1. Introduction
Musical performance is a very complex human capability and requires a broad variety of skills, including precise instrument/vocal control and technique with accuracy of notes, rhythm and phrasing as well as interpretational skills such as appropriate tempo and dynamic, suitable sense of style and involvement in the music [
1].
For the assessment of musical ability, several well-designed musical perceptual measurements such as the Seashore test [
2], the Intermediate Measures of Musical Audiation [
3], the Advanced Measures of Musical Audiation [
4], the Montreal Battery of Evaluation of Amusia [
5] and the more recently developed Profile of Music Perception Skills [
6] are available. In addition, there are self-report questionnaire inventories such as the Goldsmiths Musical Sophistication Index (Gold-MSI), which measures musical sophistication and consists of five factors: active engagement, perceptual abilities, musical training, singing ability and emotions [
7]. However, to date, there are only few musical performance measures which focus either on the reproduction of rhythmical and melodic sequences [
8] or on performing familiar or unfamiliar songs [
9,
10].
Until now, research mainly concentrated on perceptual musical ability tests. Music perception measures mainly use objective measures with correct or incorrect assessment options. As an acoustic analysis of music performance relies on accuracy, it has the advantage that findings are reproducible [
11,
12]. However, computerized methods can reach their limits. For instance, when individuals play musical pieces or perform songs technically perfect but in an inaccurate pitch, they could be evaluated poorly, even though the performance may be quite good [
10,
13]. In contrast, music performance assessments are often based on very time-consuming approaches [
14] with rating scales based on certain criteria chosen by experts in the field [
15,
16,
17,
18,
19]. These rating scales may be used in flexible ways and can therefore be adapted according to specific rating criteria [
11], which has also the advantage that longer sequences can be assessed [
14]. Even though rating scales are subjective, research found a high correlation between acoustic and subjective measures of musical performance [
11]. The increased reliability of measures based on rating scales can be achieved by using more than one rater—an approach we decided to use in this investigation and which has been applied previously [
9,
17,
18,
20]. As our cohort included individuals with dyslexia, ADD and ADHD, we decided to use rating scales, since piloting has shown that individuals with diagnoses more frequently sang parts of the musical pieces out of tune.
So far, there are contradictory findings regarding the link between music perception and performance [
21,
22]. Some researchers noted a relationship between music perception and production [
23,
24], while others could not detect a relationship between both [
25]. This dissociation between the perception and production of musical stimuli gained increasing interest in impairment studies. These studies assume that if one capacity is impaired, the other could possibly be spared. In this respect, alternative explanations for deficits were put forward. For instance, while it is generally accepted that amusics’ poor singing ability stems from poor pitch perception deficits, recent research found evidence that amusics’ poor singing ability can be explained by the inability to control sensorimotor translations [
26,
27]. Conduction aphasia, which leads to spontaneous speech production impairment, is understood as a deficit in sensory–motor integration [
28], and stuttering improves alongside gaining sensory–motor control of the vocal motor apparatus [
29]. Sensorimotor synchronization is a crucial aspect of referential behavior and describes the rhythmic coordination of perception and action, which is also a fundamental aspect required in musical activities [
30]. Sensorimotor synchronization is a crucial aspect of referential behavior and describes the rhythmic coordination of perception and action, which is also a fundamental aspect required in musical activities [
30]. So far, sensorimotor skills have mainly been assessed by the finger tapping paradigm, which measures the synchrony between the tapping of the index finger and the pacing stimuli [
31]. More recently developed test batteries, such as the
Assessment of Auditory Sensorimotor and Timing Abilities (BAASTA) [
31] and the
Harvard Beat Assessment Test (H-BAT) [
32], provide information about perceptual and sensorimotor timing ability. There is evidence that musical training enhances sensorimotor synchronization [
32], and musicians show more elaborate synchronization skills, lower tapping variability and greater perceptual sensitivity compared to controls [
33].
As musical performance requires the integration of multimodal sensory and motor information, professional musicians not only demonstrate enlargements in the motor cortex but also neuroplastic changes at a cellular level [
34]. The auditory cortex is widely linked to various brain regions, including prefrontal and parietal regions, and is involved in complex auditory and non-auditory functions, such as spectral and holistic listening modes [
35], absolute and relative pitch [
36,
37], sensorimotor [
38,
39,
40], cognitive [
41] and language-related [
42,
43,
44] functions.
Neurophysiological research suggests that the neural processing of language and music may be shared since acoustic signals of speech show similarities to music in temporal and spectral complexity [
45,
46]. This may be one fundamental reason why individuals with diagnosed neurodevelopmental disorders show deficits in both music and language processing [
47,
48]. Models such as the “OPERA” hypothesis postulate that benefits in speech processing induced by musical training are based on five conditions: overlap, precision, emotion, repetition and attention [
49]. The OPERA hypothesis mainly focusses on perceptual parameters. Other models such as the Precise Auditory Timing Hypothesis (PATH) suggest that auditory–motor entrainment and phonological awareness both depend on the same mechanisms: neural timing and its integration into motor and cognitive networks [
50]. Therefore, it can be postulated that musical training with emphasis on entrainment also trains phonological skills [
50]. The
Processing Rhythm in Speech and Music (PRISM) framework defines precise auditory timing, the synchronization/entrainment of neural oscillations to external rhythmic stimuli and sensorimotor coupling as the three common mechanisms which underly music and speech rhythm processing [
47]. The PRISM model has not only been introduced to show overlaps between music and speech perception and production but also provides a framework for developmental speech disorders. This framework unites auditory processing, crucial for the detection of timing deviations, the synchronization and entrainment of neural oscillations and sensorimotor coupling, which links perception to production [
47].
There is a growing body of evidence that the anatomy and function of the auditory cortex is altered in neurodevelopmental disorders such as dyslexia, attention deficit hyperactivity disorder (ADHD) and attention deficit disorder without hyperactivity (ADD) [
51,
52,
53,
54,
55]. Dyslexia and AD(H)D belong to the most common neurodevelopmental disorders in children and adolescents, with a worldwide prevalence of about 5–10% [
56,
57], and show a high level of comorbidity [
58,
59,
60]. Dyslexia is a specific learning disability characterized by difficulties with accurate and/or fluent word recognition and by poor spelling and decoding abilities. A poor discrimination of basic sound features and sequential acoustic patterns may lead to suboptimal speech representation, constraining the development of phonological representations [
61] and reading and spelling skills [
62]. Individuals with dyslexia not only have timing difficulties in language and music perception, performance and lack motor control [
63,
64,
65,
66] but also a large variety of auditory deficits, ranging from basic to more complex auditory processing deficits [
51,
62,
65,
67,
68,
69]. In addition, they have impairments in higher-order cognitive processing (e.g., executive functions) and cognitive skills (e.g., cognitive flexibility) [
70,
71,
72]. Children with developmental disorders have been found to exhibit underlying timing deficits which were not only seen as predictors for the disorders [
70,
72] but also as triggers [
67].
AD(H)D is characterized by the key symptoms of hyperactivity, impulsivity and/or inattention. According to the International Statistical Classification of Diseases German Modification [
73], two subtypes (namely ADHD and ADD) are distinguished. Patients affected by AD(H)D show broad deficits including motor deficits, sensorimotor integration impairments, perceptual timing deficits, temporal foresight and rhythm-related deficits such as the poor differentiation of temporal auditory parameters and the desynchronization of temporal patterns [
74,
75,
76,
77,
78,
79,
80,
81,
82]. Moreover, difficulties in hearing and understanding oral instructions [
83] and a lack of the ability to move to a beat and detect deviations from a beat [
84] can be found.
There is scarce literature dealing with characterizing features of musical ability in AD(H)D subtypes/ presentations. Noreika and colleagues demonstrated that perceptual timing and temporal foresight is less impaired in ADD than in ADHD [
77]. Children with ADHD show higher-order auditory processing deficits, including impairments in perception of rhythm and melody, and children with ADD demonstrate no auditory impairment at all [
51].
In previous studies, we observed auditory neurofunctional anomalies in children with dyslexia, ADHD and ADD. While group-averaged P1 source waveform responses were well-balanced in controls, the disorder groups showed a pronounced P1-asynchrony [
51,
53]. While dyslexics showed impairments in elementary (e.g., frequency, tone onset and duration) and complex auditory sound discrimination (meter, rhythm, melody, harmonic sound perception and phoneme discrimination), children with ADHD only performed worse in sequential auditory pattern recognition. In contrast, there were no auditory deficits in children with ADD. Musical training in children with dyslexia, ADHD and ADD lead to a markedly diminished asynchrony of the primary auditory answers [
51].
To our best knowledge, to date, there is no research focusing on musical performance in neurodevelopmental disorders such as ADHD, ADD and dyslexia. Therefore, we wanted to close this research gap.
Hence, the goal of this study was to (a) evaluate the group-specific characteristics of musical performance in adolescents with dyslexia, ADHD and ADD, and (b) to investigate whether the potentially found differences in performance can be correlated to the response pattern of the auditory cortex as measured using magnetoencephalography (MEG).
Due to the abovementioned auditory impairments, we hypothesized that the disorder group would perform worse in the musical performance assessment scale than the control group. Within the disorder group, we assumed that adolescents with ADD would perform better in the musical performance than adolescents with ADHD or dyslexia, since the latter both showed auditory impairments in previous research. Additionally, we wanted to uncover whether our groups could also be differentiated based on the response pattern latencies of the auditory cortex as measured using magnetoencephalography (MEG). Based on the statistical analysis, we wanted to analyze whether the musical performance and MEG variables which discriminate our groups best are also correlated with each other.
4. Discussion
The considerable worldwide prevalence of ADHD, ADD and dyslexia (5–10%) and the known benefits of musical training on neuronal processing and behavior [
51,
53] highlight the importance of gaining a better insight into and understanding of auditory processing in order to optimize musical education and to develop new pedagogic interventions for children/adolescents with developmental and learning disorders. Therefore, this study aimed to evaluate possible characteristic differences in music performance and auditory-evoked field variables in adolescents with dyslexia, ADHD and ADD. In addition, we sought to uncover potential correlations between musical performance and MEG response patterns. Since our previous study focused on music perception and linked atypical neurofunctional patterns to individual differences in music perception [
51], we now aimed to go beyond music perception and address musical capacities from the perspective of musical performance and used musical performance assessment measurements based on already established test designs [
99] and analysis procedures of previous research [
14]. Through this, we could show that, compared to controls, dyslexic children/adolescents score lower in basic music-hearing tasks (frequency and onset ramp discrimination) and complex sound-processing tasks (meter, rhythm, and melody differentiation), and that children/adolescents with ADHD score lower in complex rhythmic and melodic perception tasks [
51]. In contrast, children/adolescents with ADD did not show any auditory impairments at all [
51].
In our current study, musical performance differed significantly across groups. In general, the control, ADD and ADHD groups scored higher than the dyslexic participants in almost all measures of musical performance, except for the rhythmic and pitch memorization task, in which all groups scored similarly. Since rhythmic and pitch memorization are based on memorizing melodic and rhythmic phrases, it could be assumed that these measures reflect not only musical performance mechanisms but also require short-term memory ability. The reason why we could not detect mean differences could be attributed to the fact that tasks were not long enough in order to uncover individual differences.
For the interpretation of our results, we mainly relied on discriminant analysis, which provided information about which of our variables separate our participants. The discriminant analysis of the musical performance measures revealed that rhythmic improvisation and musical expression discriminated the groups best. In the following, we discuss these variables and the underlying concepts in more detail. We assumed that compared to the control and ADD groups, the dyslexic participants would perform worse in rhythmic improvisation and musical expression. However, we did not expect that the ADHD group would perform better than the dyslexic group, since in previous studies, we noted music perception deficits in both the ADHD and the dyslexic participants. Even though rhythm-related and musical perception deficits have been reported in individuals with ADHD [
51,
81,
84], in our current study, adolescents with ADHD and ADD scored similarly to controls in rhythmic improvisation and musical expression. Subsequently, as results in music performance may differ from results in music perception, research findings in music performance should not be transferred to music perception and vice versa.
In former investigations, we noted that individuals with dyslexia suffer from severe auditory deficits compared to children/adolescents with ADHD, ADD and controls [
51]. In our current study, individuals with dyslexia performed worse in rhythmic improvisation than adolescents with ADHD and ADD and the controls. The ability to encode incoming temporal information is not only crucial for musical but also for phonological processing. Goswami [
67] postulates that auditory rhythmic entrainment is impaired if individuals have specific difficulties with Theta and Delta oscillators. This auditory entrainment not only affects attentional but also auditory integration. The phonological impairments of individuals with dyslexia can therefore be understood as auditory sensory impairments. This supports assumptions of language disorder frameworks such as the temporal sampling framework (TSF) [
67] and the PRISM [
47], which suggest that timing difficulties of individuals with dyslexia may be caused by auditory sensory integration impairment of incoming acoustic signals. Since the PRISM is based on shared mechanisms of language and speech, it is also applicable to musical performance. In the light of the present findings, it is plausible to assume that the rhythmic impairment of individuals diagnosed with dyslexia affects the musical and language domain in a similar way.
An indirect aspect, namely creativity, could be a further reason why subjects with ADHD and ADD perform better than adolescents with dyslexia. Musical improvisation and expression are defined by the ability to perform music in a creative and spontaneous way [
109]. As ADHD symptoms are associated with more flexible association networks [
110] and better creative performance [
111,
112,
113,
114], one could postulate that both aspects combined could serve as an explanation as to why adolescents with ADHD and ADD score higher in rhythmic improvisation and musical expression than dyslexics.
In contrast, dyslexics have not been found to be more creative or show greater variability in creativity than peers without dyslexia [
115]. It is known that due to a variety of basic auditory deficits [
68,
69,
116], dyslexics show impaired development of language abilities such as the acquisition of phonological representations, literacy skills [
53,
62,
117] and the perception of metrical structure in music [
65]. Additionally, dyslexics are impaired in controlling brief temporal components of acoustic spectra in their motor output [
118,
119] and in anticipating and maintaining the beat in rhythmic entrainment tasks [
120,
121]. These temporal impairments may lead to the abovementioned difficulties in musical and rhythmical perception and production. The discriminant analysis of auditory-evoked fields revealed that the first function distinguished the control group from the disorder groups based on P1 latency asynchrony |R-L|, which is in line with previous research [
51,
53]. The second function of P1 and N1 latencies (mean) distinguished the control, ADD and ADHD groups from the dyslexic group. We then, correlated the music performance variables rhythmic improvisation and musical expression, which discriminated the dyslexic group from all other groups, with all MEG variables. Correlational analyses on musical performance and auditory-evoked fields revealed a relationship between rhythmic improvisation and P1 and N1 latencies (mean). The primary P1 component is thought to be a marker for musical talent [
53,
104] and can already be measured in early childhood [
122]. The N1 response usually emerges later, at about 8–10 years of age [
106,
122]. The N1 component, which reflects sensory stimuli processing, is linked to attention-specific processes [
123] and shows a strong context dependency and learning-induced plasticity [
37]. The later N1 latency (mean) and the weaker rhythmic improvisation performance of adolescents with dyslexia not only could be understood as a perceptual impairment, but also as a deficit in sensorimotor motor translations, which makes individuals insensitive to accurately reproducing musical input. Hence, it is crucial to consider that sensory processing influences efficiency in motor output [
82].
As a sign of natural development and maturity, the latencies of the primary P1 and secondary N1 response component accelerate up to the age of 15 years [
105,
122,
124]. ADHD, ADD and dyslexia are characterized by specific neuroanatomical and neurofunctional differences in the auditory cortex. The MEG source waveforms of children/adolescents with ADD are known to be shifted in latency but balanced in shape, while the response patterns of children/adolescents with ADHD were temporally expanded in the left and diminished in the right hemisphere, and in the dyslexic group, the P1 peak was enlarged. Further, all disorder groups showed a higher P1 latency asynchrony |R-L| [
51,
53]. The P1 latency asynchrony |R-L|, which indicates a shift in latency, differentiated the control group from all disorder groups best in this investigation. This asynchrony corresponds to a reduced integration of left hemispheric fine-grained and right hemispheric supra-segmental signal representations, which lead to difficulties in discriminating onsets of syllables and perceiving rhythmic structures in speech and music. These difficulties are characteristic for children with dyslexia [
116,
125] and are frequently associated with AD(H)D [
126]. There is evidence that children and adolescences with AD(H)D demonstrate an atypical development of the N1 component with growing latency over time, whereas non-affected individuals are characterized by a declining latency [
127]. It seems possible that by means of attentional training, adults with AD(H)D develop compensatory mechanisms as a part of maturity and cognitive enhancement [
128,
129]. Compared with normal average readers, dyslexic children exhibit prolonged latencies of auditory-evoked potentials, possibly reflecting disturbances in written language acquisition [
130,
131]. The negative correlation between the rhythmic improvisation and P1 and N1 latencies in our study implies that the earlier the P1 and N1 latencies (mean), the better the rhythmic improvisation.
Neurophysiological studies in musicians have shown brain plasticity induced by musical training, such as enhanced activation in the auditory cortex [
132,
133,
134], more pronounced structural and functional connectivity [
34,
135,
136,
137,
138] and intracerebral synchronization [
139,
140]. Musical training is known to positively affect the accuracy of auditory perception [
53,
141,
142,
143,
144,
145], language development [
146,
147,
148,
149,
150,
151,
152,
153] and motor functions [
38,
154,
155]. There are strong links between rhythmic and linguistic abilities [
156,
157,
158,
159,
160,
161,
162,
163]. Additionally, making music is associated with beneficial influences in general cognitive and executive functions such as planning, self-control, working memory [
164,
165,
166] and the conscious control of attention [
167,
168]. In particular, frequent musical performance rehearsals optimize and strengthen neuronal interconnection by changing the timing and synchronization as well as the number and strength of stimulating and inhibiting synaptic connections and postsynaptic potentials [
34,
169,
170,
171,
172,
173,
174,
175,
176,
177,
178,
179].
Patients with ADHD and ADD benefit from music therapy using improvisational musical input, as it has been shown to improve emotional lability, psychosomatic symptoms and attention [
180,
181,
182]. The additional advantages of music-based training programs are their motivating, playful approach, the possibility of speech-free interaction and the use of resources such as the joy of movement, creativity and openness [
183,
184] that often characterize children with ADHD [
185]. Dyslexic children could benefit from music and especially rhythmic training, leading to improved brain circuitry for music and language processes. In addition, the temporal and rhythmical features of music could positively affect temporal processing deficits [
64,
186,
187].
Further studies should be based on larger numbers of participants and focus on the role of the P1 and N1 latencies in ADHD, ADD and dyslexia and how they can be influenced by specific musical training. Research outlined that the synchronization and balancing of right and left auditory responses increases musical practice in controls and adolescents with dyslexia, ADHD and ADD [
51,
53]. As balanced and reduced latencies are correlated with more efficient and enhanced auditory processing and attention and literacy skills, it can be assumed that the shorter the latency, the faster and more precise the auditory processing [
51,
53]. In this investigation, we could also detect that P1 and N1 latencies showed a negative correlation to rhythmic improvisation in music performance, which suggests that enhanced auditory processing can probably also predict individual differences in sensorimotor timing ability. In addition, future studies should shed light on how P1 and N1 latencies can be reduced by instructional musical input, which would enrich therapeutic methods for individuals with ADHD, ADD and dyslexia.
In conclusion, our data provide novel insight into differences of music processing and performance in adolescents with and without neurodevelopmental disorders. A better understanding of these distinct differences in musical performance and underlying neurobiological factors may help to develop tailored preventions or interventions for individuals with ADHD, ADD and dyslexia. These should include sensory–motor training and the training of fine-grained auditory skills such as pitch, timing and timbre perception tasks.