In this study, we reported the case of CM, a musician diagnosed with a right frontoparietal high-grade glioma with preserved affective prosody and music processing skills after the complete surgical removal of the right dorsal pathways (including arcuate fasciculus). This study provides new evidence about the relationship between right white matter circuitry and affective prosody and music perception.
Sustained prosody and music perception abilities despite the complete removal of the right dorsal stream in CM suggests that the integrity of the right dorsal white matter is not a necessary condition to support prosody and music processing. A model by Schirmer and Kotz (2006) proposes that emotional recognition in spoken language is a multi-step process with individual sub-processes that are hierarchically organized along the ventral stream. The lower step would be the acoustic analysis mediated by the primary and secondary auditory cortices of both hemispheres. This bilateral primary sensory processing is followed by the extraction of emotional meaning from acoustic cues, a process where the right ventral stream plays a major role. Finally, the emotional meaning is available for high order cognitive functions such as explicit evaluative judgments and involving the inferior frontal gyrus and orbitofrontal cortex in the right hemisphere. Subcortical structures such as the putamen, amygdala and parahippocampal gyrus have also been associated with emotional processing (Grandjean, 2020; Hellbernd & Sammler, 2018; Pannese et al., 2015). For example, auditory information is projected from the thalamus to the amygdala and from here to cortical regions, enabling rapid and rough emotional analysis of relevant acoustic information (LeDoux, 2000). This model provides a more comprehensive view of the regions involved in the sub-processes related to prosody, where the right ventral pathway is an intermediate step involved in extracting the emotional meaning. This is in line with evidence derived from studies in neurological patients that highlight the critical role of the right ventral stream structures in deficits related to affective prosody (Sheppard et al., 2020; Wright et al., 2018, Sihvonen et al., 2022). Therefore, the fact that CM preserved his prosodic abilities is in line with the literature, arguing that the functions sustained by the right dorsal stream might not be essential for emotional recognition. The dorsal stream is thought to be involved in the processing of temporal aspects of speech rhythm processing (Poeppel and Assaneo, 2020, for a review), but the primary and secondary auditory cortices of both hemispheres also process the temporal dimensions of auditory cues at a lower level (Schirmer & Kotz, 2006). Interestingly, CM’s isolated impairment in fear recognition in prosody is resembles results by Sheppard and colleagues (Sheppard et al., 2020) who showed a similar selective impairment (following strokes affecting the right ventral stream and amygdala). Moreover, Sheppard and colleagues’ results further confirmed that this selective impairment was associated with lesions affecting the ventral stream. Previously, the difficulty in recognizing fear in musical excerpts was observed by Gosselin and colleagues in patients after unilateral temporal lobes excision (Gosselin et al., 2005). In that study, the authors observed that this specific impairment followed excisions including the vicinity of amygdala, which in our patient was spared. Future studies may aim at disentangling if emotional recognition of fear in music and prosody share neural correlates, since for CM the recognition of fear in music was spared contrarily to the recognition of fear in prosody (see also: Frünholtz et al., 2015; Paquette et al., 2018; Sachs et al., 2018).
With regard to music processing, Sihvonen and colleagues (Sihvonen, Ripollés, et al., 2017) have recently shown that damage to the right IFOF is the strongest predictor for post-stroke amusia. This tract was preserved in CM, who did not show any sign of amusia. We also observed that the patient’s lesion did not overlap with the lesions of 77 stroke patients who presented amusia in a previous study by Sihvonen and collaborators (2016). The surgical cavity in our patient involved regions of frontal and parietal cortices located superiorly compared to Sihvonen and collaborators’ VLSM findings. Additionally, the lesions presented in this group of amusic stroke patients comprised subcortical structures. The evidence coming from individuals with congenital amusia suggests that malfunction of the dorsal stream is involved in this condition (Peretz et al., 2016, Loui et al., 2009). Hovewer, this last evidence requires further research, as Chen and colleagues observed that these relationships may vary depending on neuroimaging methodology applied (2015).
More importantly, a recent study (Sihvonen et al., 2022) investigated for the first time the overlap of lesions resulting in both aprosodia and amusia in a large sample of subacute stroke patients. These authors showed that both acquired aprosodia (linguistic and affective) and acquired amusia (pitch and rhythm) were associated with significantly overlapping lesion patterns in the right frontal and basal ganglia regions. Importantly, supporting the present findings, Sihvonen and colleagues also observed that the right IFOF was the strongest predictor for both aprosodia and amusia (2022). Therefore, this study and the present one provides crucial information about the role of the right ventral stream in sustaining these functions. Additionally, the same authors found no evidence for dorsal stream involvement (see also Sheppard et al. 2020) in prosody perception.
Previous, very limited research conducted in patients with brain tumors and using intraoperative electrical stimulation (ESM) tends to converge with the idea that disruption of music processing and singing is associated to right fronto-temporal regions. In this sense, electrical stimulation applied to the anteroposterior pars opercularis of the right inferior frontal gyrus elicited spontaneous switch from a speaking to a singing mode in a single case patient (Herbet et al., 2015). Katlowitz and colleagues (2017) also reported a case of a professional male vocalist who showed that ESM applied to the right posterior superior temporal gyrus suppressed singing but not speaking. Similar findings involving right superior temporal regions in music processing were reported by Garcea and collaborators (2017) using ESM in musician who had a tumor in the right temporal lobe (see similar results by Hohmann et al., 2018 using transcranial direct current stimulation). Disentangling musical processing from emotional musical processing, as well as further ESM studies involving right temporal lobe in brain tumor patients could be an interesting avenue for future research.
An important aspect to consider is that CM has unilateral damage, and the intact regions of the left hemisphere may contribute to compensating for the damage to the right arcuate fasciculus. Despite the right hemisphere playing a major role in affective prosody and music processing, neuroimaging studies show bilateral activations and suggest that the left hemisphere subsystems support these functions (Belyk & Brown, 2013; Witteman et al., 2012). For instance, the left inferior frontal gyrus is proposed to integrate prosodic information with linguistic information (Schirmer & Kotz, 2006), and several regions of the left hemisphere are involved in interval and rhythm identification in music (Patel, 2010). These data provide evidence of a clear cross-talk between hemispheres regarding music processing, which may suggest also the potential emergence of compensatory mechanisms after injury (Schuppert et al., 2000). In relation to linguistic prosody, the interhemispheric interaction (and left-hemispheric) contribution is even clearer (Sammler et al., 2010; van der Burght et al., 2019). However, a recent study using VLSM and deterministic DTI tractography in stroke patients did not reveal clear associations between left-hemispheric structures and linguistic and affective prosody (Sihvonen et al., 2022).
Functional plastic changes following tumoral lesions are common and include reorganization within injured areas, and reshaping via recruitment of adjacent functional areas and contralateral homologous regions (Duffau, 2008). Better functional compensation is associated with slowly growing lesions. The tumor of this patient was of high-grade, meaning that the lesion expanded relatively fast in time. This might rule out the explanation that prosody and music processing are maintained because of a rapid functional reorganization involving intact regions (Noble & Dietrich, 2004) and implied that the preserved abilities in the patient are not just a matter of reorganization/plasticity, but may indeed suggest that the right dorsal stream is not substantially involved in prosody/music processing. However, in some cases high grade gliomas may be detected after a long period of less malignant tumor (e.g., WHO grade II) growth, which would in turn support the plasticity hypothesis.
Finally, it is worth mentioning that our patient did not report language and cognitive complaints, which is in agreement with what we observed during the neuropsychological evaluation. However, the patient reported that he needed to put more effort to follow the rhythms during choir sessions. This subtle deficit in rhythm processing could be associated to the lesion in the right dorsal stream, as this stream has been previously associated to amusia (Sihvonen et al., 2017a), being also implicated in rhythm perception in music (Chen et al., 2006; Merchant et al., 2015) and short-term music learning (Vaquero et al., 2018). It is also important to note that the evaluation carried out in our patient might not have been sensitive enough to detect more subtle music processing deficits, especially in an experienced musician. One possibility is that other tasks covering the broad spectrum of music processing abilities could have revealed errors related to the patient's complaints.
In conclusion, this study describes the case of an experienced musician without the right dorsal stream who preserved prosody and music processing skills after tumor surgery. The evidence provided here may support models of musical processing which involve (mainly) ventral processing stream in this regard.
LIMITATIONS
Given that the rhythmic and metric subtests of the MBEA are not the measure of musical beat processing per se (Tranchant et al., 2021) we cannot discern if CM presented with beat-perception deficits. If encountering patients with a neural profile similar to CM, we would complement their assessment by the Beat Alignment Test (BAT, Yversen & Patel, 2008). Another limitation is that we could neither consider the effect of CM’s extensive musical training on the premorbid brain organization, nor its role in brain reorganization after surgery.