Final Report Summary - MUSICMOVES (Let the music move you: involvement of motor networks of the brain in music processing.)
The neural processing of an auditory musical stimulus has characteristics that may prove to be beneficial to movement learning, with subsequent potential to be applied to a rehabilitation setting or to expert skill acquisition. Activation of the striatum and cerebellum, known to be involved in movement and specifically movement learning, is also reported for music (and specifically rhythm) perception [1], while motor area activations are shown to be modulated by rhythmic complexity [2]. Music imagery, demonstrated to have notable overlap with music perception in the EEG response [3], and reported to recruit similar motor networks in the brain [4], may also be useful in rehabilitation, as has been suggested by anecdotal reports from work with patients [5]. Replacing an auditory cue with auditory (music) imagination could represent a significant step forward in making rehabilitation setups more flexible and portable. However, it is currently unclear if such a mechanism is generalizable, and thus broadly useful in movement rehabilitation.
Acquiring a new motor skill with a healthy brain is not the same as rehabilitating basic skills after brain damage. However, proof-of-principle studies are required to make informed decisions in developing novel rehabilitation methods. Thaut et al [6] reported that the profiles of reaching movements showed reduced variability in timing and movement trajectories when performed to a cue. Ackerley et al [7] found that training wrist extension movements to a metronome (at a comfortable speed) results in greater use-dependent plasticity than self-paced (or more demandingly paced) movement, as measured via Motor Evoked Potentials after TMS stimulation. Interestingly, stimulating the cerebellum with TMS has been shown to increase plasticity in the sensorimotor cortex [8], which might provide an explanatory mechanism for the findings of increased plasticity with rhythmic cueing. While these studies together provide support for the idea that pacing facilitates motor learning, the cerebral activation pattern of moving to a cue (as compared to self-paced movement) is currently not known, and its effect on plasticity unclear. The extrapolation of these findings, namely that with more elaborate cuing (musical rhythms instead of a metronome) comes better motor performance, has been seen for sports situations (using endurance as the outcome measure, [9]), however in this study, no measurement of neural plasticity was taken.
The research findings discussed above led us to questions about the interaction of sound and movement when moving to a rhythm, as well as questions about the possible effect of this interaction on cerebral plasticity. Thus, two experiments were devised, in which first, functional brain activation could be investigated for movement to a metronome, compared with movement to heard or imagined music and second, brain plasticity could be measured related to the learning of a cued motor learning paradigm, contrasted with movement learning in silence. To maintain relevance for rehabilitation practice, the first experiment was based on an existing movement rehabilitation paradigm. Bilateral arm training with rhythmic auditory cues (BATRAC) is a cued upper-limb function rehabilitation protocol, in which bilateral arm movements are cued with a metronome, using a device that constrains movement direction [10]. The exact movements are sometimes varied [cf. 11], but the core assumption is that the bilateral aspect of moving either in- or out of phase with two hands exploits interhemispheric interactions, while the role of the auditory rhythm provided by the metronome cue is generally underemphasized. The possible extra effects of using natural music containing a pulse but also additional temporal structuring, has not been previously reported, while the cognitive aspects of cueing, especially salient in a condition where auditory imagery is the cue, has not yet been discussed. Previous reports of cued movement have led to inconsistent results, including decreased cortical activations, extra activations in cerebellum and putamen, and decreased thalamic activation [12]. To address these issues, we collected fMRI data while participants moved in silence, moved to a metronome, and listened to a metronome in a first data acquisition session, and then moved in silence, moved to music or moved to imagined music in a second session. Results from 17 non-musician volunteers indicated that both metronome cueing and music cueing result in increased activation in bilateral precuneus and posterior cingulate cortex, areas that are known at transmission hubs between motor and sensory areas [13], and the bilateral insula, known to be involved in sound-action associations [14]. However, no difference was found in the motor areas of the brain during the movement conditions, and importantly, the motor activation for simply listening to rhythms found in previous studies was not replicated. The imagery condition importantly did not show either insula or precuneus activation, but instead led to decreases in a range of temporal and parietal areas (including precuneus and insula), thought to be related to cognitive control mechanisms. Two papers describing these results are currently in preparation for submission to peer-reviewed journals.
For the second part of the project, the motor learning task, the BATRAC movements were not considered appropriate - a more complex motor task was required on which participants could show a learning curve. Here, a finger-to-thumb opposition sequence learning paradigm was developed that included bespoke music, composed specifically for the motor sequences to be learned. The training paradigm was implemented online so that participants could train at home with online training videos. Two groups of 15 participants are now being recruited to undergo movement training, using either silent videos, or videos with music. Participants train for 20 minutes, three times a week, for four weeks, resulting in 12 full training sessions. Changes in motor performance are measured using motion capture of finger touches as measured with a dataglove orginally developed as a gaming interface. Performance is assessed at training sessions 1, 7 and 12, and MR scanning takes place before session 1 and after session 12. The aspects of brain change being assessed comprise grey matter changes, white matter diffusivity changes (DTI) and functional connectivity (resting state fMRI).
The societal impact of these studies speaks mostly to current rehabilitation practice, specifically for upper-limb rehabilitation after stroke, but eventually also other types of rehabilitation and skill acquisition. Given the huge burden of upper-limb movement impairment after stroke, a better understanding of how to improve functioning represents an excellent opportunity for improved rehabilitation approaches. From the fMRI studies, we have already learned that the metronome cue, which is often used in practice, does not necessarily elicit the expected motor activation. Previous reports of motor activations for a pulse-based rhythm may be an effect of the very different experimental paradigms from which these finding emerge, and should be further investigated. Instead, the precuneus activation found here for cued movement suggests that audiomotor connectivity increases during sensorimotor synchronization. However, imagery-cued movement was found to lead to substantial decreases in brain activation, suggesting a possible facilitation of movement through the internal generation of temporal structure. These findings suggest a need for further research into the development of auditory and auditory imagery-based movement rehabilitation techniques. We anticipate that the results from the training study will further inform our knowledge of the mechanisms of music-based motor learning, and also provide a basis for further investigations of motor learning with auditory imagery. Ongoing links with patient groups and researchers are currently planned to further these aims, as an outcome of the current research.
References
1 - Grahn & Brett (2007). J. Cog. Neurosci. 19(5):893-906, Bengtsson et al. (2009). Cortex 45:61-72.
2 - Chen et al. (2008). Cer. Cortex 18, 2844-54.
3 - Schaefer et al. (2011a). Psych. Res. 75(2), 95–106; Schaefer et al. (2011b). Int. J. Psychophys. 82(3), 254-9
4 - Halpern (2001). Ann NY Ac Sci 930, 179-92
5 - Schauer & Mauritz (2003). Clinical Rehabilitation 17:713-22.
6 - Thaut et al. (2002). Neuropsychologia 40:1073-81
7 - Ackerley et al. (2011). Clin. Neurophys. 122:2462-68
8 - Popa et al. (2012). Cereb. Cortex 23(2), 305-14
9 - Karageorghis & Priest (2012). Int. Rev. Sports & Exercise Psych. 5:37–41.
10 - Whitall et al. (2000). Stroke 31, 2390–5
11 - Delden et al. (2009). BMC Neurology 9:57-71
12 - Schaal et al. (2004). Nat. Neurosci. 7:1137-45; Brown et al. (2006). Cereb. Cortex 16:1157–67; Toyomura et al (2012). Neurosci. Letters 516:39-44.
13 - Bullmore & Sporns (2009). Nat. Neurosci. 10:186-98
14 - Mutschler et al (2007). PloS One 2, e259
Acquiring a new motor skill with a healthy brain is not the same as rehabilitating basic skills after brain damage. However, proof-of-principle studies are required to make informed decisions in developing novel rehabilitation methods. Thaut et al [6] reported that the profiles of reaching movements showed reduced variability in timing and movement trajectories when performed to a cue. Ackerley et al [7] found that training wrist extension movements to a metronome (at a comfortable speed) results in greater use-dependent plasticity than self-paced (or more demandingly paced) movement, as measured via Motor Evoked Potentials after TMS stimulation. Interestingly, stimulating the cerebellum with TMS has been shown to increase plasticity in the sensorimotor cortex [8], which might provide an explanatory mechanism for the findings of increased plasticity with rhythmic cueing. While these studies together provide support for the idea that pacing facilitates motor learning, the cerebral activation pattern of moving to a cue (as compared to self-paced movement) is currently not known, and its effect on plasticity unclear. The extrapolation of these findings, namely that with more elaborate cuing (musical rhythms instead of a metronome) comes better motor performance, has been seen for sports situations (using endurance as the outcome measure, [9]), however in this study, no measurement of neural plasticity was taken.
The research findings discussed above led us to questions about the interaction of sound and movement when moving to a rhythm, as well as questions about the possible effect of this interaction on cerebral plasticity. Thus, two experiments were devised, in which first, functional brain activation could be investigated for movement to a metronome, compared with movement to heard or imagined music and second, brain plasticity could be measured related to the learning of a cued motor learning paradigm, contrasted with movement learning in silence. To maintain relevance for rehabilitation practice, the first experiment was based on an existing movement rehabilitation paradigm. Bilateral arm training with rhythmic auditory cues (BATRAC) is a cued upper-limb function rehabilitation protocol, in which bilateral arm movements are cued with a metronome, using a device that constrains movement direction [10]. The exact movements are sometimes varied [cf. 11], but the core assumption is that the bilateral aspect of moving either in- or out of phase with two hands exploits interhemispheric interactions, while the role of the auditory rhythm provided by the metronome cue is generally underemphasized. The possible extra effects of using natural music containing a pulse but also additional temporal structuring, has not been previously reported, while the cognitive aspects of cueing, especially salient in a condition where auditory imagery is the cue, has not yet been discussed. Previous reports of cued movement have led to inconsistent results, including decreased cortical activations, extra activations in cerebellum and putamen, and decreased thalamic activation [12]. To address these issues, we collected fMRI data while participants moved in silence, moved to a metronome, and listened to a metronome in a first data acquisition session, and then moved in silence, moved to music or moved to imagined music in a second session. Results from 17 non-musician volunteers indicated that both metronome cueing and music cueing result in increased activation in bilateral precuneus and posterior cingulate cortex, areas that are known at transmission hubs between motor and sensory areas [13], and the bilateral insula, known to be involved in sound-action associations [14]. However, no difference was found in the motor areas of the brain during the movement conditions, and importantly, the motor activation for simply listening to rhythms found in previous studies was not replicated. The imagery condition importantly did not show either insula or precuneus activation, but instead led to decreases in a range of temporal and parietal areas (including precuneus and insula), thought to be related to cognitive control mechanisms. Two papers describing these results are currently in preparation for submission to peer-reviewed journals.
For the second part of the project, the motor learning task, the BATRAC movements were not considered appropriate - a more complex motor task was required on which participants could show a learning curve. Here, a finger-to-thumb opposition sequence learning paradigm was developed that included bespoke music, composed specifically for the motor sequences to be learned. The training paradigm was implemented online so that participants could train at home with online training videos. Two groups of 15 participants are now being recruited to undergo movement training, using either silent videos, or videos with music. Participants train for 20 minutes, three times a week, for four weeks, resulting in 12 full training sessions. Changes in motor performance are measured using motion capture of finger touches as measured with a dataglove orginally developed as a gaming interface. Performance is assessed at training sessions 1, 7 and 12, and MR scanning takes place before session 1 and after session 12. The aspects of brain change being assessed comprise grey matter changes, white matter diffusivity changes (DTI) and functional connectivity (resting state fMRI).
The societal impact of these studies speaks mostly to current rehabilitation practice, specifically for upper-limb rehabilitation after stroke, but eventually also other types of rehabilitation and skill acquisition. Given the huge burden of upper-limb movement impairment after stroke, a better understanding of how to improve functioning represents an excellent opportunity for improved rehabilitation approaches. From the fMRI studies, we have already learned that the metronome cue, which is often used in practice, does not necessarily elicit the expected motor activation. Previous reports of motor activations for a pulse-based rhythm may be an effect of the very different experimental paradigms from which these finding emerge, and should be further investigated. Instead, the precuneus activation found here for cued movement suggests that audiomotor connectivity increases during sensorimotor synchronization. However, imagery-cued movement was found to lead to substantial decreases in brain activation, suggesting a possible facilitation of movement through the internal generation of temporal structure. These findings suggest a need for further research into the development of auditory and auditory imagery-based movement rehabilitation techniques. We anticipate that the results from the training study will further inform our knowledge of the mechanisms of music-based motor learning, and also provide a basis for further investigations of motor learning with auditory imagery. Ongoing links with patient groups and researchers are currently planned to further these aims, as an outcome of the current research.
References
1 - Grahn & Brett (2007). J. Cog. Neurosci. 19(5):893-906, Bengtsson et al. (2009). Cortex 45:61-72.
2 - Chen et al. (2008). Cer. Cortex 18, 2844-54.
3 - Schaefer et al. (2011a). Psych. Res. 75(2), 95–106; Schaefer et al. (2011b). Int. J. Psychophys. 82(3), 254-9
4 - Halpern (2001). Ann NY Ac Sci 930, 179-92
5 - Schauer & Mauritz (2003). Clinical Rehabilitation 17:713-22.
6 - Thaut et al. (2002). Neuropsychologia 40:1073-81
7 - Ackerley et al. (2011). Clin. Neurophys. 122:2462-68
8 - Popa et al. (2012). Cereb. Cortex 23(2), 305-14
9 - Karageorghis & Priest (2012). Int. Rev. Sports & Exercise Psych. 5:37–41.
10 - Whitall et al. (2000). Stroke 31, 2390–5
11 - Delden et al. (2009). BMC Neurology 9:57-71
12 - Schaal et al. (2004). Nat. Neurosci. 7:1137-45; Brown et al. (2006). Cereb. Cortex 16:1157–67; Toyomura et al (2012). Neurosci. Letters 516:39-44.
13 - Bullmore & Sporns (2009). Nat. Neurosci. 10:186-98
14 - Mutschler et al (2007). PloS One 2, e259