Continuous beat-to-beat non-invasive blood pressure recording was obtained with a Bmeye Nexfin® photoplethysmography system. The sensor cuff was located in the second phalanx of the middle finger of the right hand with a hydrostatic compensator in the little finger. The system was calibrated and contrasted with the data recorded using an OMROM M6 Comfort IT blood pressure monitor with PC connection (Figure 1).

ECG in DII (Cardioline-Delta 1 Plus®) was recorded with disposable adhesive button foam standard ECG Ag/AgCl electrodes. The recording of sympathetic cutaneous cholinergic activity was performed on the tip of the first phalanx of the ring and index fingers of the right hand with two non-polarizable Ag/AgCl electrodes (TSD203) through changes in skin conductance (GSR 100C amplifier) (Figure 1).

Signals were digitalized with the Biopac® MP160 CE multi modular system converter and analysed with the Acqknowledge 5.04 signal analysis software.

Mean heart rate, RRI, systolic blood pressure (SBP), diastolic blood pressure (DBP), mean blood pressure (MBP) and cutaneous conductance values were measured continuously during each 5 min.

In order to assess cardiac autonomic modulation, we performed Wavelet analysis using the Wigner-Ville transform (W-V) for time/frequency domain analysis with standard MatLab software. We determined sympathetic and parasympathetic modulation of RR intervals (RRI) in the low frequency (LF; 0.04–0.14 Hz) and high frequency (HF; 0.15–0.50 Hz) ranges (Task Force of the European Society of Cardiology the North A, 1996; Hilz and Dütsch, 2006; Li and Zheng, 2022). LF modulation of RRI reflects mainly sympathetic tone and, to an undetermined degree, also parasympathetic modulation; HF oscillations of RRIs reflect parasympathetic modulation (Task Force of the European Society of Cardiology the North A, 1996; Hilz and Dütsch, 2006; Li and Zheng, 2022). We also calculated the ratios between RRI modulations in the LF and HF ranges, using the LF/HF ratios as index of the sympathetic-parasympathetic balance (Task Force of the European Society of Cardiology the North A, 1996; Hilz and Dütsch, 2006; Li and Zheng, 2022). To reduce the parasympathetic component of LF modulation, we normalized LF and HF powers of RRI by calculating percentage values of LF and HF powers, with RRI LFnu = [LF/(LF + HF)] × 100%, and RRIHFnu = [HF/(LF + HF)] × 100% (Task Force of the European Society of Cardiology the North A, 1996; Hilz and Dütsch, 2006; Li and Zheng, 2022). The baroreceptor reflex sensitivity was assessed with the coherence-thresholder transfer function analysis method, computed from the Continuous Wavelet Transform (CWT) using the R package biwavelet.

Data recording and storage was obtained using 3 computers with Windows 10 operating system connected in a network.

The instrument of choice for musical study was a Yamaha P125 digital piano, and for performance was a Kawai CA-63 upright electric piano.

Throughout the process, volunteers were videotaped with a Logitech C-390 camera.

This study in no case can be considered as clinical experimentation in human beings. The ethical principles contained in the Declaration of Helsinki of the World Medical Association (2013) have been followed at all times. The volunteers granted a valid written informed consent and the protocols followed were approved by the Research Ethics Committee of the province of Málaga (Andalucía). Personal and clinical data are confidential and treated in accordance with the provisions of Regulation (EU) 2016/679 “General Data Protection” and Organic Law 3/2018, of December 5, on “Personal Data Protection.”

Data processing was carried out using statistical package GraphPad Prism 9. Descriptive analysis of continuous quantitative variables was performed, obtaining mean, standard deviation, minimum and maximum for each parameter.

For analytical statistics, data were grouped into five comparison groups according to the moment of the study (Initial rest, Warm-up, Pre-Concert rest, Concert, Post-Concert rest). The significance level was established at p < 0.05. Repeated measures of one-way ANOVA with the Geisser–Greenhouse correction were used. Tukey multiple comparisons test, was used as a post hoc test for multiple comparisons after ANOVA (adjusted p values <0.05 were considered significant). In all cases, graphical representations in the form of box plot have been used.

The mean heart rate during the “Initial rest” was 89.67 ± 4.1 bpm (from 70.56 bpm to 98.4 bpm), “Warm-up” 95.71 ± 3.56 bpm (from 85.2 to 107.11 bpm), “Pre-Concert rest” 90.75 ± 3.44 bpm (from 75.3 to 100.12 bpm), “Concert” 109.1 ± 3.85 bpm (from 94.27 to 120.91 bpm), and “Post-Concert rest” 90.20 ± 3.76 bpm (from 77.79 to 106.46 bpm).

The following significant changes in heart rate were obtained (Table 1):

The results show an increase in heart rate during the concert (Figure 2).

The mean cutaneous conductance during “Initial rest” was 8.875.10⁻² ± 8.10⁻⁴ μΩ, “Warm up” 8.833.10⁻² ± 6.2.10⁻⁴ μΩ, “Pre-Concert rest” 8.797.10⁻² ± 5.5.10⁻⁴ μΩ, “Concert” 9.341.10⁻² ± 4.1.10⁻³ μΩ, and “Post-Concert rest” 8.852.10⁻² ± 1.58.10⁻³ μΩ.

No significant differences were observed between groups.

However, conductance rose outside these predefined periods. This elevation coincides with the 30 s prior to the concert, and corresponds to the start of the metronome activity. We have named this period “Pre-Concert metronome.” The mean skin conductance values measured in this new period were the following: 1.684.10⁻¹ ± 2.18.10⁻² μΩ. Using these new data, the following significant differences appear (Table 4):

The results show an increase in sympathetic cholinergic sudomotor activity at the start of the metronome activity before the beginning of the concert.

The variability values for sympathetic activity in power normalized units (LFnu) were in “Initial rest” 79.00 ± 5.43 nu, “Warm-up” 80.43 ± 2.72 nu, “Pre-Concert rest” 79.44 ± 2.72 nu, “Concert” 88.35 ± 3.29 nu, and “Post-Concert rest” 77.86 ± 2.67 nu.

The following significant changes were obtained (Table 5):

The results show an increase in sympathetic activity (LF) during the concert.

The variability values for the parasympathetic activity in power normalized units (HFnu) were in “Initial rest” 21.00 ± 5.43 nu, “Warm-up” 23.77 ± 2.72 nu, “Pre-Concert rest” 21.13 ± 4.53, “Concert” 11.79 ± 3.29 nu, and “Post-Concert rest” 22.14 ± 2.67 nu.

The following significant changes were obtained (Table 6):

The results show a decrease in parasympathetic activity (HF) during the concert.

The variability values for the sympathetic/parasympathetic ratio (LF/HF) at “Initial rest” were 4.10 ± 1.34, “Warm up” 3.27 ± 0.52, “Pre-Concert rest” 4.08 ± 1.41 “Concert” 8.39 ± 3.26, and “Post-Concert rest” 3.59 ± 0.61.

The following significant changes were obtained (Table 7):

The results show that sympathetic activity was increased at rest. During the concert an increase of sympathetic activity (LFnu) with a decrease in parasympathetic activity (HFnu) is observed (Figure 4).

The vagal baroreceptor sensitivity values for LF at “Initial rest” were 8.83 ± 1.48 ms/mmHg (from 4.02 to 15.3 ms/mmHg), “Warm up” 6.77 ± 0.73 ms/mmHg (from 4.17 to 8.73 ms/mmHg), “Pre-Concert rest” 8.18 ± 1.25 ms/mmHg (from 4.09 to 13.41 ms/mmHg), “Concert” 5.65 ± 0.6 ms/mmHg (from 3.62 to 6.93 ms/mmHg), and “Post-Concert rest “7.9 ± 1.18 ms/mmHg (from 2.8 to 10.78 ms/mmHg).

No changes were observed in vagal baroreceptor sensitivity during the concert (Figure 5).

The sympathetic baroreceptor sensitivity values for LF at “Initial rest” were 9.17 ± 1.32 ms/mmHg (from 4.75 to 14.8 ms/mmHg), “Warm up” 6.79 ± 1.06 ms/mmHg (from 4.14 to 11.57 ms/mmHg), “Pre-Concert rest” 7.46 ± 1.15 ms/mmHg (from 4.68 to 12.52 ms/mmHg), “Concert” 4.64 ± 0.77 ms/mmHg (from 2.8 to 8.1 ms/mmHg), and “Post-Concert rest “7.91 ± 1.31 ms/mmHg (from 3.15 to 13.01 ms/mmHg).

The following significant changes were obtained (Table 8):

The results show a decrease in sympathetic baroreceptor sensitivity during concert. (Figure 6).

As previously expressed, we needed to select a work for piano with a non-operative right hand due to the presence of the recording sensors. Blood pressure was recorded by means of a pressure cuff placed on the second phalanx of the middle finger of the right hand. This circumstance was also used to record sympathetic skin conductance between the tips of the index and ring fingers of the right hand. We selected the “Prelude et Nocturne op. 9 no 1” written by Alexander Scriabin, who had chronic pain in the right hand and was therefore prone to writing works for the left hand (Altenmüller, 2015). There are other cases in which works are written for one hand with the aim of technically and expressively reinforce one of the two hands. In other cases, the objective is to be able to facilitate piano performance in situations in which a pianist loses the functionality of the hand due to an accident or a neuropathy (Boller and Bogousslavsky, 2015).

Initially, being all pianists right-handed, we supposed that a piece for the left hand could increase anxiety. However, one of the pillars of piano training is to develop ambidexterity in the performer, which is the ability to use both the right and left hand to play equally well, to tackle works that are part of the piano literature. The piece was chosen, in addition to leaving one hand free for data recording purposes, to be a work with technically undemanding interpretative characteristics for a professional pianist, so it should be performed with ease. Therefore, the impact of selecting the left hand for interpretation in the current study should have minimal effects. However, technically demanding pieces may be perceived with anxiety depending on the hand for which they were written. This is a question should be addressed in future studies.

The average heart rate increases during the concert and decreases again at final rest. No significant changes are found in warm up.

These changes in heart rate during musical performance are widely described in the literature (Yoshie et al., 2009a,b; Nakahara et al., 2010; Jha et al., 2022). Our results corroborate data previously published and reveal an increase in cardiac adrenergic sympathetic activity together with decrease in vagal activity that occurs during the performance of the piece in concert. The increase in heart rate is indicating the anxiety or stress to which the interpreter is subjected.

After the concert, at the final rest, cardiac sympathetic tone decreases and parasympathetic tone increases. During the Warm-up, we did not observe an increase in cardiac sympathetic or parasympathetic flow, which indicates that the interpreter is more relaxed.

To our knowledge this is the first study in which continuous beat-to-beat non-invasive blood pressure is recorded during a piano musical performance.

Diastolic and systolic blood pressures increase from rest to Warm-up, which is consistent with the need for increased blood flow to the structures involved in musical performance, the nervous and muscular systems.

During the concert, the increase in blood pressure is even greater, indicating an increase in the activity of the adrenergic sympathetic nervous system, probably secondary to the stress and anxiety of performing in front of an audience. This effect is observed both at the level of systolic pressure, which would indicate an increase in cardiac adrenergic sympathetic activity, and at the level of diastolic pressure, which indicates and increase in vascular tone and resistance due to an increase in peripheral adrenergic sympathetic activity.

After the concert, final rest, cardiac and peripheral sympathetic tone decreases to normal values.

The importance of this study lies in demonstrating for the first time what empirical knowledge pointed out during musical performance on blood pressure.

It is striking that there were no significant differences in sudomotor activity in the four periods studied. As in previous papers (Nakahara et al., 2010), we were expecting an increase in sudomotor activity during the performance of the concert as a result of anxiety. However, and although not significantly, sweating decreases during the concert. It would be interesting to know if this decrease in sweating is due to the concentration of the interpreter on the work that he is executing at that moment. The increase in cutaneous cholinergic sympathetic activity during the preconcert metronomic phase is very significant. Cholinergic sympathetic activity is not modified at any other time and could be a useful indicator to assess the level of anxiety to which the interpreter is subjected before the performance of the work.

Once it was verified that during the concert there was a significant increase of heart rate and blood pressure, we analysed which changes in autonomic flow were responsible for these responses. To achieve this, we used the Wigner-Ville analysis, a technique based on Wavelet’s analysis, which analyses frequency changes in the time domain.

In musical performance, the changes in the recorded variables depend to a great extent on the emotional involvement and the technical difficulty of the elements that constitute a musical work. These changes are sometimes so fast that their effects cannot be detected by conventional methods of heart rate variability analysis Fast Fourier Transform (FFT) or Regression Analysis (AR). It is necessary to use a different technique that works not only in the frequency domain but also in the domain of the temporality of the work (Jha et al., 2022).

The Wigner-Ville analysis evaluates frequency changes in the time domain, demonstrating greater sensitivity to detect fast changes in cardiac sympathetic and parasympathetic activity variability in very short periods of time like in music performance. This analysis has been previously used to evaluate the variability of heart rate in non-musically trained listeners as a response to emotions induced by music, resulting in a suitable method of study (Orini et al., 2010).

At rest, we observed high values of sympathetic flow (74,05%) associated with low parasympathetic activity (26,91%), thus the sympathetic/parasympathetic relationship was bigger than 1, showing that volunteer musicians presented a certain degree of stress induced by the experimental conditions.

The W-V analysis also shows that during the concert the increase in heart rate and blood pressure is produced by an increase in cardiac adrenergic sympathetic tone together with a decrease in cardiac parasympathetic activity. To allow this increase in blood pressure, baroreceptor sensibility, which relates blood pressure and heart rate changes, decreases, thus resetting the baroreceptor response and allowing simultaneous increases of blood pressure and heart rate. This allows for adequate blood perfusion to meet the demand of active organs.

The results also show the appearance of an increase in cardiac adrenergic sympathetic activity during warm up. This increase in sympathetic tone is due to an increase in the demand for blood flow to the structures involved in musical performance, mainly brain and muscles of the left limb. During the concert, cardiac sympathetic flow increases even more, while parasympathetic activity decreases, indicating an elevated state of anxiety or stress in the interpreter. After the concert, during the post-concert rest, both cardiac sympathetic and parasympathetic tone returned to Initial rest values.

Our results are consistent with those of other studies reported in the literature. These authors (Yoshie et al., 2009a,b) found, with piano performances in stressful context, significant variations in the following parameters: increased heart rate and a greater electromyography magnitude in the muscles of the upper limb: biceps brachialis, common digital flexor of the fingers, biceps brachii and upper trapezius. However, they found an increase in the sweating rate in the concert situation (Yoshie et al., 2009a,b).

We also reinforce the studies that demonstrate emotional involvement in musical performance. These works show an increase in the heart rate and the sweating rate just at the beginning of the execution of the work. These responses are due to an increase in the activity of the sympathetic adrenergic and cholinergic nervous system, in addition, they found a decrease in parasympathetic tone (Nakahara et al., 2010).

Our work has shown to be effective in the quantitative and qualitative study of neurovegetative activation during musical performance. Judging from our study and previous studies in the literature, it appears that the degree of autonomic activation correlates with MPA levels (Yoshie et al., 2009a,b; Nakahara et al., 2010; Jha et al., 2022). The methodology employed could be used to categorise MPA by better distinguishing those subjects where MPA generates pathological responses and offering appropriate treatments for them. For example, musicians with focal dystonia are known to have higher levels of anxiety than musicians without focal dystonia, regardless of whether they are musicians or not. Taking this into consideration, treatment could be optimized according to the autonomic activation they generate since, as mentioned above, MPA increases motor tone in the upper extremity (Spahn, 2015).

Currently, there is a range of treatment options that can be offered for MPA: psychoanalytic or psychodynamic therapy, cognitive behavioral therapy, multimodal therapy, and in more refractory or severe cases, pharmacotherapy (Spahn, 2015; Fernholz et al., 2019).

Multimodal therapy is shown to be an effective treatment approach for MPA. This therapeutic approach is customized to meet the unique needs of each patient and combines in-depth psychodynamic insight into the individual’s prior performance experiences and their management, with cognitive-behavioural therapy to promote desensitization. The latter involves gradually exposing the musician to progressively more challenging and stressful performance scenarios (Guyon et al., 2022). The aim is to promote a positive sense of self-awareness by addressing experiences associated with MPA proactively instead of resorting to anxiety avoidance techniques (Spahn et al., 2021b).

Multimodal therapy integrates biofeedback of neurovegetative parameters to support the treatment. The easiest to use are heart rate and respiratory rate. More difficult, but possible, is to use blood pressure, skin conductance or muscle tone recording. This therapy is supported by exercises with the instrument or voice, as well as relaxation techniques to avoid the motor and neurovegetative response derived from MPA (Spahn et al., 2021a).

Lastly, in refractory or severe cases, beta-blockers may be used to reduce the neurovegetative responses generated by MPA. Nevertheless, it is common to rely on beta-blockers as the primary therapeutic intervention (Spahn, 2015). The use of beta-blockers represents an abortive treatment approach to alleviate anxiety symptoms. Since it does not address the underlying psychological factors that contribute to the problem, dependence and abuse can occur in long term use, in situations when the musician is required to perform under external pressure, such as in a concert or exam. For this reason, in most cases, they are contraindicated (Spahn et al., 2021a).

Finally, the prevention of MPA is important for the performer to develop without the limitations that this condition supposes. To achieve this, teachers should promote awareness of performance-related anxiety to their students from childhood and adolescence onwards and provide them with the necessary tools to address it from multiple perspectives in a personalized way (Spahn, 2015).

The main limitation in this study is the reduced number of participants due to the difficulty of finding professional pianists as volunteers, however, the sample size is comparable to other published studies. Another limitation is that the sample comprises only young subjects. We think it could be interesting to extend this study to healthy musicians of different age ranges and performing experience. It is well-known that autonomic responses decline with aging (Takla et al., 2023). Another limitation of this study is that the subjective assessment of the anxiety experienced by the performers was not evaluated, since it would have allowed to verify a possible correspondence between physiological changes and the perceived anxiety. Catecholamine plasma measurements could be of interest to asses objective anxiety and to solve pending issues regarding the physiological mechanisms involved in our findings (Yamasaki et al., 2021).