = 8.96), was selected from a pool of 200 healthy, right-handed subjects examined in our laboratory to best correspond in terms of age and sex to each musician in the experimental group. All subjects were preliminarily examined in a clinical interview aimed at assessing their general state of health, general cognitive functioning, and independence in daily life. The general exclusion criteria were the following: an adjusted score of <24 on the Mini-Mental State Examination (Magni, Binetti, Bianchetti, Rozzini, & Trabucchi, Reference Magni, Binetti, Bianchetti, Rozzini and Trabucchi1996); the presence of cerebrovascular, cardiovascular, metabolic, oncologic, or neuropsychiatric diseases (pharmacological treatments, such as antihypertensive therapy, for keeping clinical conditions under control, were not considered a criterion for exclusion); and the presence of difficulty with hand movements (e.g., people suffering from arthritis). Additionally, control subjects were excluded if they were amateur musicians or if their profession was based on digital ability (e.g., typists). Musicians were also given a semi-structured interview aimed at evaluating some variables associated with musical performance. This interview took handedness into account as measured by the Edinburgh Inventory (Oldfield, Reference Oldfield1971), type of professional activity (orchestral players or teachers), the instrument used (divided into four categories: keyboards, strings, brass, and multi-instrumentalist), the age at which musical training began, the average duration (in hours) of daily training extrapolated across their lifetime, and the presence of family musicians during childhood. All participants gave informed consent. The research was completed in accordance with the Declaration of Helsinki.
The Fingertip Cross-Localization Test
To properly study communication between the two hemispheres, it is necessary to utilize tasks in which the correct execution compels the subject to use interhemispheric transfer. In the fingertip cross-localization test, subjects must respond to a tactile stimulus presented to one hand using the ipsilateral or contralateral hand. With hands out of view, the tip of a finger is lightly touched with a sharpened pencil point. Subjects are then asked to indicate which finger was touched by using the thumb of the same hand (uncrossed condition) or by using the corresponding finger of the opposite hand to touch the thumb of the opposite hand (crossed condition). The crossed condition requires interhemispheric transfer of information because one hemisphere is presented with the stimulus, which the other hemisphere has to use to respond correctly. The uncrossed conditions do not require any transfer since both the stimulus and response are lateralized to the same hemisphere. Performance can be measured in terms of accuracy by calculating the difference in the number of correct answers to crossed and uncrossed tests (crossed–uncrossed difference; CUD). Theoretically, if the transfer between the two hemispheres was entirely efficient, the performance under crossed (interhemispheric) and uncrossed (intrahemispheric) conditions should not be different. Conversely, the value of the CUD is expected to increase proportionally with reduction in functional efficiency.
Data from the literature indicate that the number of errors, measured as the CUD, gradually decreases during childhood and adolescence due to progressive improvement in the interaction between the two hemispheres (Innocenti, Reference Innocenti1981); progressive changes concern the number and myelination of fibers on the one hand and synaptic organization on the other, until the specificity of connections typical of adulthood is reached around age 14 (Galin, Johnston, Nakell & Herron, Reference Galin, Johnston, Nakell and Herron1979; O’Leary, Reference O’Leary1980; Quinn & Geffen, Reference Quinn and Geffen1986). Healthy adult subjects make a number of errors that varies from 0% to 10% and increases progressively with increasing age (Bogen, Reference Bogen, Heilman and Valenstein1979; Fabri et al., Reference Fabri, Del Pesce, Paggi, Polonara, Bartolini, Salvolini and Manzoni2005; Piccirilli, Finali, & Sciarma, Reference Piccirilli, Finali and Sciarma1989; Piccirilli et al., Reference Piccirilli, D’Alessandro, Germani, Boccardi, Pigliautile, Ancarani and Dioguardi2020). In patients, the number of errors varies depending on pathological condition (Brown & Paul, Reference Brown and Paul2019; Fabbro et al., Reference Fabbro, Pesenti, Facoetti, Bonanomi, Libera and Lorusso2001; Roebuck, Mattson & Riley, Reference Roebuck, Mattson and Riley2002). As expected, the performance of patients with a total callosotomy does not differ from a random response, that is the number of errors is about 75% (Funnell et al., Reference Funnell, Corballis and Gazzaniga2000; Geffen et al., Reference Geffen, Nilsson, Quinn and Teng1985; Myers & Sperry, Reference Myers and Sperry1985), However, patients with a partial callosotomy show various degrees of abnormalities depending on the site of the cut (Caillé, Sauerwein, Schiavetto, Villemure, & Lassonde, Reference Caillé, Sauerwein, Schiavetto, Villemure and Lassonde2005; Devinsky & Laff, Reference Devinsky and Laff2003; Lassonde, Sauerwein, Geoffroy, & Décarie, Reference Lassonde, Sauerwein, Geoffroy and Décarie1986). In particular, after surgical resection of different portions of the midbody, in subjects with intact genu and splenium, the occurrence of poor transfer has been documented only when the lesion was located in the portion posterior to the interventricular foramen (Bentin, Sahar, & Moscovitch, Reference Bentin, Sahar and Moscovitch1984). Also, data derived from functional MRI confirm that the tactile channel passes through the corpus callosum at the posterior midbody before the splenium (Hofer & Frahm, Reference Hofer and Frahm2006; Polonara et al., Reference Polonara, Mascioli, Foschi, Salvolini, Pierpaoli, Manzoni and Barbaresi2015).
Procedure
Subjects were seated at a table, placing their forearms directly in front of them on the table, with palms facing upwards. They looked at a blank screen, which kept their hands out of view. The experimenter touched the tip of one of four fingers with the point of a sharpened pencil, asking the subject to then touch that same finger with the tip of the thumb of the same hand (uncrossed condition) or to touch the corresponding finger on the opposite hand with the thumb of that opposite hand (crossed condition). Subjects responded immediately after stimulation, and the examiner recorded each response on his answer sheet. No feedback was given concerning performance accuracy. The test was first demonstrated with the subject’s hand in full view until he/she understood what was required.
There were 192 trials altogether, in sets of 48 trial sequences for each hand/task (right-to-right, right-to-left, left-to-left, and left-to-right). Each finger was stimulated the same number of times (12 times each for the four hand/task conditions). All subjects performed the uncrossed localization condition trials first and the crossed localization condition trials afterward. Choice of the hand on which stimulation was carried out was counterbalanced across all subjects. Trial sequences for each condition were different. The examination took around 10–15 minutes.
The score was based on the number of correct responses (with 48 as the maximum value for each condition). The CUD (difference in the overall number of errors between uncrossed and crossed tasks) was calculated by using the following formula: (uncrossed test on the right hand − crossed test on the right hand) + (uncrossed test on the left hand − crossed test on the left hand).
Data Analysis
First, a mixed ANOVA with Group (controls vs. musicians) as the between factor and Hand (right vs. left) and Condition (uncrossed-hand vs. crossed-hand) as repeated measures was run using the number of correct responses. The effect size was measured using partial eta-squared, with 0.0099, 0.0588, and 0.1379 being small, medium, and large effects, respectively (Cohen, Reference Cohen1988). Post hoc analyses were run using two independent-sample and two dependent-sample t tests. To account for multiple comparisons, the threshold at which p-values were deemed significant was halved (.05/2 = .025).
Second, to compare the CUD between controls and musicians, an independent-sample t test was carried out. The effect size was measured using Cohen’s d for independent and dependent samples, respectively, in which benchmarks are small (d ≤ .2), medium (d ≤ .5), and large (d ≤ .8). As a further index of the efficiency of interhemispheric transfer, the number of control and musician subjects who performed the test without making any mistakes was calculated.
A ceiling effect in musicians made it difficult to analyze the potential impact of personal variables on their test performance (see below). Thus, musicians were split into two separate groups (errors vs. no-errors) based on the presence or absence of differences in scores between crossed and uncrossed conditions. A series of independent-sample t tests were run to compare the errors vs. no-error groups in terms of handedness, the age at which musical training began, and the average daily number of hours practiced. A Chi-squared test was carried out to test associations between errors and the type of professional activity, instrument played, and presence of family musicians during childhood.
Finally, since there was no ceiling effect in the control group and thus greater sensitivity to age-related changes, Pearson’s correlations between age and score (i.e., number of corrected responses and CUD) were calculated to test for age-related changes in performance.
RESULTS
Number of Correct Responses
There was a significant main effect of Group (F
(1) = 63.27, p < .001,
$\eta _p^2$
= .331) on the number of correct responses. As shown in Table 1, the musicians reported higher scores than the controls. The main effect of group indicates that, if we ignore all other variables, musicians’ scores were significantly higher than controls. The main effect of Hand (F
(1) = 1.00, p = .318,
$\eta _p^2$
= .008) was not significant, showing that there were no differences between performance with the right and left hand. There was a significant main effect of Condition (F
(1) = 117.04, p < .001,
$\eta _p^2$
= .478), with higher scores for the uncrossed than the crossed condition.
Fingertip localization test, with mean correct scores (standard deviations and percentages) for each hand condition
With regard to interaction effects, there was no significant interaction between Hand and Condition (F
(1) = 0.66, p = .416,
$\eta _p^2$
= .005), or between Hand and Group (F
(1) = 1.00, p < .318,
$\eta _p^2$
= .008), while there was a significant interaction between Condition and Group (F
(1) = 57.89, p < .001,
$\eta _p^2$
= .311) with respect to the number of correct responses. The latter interaction means that uncrossed and crossed conditions had a different effect on performance in musicians and controls. Finally, there was no significant interaction between Hand, Condition, and Group (F
(1) = 0.99, p = .321,
$\eta _p^2$
= .008).
Post hoc analyses indicated significant differences between musicians and controls, both in uncrossed [t(128) = –2.55, p = .012, d = 0.45] and crossed [t(128) = –8.02, p < .001, d = 1.41] conditions. Furthermore, two paired-sample t tests showed significant differences between uncrossed and crossed conditions, both among the controls [t(64) = 11.09, p < .001, d = 1.37] and the musicians [t(64) = 2.88, p = .005, d = 0.35]. Although musicians had higher scores than controls in both conditions (group effect), the effect size was higher in the crossed condition (large) than the uncrossed condition (small). Furthermore, although the differences between the uncrossed and crossed conditions were significant both for controls and musicians (condition effect), the effect size of this difference was small for musicians and large for controls (Figure 1).
. Number of correct responses related to Group × Condition. The error bars depict the standard error of the mean.
CUD (Difference in the Overall Number of Errors between Uncrossed and Crossed Tasks)
There was a significant difference with a large effect size between the controls and musicians in the number of errors between uncrossed and crossed tasks [t(128) = 7.85, p < .001, d = 1.38]. The CUD was lower in the musicians (M = 0.91, SD = 2.27) than in the controls (M = 4.85, SD = 3.34).
The comparison of the CUD values explicitly shows that the two groups differed in the crossed condition.
Moreover, the distribution of participants who performed the task without making any mistakes across groups was not equal [
$\chi _{(1)}^2$
= 60.27, p < .001], with 7 (10.8 %) controls and 51 (78.5%) musicians; 14 musicians and 54 non-musicians made at least one error.
Individual Variables
As for the potential effect of personal variables on performance within the musician group (error vs. no-error), there were no significant differences in terms of handedness, age at onset of musical practice, and daily duration of training (Table 2). Regarding handedness, 54 of the 65 musicians had the Edinburgh Handedness Inventory laterality index ≥ 80; the musician group included 7 ambidextrous and 4 left-handed subjects. Moreover, there was no significant association between errors vs. no-errors and type of professional work [
$\chi _{(1)}^2$
= 0.11, p = .736], instrument played [
$\chi _{(3)}^2$
= 2.05, p = .563], or the presence of family musicians during childhood [
$\chi _{(1)}^2$
= 0.53, p = .468] (Table 3).
No effect of individual variables on interhemispheric transfer performance within the musician group (no-error vs. error group)
. Contingency table showing no effect of individual variables on interhemispheric transfer performance within the musician group (no-error vs. error group)
Based on these findings, musicians’ performance is apparently not influenced by individual variables.
Correlations between Age and Scores in the Control Group
Within the control group, age was not associated with the number of correct responses in the uncrossed condition (r = −.09, p = .487), but was significantly and negatively associated with the number of correct responses in the crossed condition (r = −.31, p = .011) and significantly and positively associated with CUD (r = .27, p = .028).
These findings indicate that, in the control group, the number of errors in the crossed condition increases with increasing age.
DISCUSSION
In the present research, efforts were made to verify if playing a musical instrument is associated with enhanced interaction between the two hemispheres in older healthy subjects. Our results indicate that, in contrast to age-matched control subjects, professional musicians perform significantly better on the test of interhemispheric transfer.
It is worth mentioning that the fingertip cross-localization test has proven to be a reliable behavioral index of the functional effectiveness of interhemispheric communication. Since the scores obtained from the crossed test were significantly lower than the scores obtained from the uncrossed test, regardless of the hand used, the experimental paradigm supports the hypothesis of greater difficulty in the crossed interhemispheric condition and suggests that transfer of information from one hemisphere to another implies a significant loss of accuracy. Therefore, the degree of difference between the uncrossed and crossed conditions can be considered an adequate measure of information transfer processes.
In interpreting our results, it must be underlined that the musicians we examined showed a ceiling effect, which could indicate that the finger cross-localization test lacked sensitivity. The significant group difference in CUD indicates that the test was adequate to reveal differences between musicians and control subjects in interhemispheric transfer efficiency. Because the task measures (number of crossed errors and CUD) were significantly associated with age in the non-musician group, it is likely that the non-musician group declined in this measure of interhemispheric transfer as a consequence of increasing age. Given the evidence in the literature that music affects the corpus callosum, the data could be interpreted as demonstrating that in musicians, the efficiency of interhemispheric communication is protected from age-related decline. However, it should be noted that the ceiling effect may have obscured age-related decline in musicians. In short, the ceiling effect does not allow us to distinguish between a stable non-age-related group difference and a group difference related to significantly less impairment in performance in musicians than non-musicians. In other words, younger musicians may already enjoy a much higher level of interhemispheric efficiency which also reduces with age. A longitudinal investigation or the use of a more complex transfer test would help determine whether music playing confers specific protection from age-related changes in interhemispheric connectivity.
It should also be noted that the test requires the transfer of both tactile and motor information and that, in our investigation, musicians obtained higher scores than controls, not only in the crossed condition but also in the uncrossed condition. Superior tactile and motor performance have previously been reported in professional musicians in comparison to non-musicians, and these findings were assumed to arise from cortical reorganization induced by musical training (Fling & Seidler, Reference Fling and Seidler2011; Ragert, Schmidt, Altenmüller, & Dinse, Reference Ragert, Schmidt, Altenmüller and Dinse2004). However, since this study is concerned with CUD and the crossed scores are significantly lower than the uncrossed ones, it is likely that this suggestion is not critical to data interpretation: tactile acuity and motor ability cannot explain the excellent performance in the uncrossed task, regardless of hand used and the decrease of scores in the crossed task.
Based on this experimental paradigm, we can conclude that the data of our study present evidence for more efficient interhemispheric transfer of tactile sensory information in older musicians compared to non-musicians; moreover, they suggest that musical practice improves connectivity between the two hemispheres.
Our findings are in accordance with studies showing that playing an instrument may influence neuroplasticity (Münte, Altenmüller, & Jäncke, Reference Münte, Altenmüller and Jäncke2002; Wan & Schlaug, Reference Wan and Schlaug2010). By now, it is well documented that making music is accompanied by a remarkable rearrangement of brain structures (Elbert, Pantev, Wienbruch, Rockstroh, & Taub, Reference Elbert, Pantev, Wienbruch, Rockstroh and Taub1995; Herholz & Zatorre, Reference Herholz and Zatorre2012; Hyde et al., Reference Hyde, Lerch, Norton, Forgeard, Winner, Evans and Schlaug2009; Schlaug, Reference Schlaug2015). This brain reorganization concerns both gray and white matter and seems to be justified by the peculiar complexity of musical practice: performing music necessitates a wide array of brain regions, which must work in a coordinated manner throughout both hemispheres (Peretz & Zatorre, Reference Peretz and Zatorre2005). In this perspective, the accuracy of interhemispheric exchanges plays an irreplaceable role (Schulte & Müller-Oehring, Reference Schulte and Müller-Oehring2010; Moore, Schaefer, Bastin, Roberts, & Overy, Reference Moore, Schaefer, Bastin, Roberts and Overy2014).
Changes induced by musical training may justify the protective benefits on cognition documented by current research (Bugos, Perlstein, McCrae, Brophy, & Bedenbaugh, Reference Bugos, Perlstein, McCrae, Brophy and Bedenbaugh2007; Diaz Abrahan, Shifres, & Justel, Reference Diaz Abrahan, Shifres and Justel2019; Hanna-Pladdy & MacKay, Reference Hanna-Pladdy and MacKay2011; Koelsch, Reference Koelsch2009). For example, individualized piano instruction enhanced executive functioning and working memory in musically naïve older adults aged 60–85 years (Seinfeld, Figueroa, Ortiz-Gil, & Sanchez-Vives, Reference Seinfeld, Figueroa, Ortiz-Gil and Sanchez-Vives2013). In this view, making music is considered a factor that can mitigate or even prevent the consequences of brain aging (Fauvel, Groussard, Eustache, Desgranges, & Platel, Reference Fauvel, Groussard, Eustache, Desgranges and Platel2013; Jäncke, Reference Jäncke2013). Furthermore, musical training protects aging individuals from pathological cognitive decline. In a population-based twin study (Balbag, Pedersen, & Gatz, Reference Balbag, Pedersen and Gatz2014), playing a musical instrument was significantly associated with less likelihood of dementia and cognitive impairment, supporting suggestions that differences observed between musician and non-musician twins are likely due to musical training rather than pre-existing biological differences (de Manzano & Ullén, Reference de Manzano and Ullén2018). From this point of view, music playing has been considered a transformative technology of the mind (Patel, Reference Patel and Honing2019). In other words, making music appears to preserve brain functions (Rogenmoser, Kernbach, Schlaug, & Gaser, Reference Rogenmoser, Kernbach, Schlaug and Gaser2017).
These effects of individual musical expertise on brain connectome and cognitive functions can be of great significance with regard to brain aging. There is noteworthy inter-individual variability in aging (Raz, Ghisletta, Rodrigue, Kennedy, & Lindenberger, Reference Raz, Ghisletta, Rodrigue, Kennedy and Lindenberger2010), a finding that has been interpreted in relation to the notion of reserve capacity: brain reserve is thought to be related to the structural characteristics of the individual nervous system, and cognitive reserve to more efficient configuration of the neural network (Stern, Reference Stern2009; Valenzuela, Reference Valenzuela2008). In this view, the same number of neurons can form networks that are connected in diverse ways and a well-connected network can more easily mitigate the effects of aging (Madden et al., Reference Madden, Spaniol, Costello, Bucur, White, Cabeza and Huettel2009; Raz et al., Reference Raz, Ghisletta, Rodrigue, Kennedy and Lindenberger2010). Thus, the same degree of brain atrophy may manifest itself in distinctly different ways in different individuals (Barulli & Stern, Reference Barulli and Stern2013; Park & Reuter-Lorenz, Reference Park and Reuter-Lorenz2009). A growing body of literature indicates that interventions with the capacity to integrate multiple neural networks may be an effective way to enhance reserve capacity (Stern, Reference Stern2009), as occurs with musical training (Boyke, Driemeyer, Gaser, Büchel, & May, Reference Boyke, Driemeyer, Gaser, Büchel and May2008). Through musical exercise, neuronal networks began to operate in an integrated manner (Bangert & Altenmuller, Reference Bangert and Altenmuller2003; Bangert et al., Reference Bangert, Peschel, Schlaug, Rotte, Drescher, Hinrichs and Altenmüller2006; Baumann et al., Reference Baumann, Koeneke, Schmidt, Meyer, Lutz and Jäncke2007). It is reasonable to assume that differences in the efficiency of the configuration of neural networks induced by musical training may lead to the differences in brain-aging profiles found for musicians (Balbag et al., Reference Balbag, Pedersen and Gatz2014; Rogenmoser et al., Reference Rogenmoser, Kernbach, Schlaug and Gaser2017). Musical engagement seems to shape brain structures in such a way as to maintain better reserve capacity (Hanna-Pladdy & Gajewski, Reference Hanna-Pladdy and Gajewski2012). For example, musicians showed no or smaller decreases in grey matter density in the frontal cortex compared with non-musicians (Jäncke, Reference Jäncke2013; Sluming et al., Reference Sluming, Barrick, Howard, Cezayirli, Mayes and Roberts2002). In our opinion, musical performance can promote a similar use-dependent retention of white matter involving the corpus callosum: making music may be able to protect the efficiency of the corpus callosum (Moore et al., Reference Moore, Schaefer, Bastin, Roberts and Overy2014). Brain plasticity depends, in large part, on experiences that have taken place over the course of a lifetime and continue throughout the individual’s life if adequate incentives are available (Park & Bischof, Reference Park and Bischof2013). Music could be an incentive of this kind (Fukui & Toyoshima, Reference Fukui and Toyoshima2008).
A limitation of our study is that the participants did not undergo in-depth neuropsychological assessment, which prevents any consideration of the relationships between individual interhemispheric information transfer capacity and cognitive performance. The findings of our study do not allow us to ascertain the cognitive significance of interhemispheric transfer efficiency. However, it should be underlined that, in comparison to the control subjects, a significant percentage of the musicians performed the transfer test without errors. We hypothesize that individuals who maintain interhemispheric transfer test performance at levels similar to those achieved at younger ages may have avoided any negative influence of aging on the efficiency of their brain function. To confirm this hypothesis – that musicians are able to demonstrate both successful aging and perform significantly better on interhemispheric transfer tests than subjects with typical aging – further research is required.
Furthermore, it is important to bear in mind that our data refer only to the transfer of tactile information and cannot be generalized to other functions. In fact, the corpus callosum should be considered a collection of fiber bundles with specific functions and structural properties, rather than a single bundle of fibers (Aboitiz & Montiel, Reference Aboitiz and Montiel2003; Sperry, Gazzaniga, & Bogen, Reference Sperry, Gazzaniga, Bogen, Vinken and Bruyn1969). Therefore, it is possible that the analysis of other modes of transfer might provide different results in relation to the functional differences of the different subunits contained within the corpus callosum. Further research would be useful in order to establish whether interhemispheric transfer tests are reliable and easy-to-use clinical tools for monitoring aging-related brain changes.
Another question raised by our study concerns the relationship between musical expertise and the efficiency of callosal functions (Merrett, Peretz, & Wilson, Reference Merrett, Peretz and Wilson2013). In the available scientific literature, brain changes due to musical training have been shown to correlate with a number of variables linked to individual musical competence: musical instrument played (e.g., the sensorimotor representation of the hands is different for violinists and pianists); age of onset (which distinguishes musicians who began their musical training before or after 7 years of age); duration and intensity of training (which could differentiate professional musicians from those who just play for fun); and so on (Bangert & Schlaug, Reference Bangert and Schlaug2006; Skoe & Kraus, Reference Skoe and Kraus2012; Vaquero et al., Reference Vaquero, Hartmann, Ripollés, Rojo, Sierpowska, François and Altenmüller2016). Our study has not been able to discriminate between the roles of these different variables in individual expertise. It must be underlined that the musicians we examined showed a ceiling effect, which can prevent correct analysis of the data, that is, differences may have been non-significant due to the ceiling effect. It remains to be assessed whether the role of the various individual factors can be better revealed using a more complex interhemispheric transfer test. However, it is also possible that the expertise acquired over years of continuous musical practice by professional musicians can reduce the importance of individual differences (Oechslin, Van De Ville, Lazeyras, Hauert, & James, Reference Oechslin, Van De Ville, Lazeyras, Hauert and James2013).
CONCLUSION
One of the most promising areas of current research is the identification of strategies, which can favorably alter plasticity to cope with age-related changes (Christie et al., Reference Christie, Hamilton, Manor, Farb, Farzan, Sixsmith and Moreno2017; Gow et al., Reference Gow, Bastin, Muñoz Maniega, Valdés Hernández, Morris, Murray and Wardlaw2012; Piccirilli, Pigliautile, Arcelli, Baratta & Ferretti, Reference Piccirilli, Pigliautile, Arcelli, Baratta and Ferretti2019; Wirth, Haase, Villeneuve, Vogel, & Jagust, Reference Wirth, Haase, Villeneuve, Vogel and Jagust2014). The literature suggests that music is one of the most effective interventions for positively shaping organization in the aging brain (Diaz Abrahan et al., Reference Diaz Abrahan, Shifres and Justel2019; Hanna-Pladdy & Gajewski, Reference Hanna-Pladdy and Gajewski2012; Jäncke, Reference Jäncke2013; Rogenmoser et al., Reference Rogenmoser, Kernbach, Schlaug and Gaser2017; Seinfeld et al., Reference Seinfeld, Figueroa, Ortiz-Gil and Sanchez-Vives2013). In our study, older professional musicians made significantly fewer errors on the tactile interhemispheric transfer test than age-matched non-musicians. This finding is in line with a great deal of investigations showing that individuals with extensive musical training have a larger corpus callosum. It may indicate a relationship between music playing, music-induced changes in plasticity, and the efficiency of interhemispheric communication.
ACKNOWLEDGMENTS
We thank all the study participants and staff at the Academy of Santa Cecilia in Rome and the Casa della Musica Giuseppe Verdi in Milan.
CONFLICT OF INTEREST
The authors have nothing to disclose.
FUNDING
This research did not receive any specific grants from funding agencies in the public, commercial, or not-for-profit sectors.
