A Protective Effect of Musical Expertise on Cognitive Outcome Following Brain Damage?

Neuropsychol Rev (2014) 24:445 460 DOI 10.1007/s11065-014-9274-5 REVIEW A Protective Effect of Musical Expertise on Cognitive Outcome Following Brain Damage? Diana Omigie & Severine Samson Received: 11 July 2014 /Accepted: 22 October 2014 /Published online: 8 November 2014 # Springer Science+Business Media New York 2014 Abstract The current review examines the possibility that training-related changes that take place in the brains of musicians may have a beneficial effect on their cognitive outcome and recovery following neurological damage. First, we propose three different mechanisms by which training-related brain changes might result in relatively preserved function in musicians as compared to non-musicians with cerebral lesions. Next, we review the neuropsychological literature examining musical ability in professional musicians following brain damage, specifically of vascular, tumoral and epileptic aetiology. Finally, given that assessment of musician patients can greatly inform our understanding of the influence of premorbid experience on postmorbid recovery, we suggest D. Omigie: S. Samson Laboratoire de Neurosciences Fonctionnelles et Pathologies, EA4559, Université de Lille, Villeneuve d Ascq, France S. Samson Unité d Epilepsie, GHU Pitié-Salpêtrière, Paris, France D. Omigie Centre de Recherche de l Institut du Cerveau et de la Moelle Epinière (CRICM), UMR_S 975, and Centre MEG-EEG - CENIR, Université Pierre et Marie Curie-Paris 6, Paris, France D. Omigie CNRS, UMR 7225, CRICM and Centre MEG-EEG, Paris, France D. Omigie Inserm, U 975, CRICM and Centre MEG-EEG, Paris, France D. Omigie (*) CRICM, UMR7225 / U975, CNRS / UPMC / Inserm, Institut du Cerveau et de la Moëlle Épinière (ICM), GHU Pitié-Salpêtrière, 47 Boulevard de l Hopital, Paris, France 75013 e-mail: d.omigie@gmail.com D. Omigie Music Department, Max Planck Institute for Empirical Aesthetics, Frankfurt am Main, Germany some basic guidelines for the future evaluation of relevant patients. Keywords Musicians. Brain damage. Neuroplasticity. Lesion. Training. Cognitive outcome Introduction In the general population, music listening constitutes a ubiquitous activity that fulfils a range of functions, from passing the time to managing a listener s mood(denora 2000). However, a subset of the population also undergo various forms of intense training that allow them to master complex musical activities such as singing, skillfully playing a musical instrument, composing musical pieces or directing an orchestra. Such musical training necessarily involves the simultaneous recruitment of a diverse range of neural processes including those involved in perception and cognition and those concerned with motor planning and execution. Consequently, it may be hypothesised that individuals who, through intense practice, have mastered their musical skills to a professional level, should differ from those who have never undergone such intense training, and thus lack such expertise. In the words of Santiago Ramon y Cajal, Every man, if he so desires, can become the sculptor of his own brain. The highly influential histologist, often referred to as the father of modern neuroscience, argued that the many hours of mental and muscular gymnastics that musicians undergo during their training brings about neuroanatomical changes that render them different to their non-musician counterparts (Ramon y Cajal 1904/1999). A century later, this proposal is supported by numerous studies that exploit advancements in anatomical scanning and image processing techniques to show that the structure of the musician brain does indeed differ from that of non-musicians. Thus, adding to the ever growing literature

446 Neuropsychol Rev (2014) 24:445 460 demonstrating structural and functional neurological differences in various populations as a result of their continued engagement in a specific, intense and complex activity (Maguire et al. 2000; Draganski et al. 2004; Draganski and May 2008; Janckeet al. 2009), musicians provide an excellent example of how the brain changes with training and an invaluable model of cerebral plasticity (Driemeyer et al. 2008; Hyde et al. 2009; see Herholz and Zatorre 2012 for review). In the current review, we examine the possibility that musical training might not only bring about structural and functional changes in the brain of healthy individuals, but might also have an influence on their cognitive outcome and recovery following cerebral damage. The existing literature tends to focus on cognitive outcomes of musical functions in musicians. However, despite the lack of evidence in the literature as it currently stands, one prediction that could be made is that these structural and functional changes may also influence cognitive outcomes of non-musical functions. With regard to content, the first section of the paper suggests three specific hypotheses regarding how training related differences in the musician brain may be expected to bring about preserved musical ability following brain damage. In the following section, we then present an overview of the neuropsychological literature examining cognitive outcomes in musical functioning in brain-damaged professional musicians, specifically those with lesions or who have undergone surgical resection for epilepsy. Critically, we suggest that the examination of this neuropsychological literature can provide important insights: firstly, into the extent to which a patient s recovery of a given function is based on their premorbid experience and secondly, into the incidence and rate of more general functional recovery in musicians as well as their professional prospects following surgery. Finally, in the third section of the paper, given the important insights, which, we argue, may arise from the current approach of assessing cognitive outcomes following brain damage, we suggest some guidelines for future research. 1) Potential implications of training-related differences in the musician brain. There are various ways in which the structure and function of the musican brain differs from that of nonmusicians as a result of training. Based on these, at least three hypotheses can be proposed regarding factors that may bring about preserved musical function in brain damaged musicians. a) Anatomical changes resulting in a greater likelihood of preserved substrates Recent advancements in anatomical scanning and image processing techniques have greatly encouraged the study of the ways in which the musician brain is structurally different from that of non-musicians. Specifically, with the technique known as Voxel Based Morphometry, it is now possible to carry out quantitative and statistically rigorous analysis of the grey and white matter concentration of the whole brain (Wright et al. 1995), while other techniques like Diffusion Magnetic Resonance Imaging (dmri) allow researchers to carry out detailed analysis of white matter anatomical features (Basser and Jones 2002; Johansen- Berg and Behrens 2009). These techniques, along with functional imaging methods, have shown that musicians brains are characterized by increased volume in several areas. For instance, in one of the first studies to demonstrate training-specific changes in cortical anatomy, Elbert and colleagues (Elbert et al. 1995) used magnetic source imaging to show that in line with the increased finger movements string players carry out with the left hand, cerebral representations of the digits of the left hand were substantially larger than those of controls. The authors went as far as showing that this group effect was absent in the right hand and smallest in the left thumb (in line with extent to which these engage in fine finger movements). They also demonstrated a relationship between the observed changes in cortical territory and the age at which the musicians begantoplaytheirinstrument. Schneider and colleagues (2002) showed that the grey matter volume of the antero-medial portion of the Heschl s gyrus, (considered the primary auditory cortex), was not only 130 % larger in musicians than nonmusicians but also correlated with musical aptitude as measured by the Advanced Measures of Music Audiation (AMMA) tonal test (Gordon 1989). Gaser and Schlaug (2003), in addition to confirming grey matter volume differences in auditory brain regions such as the left Heschl s gyrus between musicians (professional keyboard players) and a matched group of amateur and non-musicians, also provided evidence that motor-related structures such as the primary motor and premotor regions as well as the cerebellum (Hutchinson et al. 2003) is expanded in musicians. Also striking is that the anterior half of the corpus callosum (Schlaug et al. 1995; Schlaug 2001) has been shown to be larger in musicians compared to nonmusicians suggesting greater communication between the hemispheres by virtue of a larger number of fibres crossing through this region. More recently, in further demonstration that musical training brings about a range of morphological and anatomical changes, white matter structure has been shown to change with degree of musical expertise (for instance, with years of piano training: Bengtsson et al. 2005) and results suggest that musicians have a larger volume and better organization of connections between temporal and frontal lobe

Neuropsychol Rev (2014) 24:445 460 447 regions (Halwani et al. 2011). Interestingly Halwani et al. (2011) not only found differences between musicians and non-musicians in the macro and micro structure of the arcuate fasciculus but they were able to show further differences between instrumentalists and singers that could account for the long term vocal motor training carried out by the latter. Finally, that anatomical changes which were observed at the sensorimotor level after approximately a year of musical practice in children were accompanied by improvements in auditory perception and motor skills (Hyde et al. 2009) demonstrates the relationship between anatomical variations and musical abilities in musical experts (Foster and Zatorre 2010). The volume changes observed as a function of expertise are particularly pertinent when one considers that the volume of preserved neural substrate following brain damage may be related to preservation of cognitive function. Such a hypothesis has already been put foward in the vast literature on Brain and Cognitive Reserve (Stern 2002). Models described in this literature have sought to explain certain paradoxical phenomena in cerebral aging and stroke patients; such as why some older adults fail to show typical clinical symptoms of Alzheimers disease in life despite posthumous evidence that they had the pathology, or why a stroke of a given magnitude may result in severe impairment in one patient but not another. In one group of models, reserve has been taken to describe an active process whereby alternative strategies are used to carry out a function (Stern et al. 2005;Stern2006). However, in another group of models, reserve is simply associated with the extent to which cerebral damage has depleted neural substrates (Mortimer et al. 1981; Katzman 1993; Satz 1993). This latter group of models tends to allude to a theoretical construct known as brain reserve capacity, which has been associated with such biological measures as number of synapses and brain size. A major assumption of these models is that brain reserve capacity may differ across individuals but that there exists a fixed threshold that must be reached before deficits in brain function arise. Specifically, according to these models, differences in cognitive outcomes, given a lesion of a particular size (or a degenerative process), are observed due to the size of brain reserve capacity. The threshold of brain damage necessary to bring about a given deficit is more quickly reached in individuals with lesser brain reserve capacity than those with greater brain reserve capacity, resulting in greater likelihood of impairments in the former. Performing structural MRI on a large group of stroke patients, Särkämö et al. (2009) demonstrated that those patients with less extensive damage to the auditory cortex and frontal lobe also showed less extensive auditory and musical dysfunction. Such a finding supports the notion of a relationship between the volume of preserved neural substrate and cognitive outcome. By extension, all else being equal (specifically the size of a lesion), it may be predicted that preserved or near preserved cognitive outcome will be more likely in an individual who possesses, before damage, more extensive substrate (larger brain reserve capacity) than in one who possesses less extensive substrate (smaller brain reserve capacity). The notion that a greater quantity of neural substrate results in superior function also finds support in studies showing a correlation between integrity of white matter structures and both music cognitive ability and other types of behaviour (Hyde et al. 2006; Catani and Mesulam 2008; Loui et al. 2009; Johansen-Berg 2010; Johansen-Berg et al. 2010). In line with the literature on reserve, we suggest that given the greater volume of grey and white matter and consequently greater brain reserve capacity in musicians relative to non-musicians (in specific and non specific brain areas following training), a lesion of a given size may be less likely to result in clinical symptoms in the former relative to the latter. b) Greater redundancy in the musician brain and access to different strategies A growing body of literature suggests that musicians differ from non-musicians in terms of the distribution of neural substrates involved in music processing (Ohnishi et al. 2011; Munteetal.2002; Habibi et al. 2013). In a seminal study using the dichotic listening experimental technique, Bever and Chiarello (1974) observed that as musicians capacity for musical analysis increases, the left hemisphere becomes increasingly involved in the processing of music (p. 539). Indeed, a number of studies have now demonstrated that, in addition to the holistic manner (subserved by the right hemisphere) in which non-musicians process pitch patterns, musicians also, as initially proposed by Jackson (1932), carry out more analytical forms of pitch pattern processing using their left hemisphere. Peretz and Morais (1980) showed that, during a melody recognition dichotic listening task, non-musician participants who carried out more analytic processing (having become aware of the dimensions manipulated by the experimenter) had a right ear advantage (suggesting left hemisphere predominance) and in doing so demonstrated that formal musical training is not necessary for involvement of the left hemisphere of the brain. However, various lines of evidence suggest

448 Neuropsychol Rev (2014) 24:445 460 that musicians are generally less likely than nonmusicians to limit themselves to one mode of processing. For instance, Peretz and Babai (1992) demonstrated that, when required to recognise melodies in a probe recognition task, musicians were not confined to the left hemisphere based analytic mode of processing but could flexibly use either contour or interval information, involving global and analytic processing respectively. Similarly, work from Habibi et al. (2013) showed that while non-musicians will mostly rely on the right hemisphere for detecting pitch deviations, musicians will usually recruit both hemispheres while carrying out such tasks. These studies are important in showing that the musician brain possesses a greater degree of redundancy than the non-musician brain and/or may have access to multiple strategies for carrying out a musical task. The important implication of such a state of affairs is that brain damage may not have the same degree of impact on musical skills in musicians and non-musicians. Specifically, one hypothesis could be that circumscribed damage, for instance limited to the right hemisphere, might result in preserved cognitive function in musicians who can easily use the left hemisphere to carry out pitch and melody processing. In contrast, it could bring about impaired function in non-musicians who have not developed the left hemisphere based analytical processing of pitch information to an equivalent extent. Indeed, the current hypothesis shares a lot in common with the, previously discussed, so called active cognitive reserve models that suggest that preserved function may be observed in some individuals as a result of recruitment of alternative brain networks that have not been damaged (Stern 2002; Stern et al. 2005; Stern 2006). In other words, those models which suggest that individuals who can use another strategy when a more standard one is impaired will be more likely to show preserved ability compared to those individuals who have no alternative approaches. Such a notion of differences at the level of available strategies is distinct from the notion of differences in the quantity of residual cerebral substrate as referred to when describing the potential implications of differences in anatomical changes in the musician brain. Indeed, the notion that differences in cognitive outcome may be accounted for in terms of the availability of alternative cognitive strategies may also be seen in the Reorganization of Elementary Functions (REF) model of Mogensen and Malá (2009). The authors observed that, very often in the neuropsychological literature, functions may be preserved despite damage to what had been thought of as essential cerebral substrates for that function. To explain this phenomenon, these authors described cognitive ability with respect to three different levels of realization. On the most basic level, they describe so called elementary functions (EFs) as information processing modules truly localized in the sense that they are mediated by local circuitry within a structure of the brain. They argue that these EFs, however, stand in contrast to socalled algorithmic strategies (ASs), which, in turn, are made up of numerous interacting EFs. Importantly, the model proposes that the psychological functions, the third realization level, are more closely associated with ASs than with EFs. Specifically, it suggests that the same surface phenomena or psychological functions may arise from potentially numerous different ASs, which in turn are comprised of an interaction of different sets of EFs. Thus, according to Mogensen et al. (2009, 2011) the possession of a greater number of EFs may bring about greater opportunities for the creation of new ASs, resulting in a greater likelihood of preserving a cognitive function. Importantly, findings showing that musicians use more parts of the brain to carry out a given musical function would seem to suggest that, due to plasticity mechanisms following high levels of training, musicians have an increased number of elementary functions. The consequence of this, according to Mogensen and Malá (2009), is a greater potential number of algorithmic strategies resulting in a greater tendency for preserved cognitive function following brain damage. It remains to be clarified to what extent such changes due to musical training may benefit non-musical function. However, the fact that musicians can make better use of both hemispheres during musical processing has been put forward as a reason why abilities in musicians may be robust to brain surgery (Schulz et al. 2005). c) Greater metaplasticity in the musician brain. First referred to as such by Abraham and Bear (1996), metaplasticity may be defined as the plasticity of synaptic plasticity or in other words, the ability to learn how to learn. Critically, it refers to the phenomenon whereby a synapse s plasticity is influenced by its previous history. There is some initial evidence that musical training endows musicians with greater metaplasticity than non-musicians. Ragert and colleagues (2004) sought to examine the altered sensorimotor cortical representation that had been shown to be induced by musical training. Their study revealed that not only do musicians, when compared with non-musicians, have lower discrimination thresholds in terms of tactile spatial acuity, but also that musicians had a greater capacity

Neuropsychol Rev (2014) 24:445 460 449 to further reduce these thresholds over the course of a few hours. They also showed that this effect in musicians correlated with how much they practiced on a regular basis and in doing so demonstrated a reliance of metaplasticity on extent of use. An increased capacity for plastic reorganisation in musicians has also been demonstrated in a Transcranial Magnetic Stimulation (TMS) study that showed greater motor cortical excitability and plasticity in musicians than in non-musicians (Rosenkranz et al. 2007). Watanabe and colleagues (2007) went further in showing that metaplasticity might be greater in musicians who started practice during the sensitive period of musical development: within the first 7 years of life (Penhune 2011). Importantly, these findings suggest that musicians will be better than non-musicians in relearning functions that have been impaired by brain damage. However, unfortunately, while efforts are being made to relate the changes seen at a molecular level to those seen at a structural, functional and behavioural level (Zatorre et al. 2012), these links still remain unclear and further studies will be needed to understand, for instance, how synaptic changes may be related to increases in the grey and white matter observed in musicians or the increased modulation of their somatosensory and motor cortex excitability. In sum, we suggest three ways in which the musician brain may be equipped to preserve musical function, or alternatively encourage its significant recovery following brain damage. We now examine the literature for any evidence to suggest that musical function may indeed tend to be preserved in musicians despite brain damage. 2) Neuropsychological literature on cognitive outcomes A number of reviews have examined general music cognitive outcomes in non-musician patients following brain lesions and surgery (Benton 1977;MarinandPerry 1999; Samson 1999; Stewart et al. 2006; Lechevalier et al. 2007; Maguire 2012). While they do not omit referring to cases where brain lesions result in cognitive deficits like aphasia while leaving music intact (Hébert et al. 2003; Peretz et al. 2004; Wilson et al. 2006), these reviews have tended to highlight the vast literature in which non-musicians acquire disorders of music listening following vascular, epileptic or traumatic lesion. Stroke, for instance, can cause a range of cognitive deficits with respect to language, memory, attention and orientation and attempts have been made to look at patterns of incidence (e.g. Tatemichi et al. 1994). Certainly, in the reviews of postmorbid musical abilities in non-musicians, the non-musicians reported on are shown to also suffer from some of such other deficits. Unfortunately the extent to which music is affected or spared compared to these other features is not clearly detailed due to a lack of systematic documentation of the full spectrum of cognitive impairments that may be seen in stroke and brain damaged patients. However we argue that perhaps even more interesting (and more plausible, given that they are more likely to be tested) may be to specifically examine the cognitive outcomes in professional musicians. Certainly of specific interest would be the nature of these outcomes given the musicians extensive premorbid musical ability. To this end, case reports were accumulated following searches on Pubmed and Psychinfo that employed the keywords epilepsy, stroke, brain damage, lesion, tumour and musicians and were published from the year 1900 onwards. Other relevant studies cited in the papers obtained from these searches were included. Reports were limited to professional musicians and teachers, as opposed to amateur musicians, and only studies with sufficient detail (of the state of expressive and receptive musical abilities) were included. Table 1 provides a summary of case reports of 35 brain-damaged musicians following lesions of different aetiologies (stroke, tumour and epilepsy surgery), who were identified through this procedure. It is important to bear in mind that biases may exist in the literature whereby a musician who has lost function is more likely to be reported than one who has not (Stewart et al. 2006). Indeed, a first look reveals that in more than half of the cases (20 out of 35), musical abilities in musicians are impaired following brain damage. Out of these 20 cases, the loss of musical abilities is accompanied by language deficits (12 patients: e.g. Souques and Baruk 1926; 1930; Alajouanine 1948; Jellinek 1956; Brust 1980, case 2) roughly in the same number of cases as it is not (8 patients: e.g. Judd et al. 1979; McFarland and Fortin 1982). 1 Whereas, in half of the cases, there is at least some damage to both receptive and expressive musical abilities (11 cases), in cases of dissociation, the loss of receptive but not expressive musical function (7 cases) is more frequent than the reverse (2 cases). Further, in addition to studies documenting severe and global impairments at either the receptive or expressive level or both, a number of other cases demonstrate fairly isolated deficits in specific skills such as music writing, music reading or the discrimination and production of rhythm (9/20: e.g. Horikoshi et al. 1997; Midorikawa and Kawamura 2000; Di Pietro et al. 2004). 1 In line with the evidence of double dissociation of musical and language cognitive abilities, cases of loss of language ability without any loss in musical ability are also seen (8 patients).

450 Neuropsychol Rev (2014) 24:445 460 Table 1 A summary of case reports of brain damaged patients Author Details Musical expertise Aetiology Regions affected Language abilities Musical abilities Specific music difficulty Receptive Expressive Alajouanine 1948 58 years old Composer Uncertain Likely left hemisphere Impaired Spared Impaired Assal 1973 64 years old Conductor/Pianist Infarction Left hemisphere Impaired Spared Spared Assal and Buttet 1983 54 years old Pianist/Organist + Music teacher Basso and Capitani 1985 Botez and Wertheim 1959 Lesion Left posterior temporo-parietal region Impaired Spared Spared 67 years old Violinist Infarction Bilateral : Left temporo parieto-occipital lobe and right posterior temporal region 20 years old Accordionist Tumour Right frontal lobe (posterior middle and superior frontal gyrus) + likely pars triangularis + likely subcortical regions Brust 1980 (case 2) 42 years old Jazz bassist Infarction Left posterior temporal region and inferior parietal lobule Cappelletti et al. 2000 51 years old Pianist/Guitarist +Composer Lesion Bilateral: Left posterior temporal lobe + smaller right occipito-temporal-parietal junction Di Pietro et al. 2004 48 years old Singer Infarction Left superior temporal gyrus, posterior part of middle temporal gyrus, and inferior parietal lobule Fasanaro et al. 1990 72 years old Violinist Infarction Left temporo-parieto-occipital region, splenium and thalamus (medial region) Finke et al. 2012 68 years old Cellist Lesion Bilateral: Right medial temporal lobe and left temporal (large sections), frontal and insular cortex (smaller sections). Galarza et al. 2014 36 years old Jazz guitarist Resection to treat medically intractable epilepsy Horikoshietal.1997 26 years old Pianist Resection of hematoma Left temporal lobe (extensive) and possible slight atrophy of frontal and parietal lobes + abnormally small right hemisphere volume Left occipito-temporal deep white matter surrounding the trigone, posterior horn of left lateral ventricle and splenium of the corpus callosum Impaired Spared Spared Spared Spared Impaired Impaired Impaired Impaired Spared Impaired** Spared Reading, writing Impaired Impaired** Impaired** Rhythm Impaired (mild anomic aphasia) Impaired** Spared Music reading and writing (pitch ) Impaired Spared Spared Spared Spared (Fully recovered) Spared (Fully recovered) Spared Impaired** Spared Music reading Jellinek 1956 (case 1) 46 years old Singer/Guitarist Glioma Left frontal lobe Impaired Impaired Impaired Spared Impaired Impaired Judd et al. 1979 51 years old Composer + Music teacher Infarction Right fronto-parietal and posterior temporal regions Judd et al. 1983 77 years old Composer Infarction Left occipito-temporal lobe Impaired Spared Spared Kawamura et al. 2000 57 years old Trombonist Haemorrhage Left parietal (angular gyrus) regions Impaired Impaired** Spared Music reading and writing Levin and Rose (1979) 58 years old Drummer Resection of tumour Left splenio-occipital region and occipital pole Spared Impaired Impaired (mild) Luria et al. 1965 57 years old Composer Infarction Left temporal and inferior parietal regions Impaired Spared Spared Mavlov 1980 61 years old Violinist + Music teacher McFarland and Fortin 1982 Infarction Left posterior inferior parietal and parieto-temporal regions 78 years old Organist Infarction Right superior temporal and parietal (superior marginal) regions Impaired Impaired** Impaired Rhythm Spared Impaired (mild) Impaired

Neuropsychol Rev (2014) 24:445 460 451 Table 1 (continued) Author Details Musical expertise Aetiology Regions affected Language abilities Musical abilities Specific music difficulty Receptive Expressive Midorikawa and Kawamura 2000 53 years old Piano teacher Lesion Left superior parietal lobule (cortical and subcortical) Midorikawa et al. 2003 62 years old Piano teacher Infarction Left superior temporal gyrus to angular gyrus (parietal region) Russell and Golfinos 2003 (case 1) Schulz et al. 2005 (case 1) Schulz et al. 2005 (case 2) Schulz et al. 2005 (case 3) 28 years old Singer Resection of tumour Right Heschl s gyrus Spared Spared (transient 3 week deficits) 45 years old Organist Resection to treat medically intractable epilepsy (benign tumour) 45 years old Trumpeter + music teacher Resection to treat medically intractable epilepsy (hippocampal sclerosis) 34 years old Organist Resection to treat medically intractable epilepsy (hippocampal sclerosis) Right anterior and lateral temporal lobe extending to hippocampus Right median temporal lobe including hippocampus Left median temporal lobe including anterior hippocampus and inferior amygdala Spared Impaired** Spared Music writing Impaired Impaired** Impaired** Rhythm/ Singing Spared Spared (self-assessment) Spared Spared (self-assessment) Spared Spared (self-assessment) Spared (transient 3 week deficits) Spared (self-assessment) Spared (self-assessment) Spared (self-assessment) Schön et al. 2001 65 years old Organist Ischemic lesion Left parieto-temporal lobe Spared Impaired** Spared Music reading Signoret et al. 1987 77 years old Composer/ organist Souques and Baruk 1926; 1930 Infarction Left temporo-parietal region Impaired Spared Spared NS Piano teacher Infarction Left posterior superior temporal region and angular gyrus (parietal region) Stanzione et al. 1990 26 years old Music teacher Lesion Left posterior temporo-parietal lobe Impaired (slightly anomic) Terao et al. 2006 62 years old Singer Infarction Right superior temporal cortex, parietal regions (supramarginal gyrus and postcentral gyrus), posterior insula Tzortzis et al. 2000 74 years old Composer Infarction and or degenerative atrophy Wertheim and Botez 1961 Impaired Impaired Impaired Impaired (mild) Impaired** Spared Music reading Impaired Impaired Bilateral: Both hemispheres Impaired Spared Spared 40 years old Violinist Vascular injury Left posterior superior temporal gyrus Impaired (mild) Wilson et al. 2013 26 years old Singer Resection to treat medically intractable epilepsy Zatorre 1989 17 years old Violinist Resection to treat medically intractable epilepsy Right inferior and middle temporal gyrus, Parahippocampal gyrus, Hippocampus, Amygdala Left anterior temporal lobe including hippocampus. Impaired Impaired (mild) Spared Spared Spared Impaired Spared Spared Impaired** indicates general preservation of function apart from the specific difficulty reported in the specific music difficulty column

452 Neuropsychol Rev (2014) 24:445 460 Importantly, a closer examination shows that reported musical deficits are often predictable by the specific cerebral substrates affected. The superior temporal gyrus has been shown to be specialized in auditory processing more generally and the processing of pitch, timbre and melodic contours more specifically, using a range of procedures including cortical stimulation in patients during surgery (Celesia 1976), intracranial recordings in presurgical patients (Liégeois-Chauvel et al. 1991) and lesion studies in non-musician humans (Samson 1999; Stewart et al. 2006). Accordingly, it may be noted that those musicians with extensive music receptive deficits have tended to show at least some damage to the superior temporal gyrus in line with this area s importance in such processing (e.g.wertheim and Botez 1961; McFarland and Fortin 1982). Conversely, deficits of, for instance, music reading or writing tend to implicate other regions in the occipital and parietal cortex: For instance, the patient from Midorikawa and Kawamura (2000) who presented with musical agraphia in the absence of any other expressive or receptive forms of amusia had damage to the left upper parietal lobe in line with previous evidence of its specific role in agraphia, while similarly, the young pianist reported by Horikoshi et al. (1997) presented with difficulties in music reading after brain damage that affected the left occipital parasplenial region but spared the superior temporal gyrus. Taken together, these studies are important in showing that musicians are highly susceptible to deficits following cerebral damage and further that in the majority of cases, the deficits shown by musician patients are predictable by the localization of damaged regions. However, having considered these cases of loss of function in musician patients which constitute the majority of reports, it is interesting to consider those cases where preserved function has been reported. Indeed, given the well documented bias in the literature whereby a loss of function is more likely to be documented than a preservation of function, it is interesting that up to 15 out of 35 patients were spared on both receptive and expressive musical function, often despite at least some damage to the temporal lobe area(luria et al. 1965; Assal and Buttet 1983; Judd et al.1983; Basso and Capitani 1985; Signoret et al. 1987). For instance, the patient of Basso and Capitani (1985) continued to be able to play the piano and to conduct his orchestra successfully despite a severe ideomotor and ideational apraxia following damage to both the left temporo-parieto-occipital lobe as well as the posterior part of the right temporal lobe. The authors observed that, the patient was able to programme and execute gestures with a high praxic content when they were the motor expression of musical processing. Similarly interesting is the case of the blind organist who lost the ability to read the alphabet (braille) but continued to be able to read and play music after damage to the left temporoparietal cortex (Signoret et al. 1987). 2 Reviewing the neuropsychological literature on music reading, Hebert and Cuddy (2006) would observe that although there were numerous studies where musicians showed text-reading difficulties in the absence of music-reading difficulties, there was only one case of a musician showing a selective impairment in music reading i.e. without a deficit in text-reading as well (Cappelletti et al. 2000). The existence of a selection bias, whereby, for instance, less attention may be paid to a music reading disorder than to a text reading disorder (even in musicians), should be borne in mind. Nevertheless, it is relevant to consider how authors have tried to account for the specific preservation in the musical domain of an otherwise compromised function (Tzortzis et al. 2000). Indeed, to explain why their patient M.M., whose progressive aphasia after damage to both hemispheres led to a severe deficit in naming that failed to extend to the naming of musical instruments, Tzortzis et al. (2000) would suggest an effect of overlearning. Specifically, they would suggest that the intensive practice and early acquisition of music may have provided M.M. with protection against loss of musical skills in the event of brain damage. At this point, it is perhaps worth considering the body of animal work that has examined the extent to which overlearning may lead to protection following brain damage. In these experiments, more trials than the criterion needed for mastery of a skill are presented to the animal model and then the effect of such overtraining on the ability to relearn the task following brain damage is reported. Indeed while there are some reports of failures of overlearning to be protective (e.g. Bignami et al. 1968) there are nonetheless several case of overlearning resulting in preserved ability to relearn following brain damage (Chow and Survis 1958; Orbach and Fantz 1958; Weese et al. 1973) The majority of studies in the literature document musicians who have undergone damage after lesions due to stroke and tumours. However, it is important to consider the aetiology of damage given the influence this might be expected to have on outcome. Indeed, 9 out of the 35 cases reported involved resection (for intractable epilepsy or hematomas or tumours) and it is interesting to note that in all but two of these cases, a preservation or recovery of both receptive and expressive musical function was reported. Indeed, apart from the drummer patient of Levin and Rose (1979) who presented with a range of deficits (e.g. music 2 Although these cases are in the left hemisphere, it is important to note that left hemisphere damage can cause deficits as evidenced by the fact that all but 4 of the 20 previously reported to have impairments showed their deficits after damage to the left cerebral hemisphere. Why so few cases of right hemisphere damage have been reported is not clear.

Neuropsychol Rev (2014) 24:445 460 453 reading and discrimination of pitch pattern and time interval) following resection of parts of left the occipital lobe in a procedure that involved cutting part of the corpus callosum (the splenium), and the pianist from Horikoshi et al. (1997) who showed difficulty with music reading following damage to the occipital and temporo-parietal lobe, all the other musicians showed only transient difficulty, if any, and a remarkable preservation of musical ability/ recovery of musical function following lateral and medial temporal lobe resection (7/9). Zatorre (1989) reported that, following an anterior left temporal lobectomy for the relief of medically intractable seizures, a 17-year-old epileptic pianist with absolute pitch showed no loss of function but rather an improvement after surgery in the notation of single piano tones 3 as well as high proficiency in the recall of threenote sequences despite showing an impairment in the equivalent verbal task (as is typically expected following a left temporal lobe lesion with a significant hippocampal resection in the predominant hemisphere for language). Russell and Golfinos (2003) reported on a singer who immediately following resection of right Heschl s gyrus suffered very poor pitch discrimination, melody recognition and singing ability, but who, by 3 weeks post surgery, had recovered all musical abilities. In another study, Schulz and colleagues (2005) required three musicians that had undergone unilateral temporal lobe epilepsy surgery to complete a questionnaire about their musical abilities as well as to write a free report about their professional training and expertise both pre and post surgery. The authors reported that all participants became seizure-free following their surgery, that there was no deterioration in performance on neuropsychological tests of intelligence or learning, but, most importantly, that none of the patients suffered deterioration in musical abilities with two rather reporting improvements in concentration and learning amongst other abilities. 4 Similarly, Galarza and colleagues (2014) provided a retrospective report on a jazz musician patient who not only had a resection of the left temporal lobe but also whose right hemisphere was abnormally small at 2 standard deviations below controls. With respect to outcome, the authors noted that although, following surgery, the patient showed a very impaired musical ability(alongwithsevereamnesia), after only a few years the musician enjoyed a return of his virtuoso status, along with a return of his memory abilities. Finally, a singer 3 In those with temporal lobe epilepsy, cortical regions adjacent to the site of the foci are often depressed leading to a worsening of cognitive ability. As suggested by Zatorre (1989), the improved performance afterwards may be explained by the removal of the epileptogenic focus which results in less cortical interference over the brain. 4 As suggested by the authors, and as with Zatorre (1989) improvements reported by these patients are likely due to the disappearance of seizures and consequently the functional recovery of regions depressed by the epileptogenic focus. reported by Wilson et al.(2013) whose surgery for an epileptogenic lesion involved the resection of right inferior and middle temporal lobe along with parts of the hippocampus, parahippocampus and amygdala, showed no musical impairments as a result but rather much improved singing (again likely due to the impact of the epileptogenic tissue perturbing the normal network). The reason for the higher preservation of function in the cases of temporal lobe resection reported here is likely that the early lesion, cerebral malformations or slow growing tumors that necessitated surgery (for example hippocampal sclerosis and benign tumour in the case of Schulz et al. 2005; and arteriovenous malformation in the case of Galarza et al. 2014) may have allowed reorganization of brain regions over and beyond that caused by training. Indeed, it is known that slow-growing lesions, such as arteriovenous malformations bring about extensive brain re-organization (Maldjian et al. 1996; Desmurget et al. 2007) that can result in spared function after surgery (Duffau et al. 2002) due to the given function shifting to the outside of the lesion (Rosenow and Luders 2004). Certainly, it is clear that the consequences of strokes, which damage healthy tissue, are likely to differ from the consequences of surgical treatment, the aim of which is to remove only damaged non-functional tissue. Specifically, as was seen here, it could be predicted that the latter type of damage should result in little or no loss of cognitive function compared to the former. It is therefore interesting that these findings of preserved ability in musicians following temporal lobe surgery contrast somewhat with those of non-musicians. Indeed, a series of studies examining the effect of temporal lobe resection on musical perception in non-musicians have revealed deteriorated performance as a result not just of right hemisphere damage (Kester et al. 1991; Milner 1962; Zatorre 1985; Samson and Zatorre 1988; Samson and Zatorre 1994) but also of left temporal lobectomy in a range of musical abilities including learning, discrimination and memory of melodies, and discrimination of note intervals, rhythms and metres (Zatorre 1985; Samson and Zatorre 1992; Kester et al. 1991; Liégeois-Chauvel et al. 1998; for review Samson 1999; Maguire 2012). Although we do not propose to compare the two groups given the difference in the evaluation approaches taken by those carrying out case studies of musical ability loss (not to mention the difficulty in comparing the initial abilities shown by the two groups), this suggestion of a tendency of musicians and non-musicians to show differing cognitive outcomes following surgery is worthy of note. We have suggested that these differences may arise from the structural and functional changes that take place in the brain as a result of the musical training that musicians undergo. Some experimental evidence suggests that musical training can improve non-musical function (in a so-called transfer of learning) (for a review, see Besson et al.

454 Neuropsychol Rev (2014) 24:445 460 2011). However there is currently no evidence in the literature to support the hypothesis that training-related brain changes might also have a beneficial effect on general cognitive outcome and recovery following brain damage. We argue that this could be an interesting issue to address in future studies. In conclusion, while the majority of musician participants so far reported in the literature, show considerable impairment following brain damage, the preservation of function in at least a subset of musicians, particularly those who have undergone temporal lobe resection, have raised some interest in those who have encountered these patients. The question of whether these individuals provide initial evidence of preservation of function due to intense training or overlearning (Tzortzis et al. 2000), and the question of whether training-related brain changes might have a beneficial effect on general cognitive outcome remain open. However, we argue that it may be addressed by further standardized documentation of musician patients. To this end, we suggest a number of guidelines for future research that may make help to clarify the question of whether premorbid experience as seen in musical experts influences cognitive outcome and functional recovery following brain damage. 3) Suggested guidelines for the future evaluation of relevant patients Playing music depends on a number of complex cognitive and motor abilities and a loss of these abilities would have a debilitating impact on the professional life of a musician. Accordingly, assessment of musicians cognitive outcome following lesion and neurosurgery is clearly important for what it can tell us about functional recovery in musicians generally and therefore their professional prospects. However, on a more basic level, the assessment of outcome in musicians and non-musicians is important for what it can tell us about the extent to which patients chances of recovery of different functions is based on premorbid experience. Further, such assessment may also provide insights into the extent to which any advantages shown by musicians, in terms of preservation of musical ability following damage, may be seen across domains i.e. whether any such advantages may extend to the preservation or functional recovery of non-musical abilities. In this section we seek to encourage the detailed documentation of cognitive outcome that would make such analyses possible by providing guidelines for evaluation of relevant patients in future research. First, we emphasise the need for a precise description of the case in question in terms of age, duration and onset of disease or neurological difficulty and localisation and aetiology of the lesion. Manual laterality should be determined to infer cerebral representation of language. The previous musical experience and degree of expertise of the musician should also be determined. One approach to quantifying expertise is to determine the amount of musical training in terms of the number of hours of dedicated practice undergone (Foster and Zatorre 2010; Foster et al. 2013). With estimates of practice hours per week for each year, a cumulative measure of hours of musical practice may be obtained and considered alongside the age at which musical training commenced. In addition, questionnaires that determine musical ability based on multidimensional criteria may be used. A general comprehensive neuropsychological evaluation including an anamnesis of medical history to identify any developmental learning disorders (e.g. dyslexia, dysorthographia, dysphasia, dyscalculia etc.) should then be carried out so as to allow examination of cognitive outcomes of non-musical functions. A measure of general intellectual function is important to interpret other cognitive abilities including attention, orientation, perception, memory, executive, somatosensory, and motor function. Table 2 proposes a list of recommended investigations into cognitive functioning with examples of possible tests to use for each domain. Although this list is not exhaustive and should be adapted to the level of the patients, it provides suggestions about the functions that need to be investigated. A list of auditory and musical skills recommended to screen musical difficulties is detailed in Table 3. Note that the once again this list of musical functions is not exhaustive and is only to provide suggestions as to how one might initiate neuropsychological assessement of musicians. In the case of severe musical difficulties being reported, a preliminary assessment of musical abilities with the Montreal Battery for the Evaluation of Amusia (MBEA :Peretzetal.2003) may be carried out, as the failure of a musician to perform normally on this very basic test of musical abilities would provide confirmation of severe loss of ability. Timing and synchronization to musical sequences can be also assessed with tasks from the Battery of Assesment of Auditory Sensorimotor and Timing Abilities (BAASTA, Farrugia et al. 2012). We also suggest that the investigators assess processing of other auditory sounds (environmental, vocal and verbal) in addition to musical sounds to in order to determine the extent of dysfunction in the auditory sphere and to verify domain specificity of any observed deficits. Finally, in the interests of disentanging potential effects of musical training from other effects that may also influence brain and cognitive reserve and metaplasticity, factors such as education, occupation (as well as that of the parents), family income, parental language, diet, drinking and smoking habits should also be reported. Details of involvement in other intellectual and physical activities should be solicited.