Research ReportA network for audio–motor coordination in skilled pianists and non-musicians
Introduction
From several points of view, high-level music performance is an interesting topic to study the neural underpinnings of action and perception and to address the question to what extent the human cerebral cortex is modified structurally and functionally due to training. First of all, playing a musical instrument poses great demands on the human motor system. Complex movement coordination is required at tremendous speed and accuracy. In addition, musicians are arguably challenged by no other profession in their expertise in the auditory domain. Music performance requires detection of minimal changes in pitch and rhythm to ensure a perfect artistic outcome. Thus, highly skilled musicians are an ideal model to investigate the function and plasticity of the auditory and the motor cortex (Munte et al., 2002, Schlaug, 2001). The combination of proficiency in the auditory and the motor modality, however, makes musicians particularly interesting for the studying of interaction and coordination between both modalities. The information flow between sensory processing and motor planning areas is crucial in music performance (Janata and Grafton, 2003). Fast feedforward and feedback connections are required to coordinate auditory input and motor output (Bangert and Altenmuller, 2003). Furthermore, these connections depend on cortical processing circuits which are capable of transforming auditory information into a code that is appropriate for use by the motor system and vice versa.
An impressive amount of data about visuo-motor transformation processes in humans and monkeys has been published (Caminiti et al., 1991, Ellermann et al., 1998, Grefkes et al., 2004, Rizzolatti et al., 1996), reviewed by Burnod et al. (1999). A preeminent structure consistently appearing in these studies is the premotor cortex. However, little is known about comparable audio–motor transformation centers so far (Hickok and Poeppel, 2000, Hickok et al., 2003). In the context of vocal performance, Hickok et al. (2003) carried out an interesting analysis to examine an audio–motor network. fMRI data were recorded during a passive listening task and during an active auditory rehearsal task in order to reveal cortical areas involved in perception and production of speech and singing. Only brain areas being active in both tasks were considered to be audio–motor integration areas. Hickok et al. (2003) identified bilateral premotor cortex (PM), bilateral inferior frontal gyrus (IFG) and an area in the posterior part of the Sylvian fissure in the left hemisphere as belonging to the above-mentioned audio–motor network. Apart from that, the left superior temporal sulcus (STS) supported the speech task while the music task recruited the contralateral STS.
Most of the previously published studies on instrumental music performance in professionals or amateurs focused either on music perception or instrument playing. Recent studies investigated brain activation on musical motor performance (Lotze et al., 2003, Meister et al., 2004). As expected, a wide range of primary, secondary (premotor cortex (PM), supplementary motor area (SMA)) and other motor and somatosensory areas (e.g., basal ganglia, cerebellum) were involved. Interestingly, Lotze et al. (2003) also demonstrated differences between amateurs and professionals in the primary and secondary auditory cortices. Although some studies had been set out to study combined auditory and motor control processes (Kristeva et al., 2003, Parsons et al., 2005), only very few of them focused on the interaction between both modalities. To our knowledge there is only one study which consequently addresses this issue (Bangert et al., 2001, Bangert and Altenmuller, 2003). In this study DC-EEG scalp maps recorded from piano novices during an audio–motor task become increasingly similar to the maps of professional pianists during the course of an audio–motor training. Haueisen and Knösche (2001) and Popescu et al. (2004) used a different approach to address audio–motor interaction. These two MEG studies observed activation in motor areas during pure auditory music stimulation in musicians (the former) or in non-musicians (the latter). The results imply either direct or indirect connectivity between auditory and motor areas.
All the aforementioned studies investigating audio–motor interactions are based on electrophysiological recordings. The high temporal resolution of these methods is an advantage for the discovery of fast interactions between the modalities of interest. However, imaging technologies like PET or fMRI benefit from higher spatial resolution. Increased spatial resolution may be crucial to distinguish between primary and secondary sensory and motor structures. Furthermore, precise localization information helps to decide whether involuntary motor activity in musicians is indeed related to purposeful finger movement (focus of activity in hand areas) rather than unspecific motor activity which is not constrained to areas usually activated by finger movements (e.g., the urge of Jazz listeners to tap the rhythm of the music with their limbs). The tendency to react to music and rhythmic sound by tapping, drumming or even dancing is known from many cultures. However, to our knowledge there is no systematic neurophysiological study on involuntary motor activity induced by music which is not specific to a certain body part.
The value of studying music and in particular highly trained musicians as a model for cortical plasticity has been increasingly recognized in the last decade. Several anatomical and functional studies examined the effect of plastic changes due to intensive musical training. Increased grey matter volume for musicians was found in the anterior corpus callosum (Schlaug et al., 1995), the cerebellum (Hutchinson et al., 2003, Schlaug, 2001) and in primary and secondary somatosensory and motor areas (Gaser and Schlaug, 2003). All these brain regions play an important role in either fine motor control or bimanual information transfer, both vital processes for music performance. A seminal MEG study (Elbert et al., 1995) showed that fingers of the left hand of violinists show stronger representation by means of signal amplitude in the primary somato-sensory cortex than those of a control group. The same method was used to reveal increased auditory cortical representations for musical timbre in violinists and trumpeters relative to non-musicians (Pantev et al., 2001). fMRI studies, however, showed rather a decrease of intensity of activation in motor areas in musicians vs. non-musicians. Further studies observed weaker activity in primary and secondary motor areas (primary motor area (M1), supplementary motor area (SMA), preSMA and cingulate motor area (CMA)) for pianists compared to non-musicians in either bimanual (Jancke et al., 2000b) or unimanual piano playing like tapping tasks (Hund-Georgiadis and von Cramon, 1999). Even the comparison of professional vs. amateur violinists during the performance of a Mozart concerto by Lotze et al. (2003) revealed mainly stronger activity in motor areas for the amateurs. The lower activity in cortical motor control areas has commonly been attributed to a diminished neural effort needed to perform a particular motor task. Finally, Stewart et al., 2003a, Stewart et al., 2003b, Stewart et al., 2004) showed changes in activity of cortical areas in the parietal cortex after laymen learned to read and play music. The observed activity highlighted the involvement of a well known visuomotor transformation area in the translation of visually perceived musical notes into a motor program for key presses.
The functional imaging studies outlined in this section highlight the potential of music training to modify human brain functions. However, the exact neuronal mechanisms of these functional alterations remain elusive for methods which are currently used for the investigation of whole brain activity. Therefore we work with a relatively broad definition of functional plasticity in this manuscript. We define functional plasticity as any kind of altered neuronal activity evoked by an invariable stimulus or task due to a history of training or experience. Although this definition does not distinguish between modified bottom-up or top-down processes, we will nevertheless make an effort to keep the influence of both processes on plasticity apart by the introduction of attention guiding tasks in the present study.
In contrast to previously accomplished studies, the main interest of the present study is crossmodal activity, i.e., responses in the motor cortex elicited by auditory stimulation and activity in the auditory cortex triggered by piano (finger) movements. We hypothesize that these transmodal1 activities are stronger in pianists compared to non-musicians. This should especially be evident for music stimuli, an assumption based on the idea that musicians might have stronger connections between auditory and motor areas which result in an efficient translation of finger actions into auditory music representations and vice versa.
Another outstanding question is whether transmodal information transfer occurs completely involuntary or by the use of voluntary imagery of the missing modality. The aforementioned studies by Haueisen and Knösche (2001) and Popescu et al. (2004) imply that information transfer from the auditory to the motor system has at least an involuntary component. However, various studies demonstrated that imagery of sound is sufficient to activate areas in the auditory cortex (Bunzeck et al., 2005, Halpern and Zatorre, 1999, Janata, 2001, Zatorre et al., 1996) and that imagery of music performance leads to activation in motor areas (Kristeva et al., 2003, Lotze et al., 2003, Meister et al., 2004).
The aim of our study is to provide a detailed description of cortical and subcortical areas involved in an audio–motor integration network using piano performance as an example. In order to find areas communicating with each other during piano performance, we specifically search for areas which are coincidentally active during both pure motor and pure auditory tasks. Furthermore, we investigate to what extent this network is reliant on voluntary, top-down control and to what extent decades of training (as normally pertaining to musicians) change the activity pattern of the network. According to these aims we propose three hypotheses: (1) Piano performance evokes strong activity in a network of areas integrating information from the auditory and the motor modality. (2) Hearing musical pieces will automatically (implicitly) evoke activations in motor areas of musicians. Vice versa, playing musical pieces without hearing the tune will automatically evoke activations in auditory areas. (3) Pianists have stronger transmodal representation for music stimuli compared to non-musicians. In particular, we expect a much more focused activity pattern areas involved in finger movements in pianists compared to an eventually more unspecific and distributed activity of the motor cortex in non-musician controls.
Section snippets
Executed finger movements vs. OFF
In general, we observed movement-related activations in the group analyses that are in line with previous studies which have mapped musical instrument performance (Lotze et al., 2003, Meister et al., 2004, Parsons et al., 2005). Peak activations were observed in the primary motor and somatosensory cortices with the maximum at x = 40 = − 38 z = 54 (MNI coordinates). Further activations comprised bilateral secondary sensorimotor areas such as PMd, preSMA/SMA, inferior parietal lobule and motor-related
Discussion
The presented data offer a set of presumably interconnected cortical areas engaged in both piano playing and piano listening. These cortical areas are ideal candidates for audio–motor integration and transformation circuits mediating information between the two modalities. Our discussion first considers the concurrently activated auditory and motor areas in pianists and non-musicians which appear in both VOLUNTARY and INVOLUNTARY modes. Then we discuss the role of voluntary involvement in
Subjects
Seven highly skilled pianists from the Zurich Conservatory of Music (6 female; mean age: 25.7 ± 3.2 years) and 7 control subjects (3 female; mean age: 30.0 ± 5.0) took part in the study. The pianists received at least 10 years of extensive piano training (age of commencement: 5–13) and the time of exercise comes to 3–5 h per day. Control subjects were matched according to the education level of the musicians. They did not receive any special musical training in their past life. All subjects were
Acknowledgments
During the preparation of the manuscript, the authors were supported by the Swiss National Science Foundation (SNF) SNF 46234101 (S.B. and M.M.) and by the SNF priority project NCCR/TP5 “Neural plasticity and repair” (S.K.).
References (94)
- et al.
Shared networks for auditory and motor processing in professional pianists: evidence from fMRI conjunction
NeuroImage
(2006) - et al.
Parametrically dissociating speech and nonspeech perception in the brain using fMRI
Brain Lang.
(2001) - et al.
The song system of the human brain
Cogn. Brain Res.
(2004) - et al.
Neural circuits underlying imitation learning of hand actions: an event-related fMRI study
Neuron
(2004) - et al.
Scanning silence: mental imagery of complex sounds
NeuroImage
(2005) - et al.
Song and speech: brain regions involved with perception and covert production
NeuroImage
(2006) - et al.
Interactions between auditory and dorsal premotor cortex during synchronization to musical rhythms
NeuroImage
(2006) - et al.
The preparation and execution of self-initiated and externally-triggered movement: a study of event-related fMRI
NeuroImage
(2002) - et al.
Reduction of gradient acoustic noise in MRI using SENSE-EPI
NeuroImage
(2002) - et al.
Activation of visuomotor systems during visually guided movements: a functional MRI study
J. Magn. Reson.
(1998)