**4. Discussion**

This structural white matter connectometry study set out to determine structural connectome changes induced by NMT and their relation to improved executive function in patients with TBI. Our novel findings were that compared to standard care, NMT enhanced structural connectivity in right frontal dorsal and projection pathways as well as in the corpus callosum. The enhanced structural connectivity was associated with improved executive function. To our best knowledge, this is the first study on patients to link treatment-induced white matter neuroplasticity changes with improved cognitive outcome in TBI. This study provides novel and crucial information about the neural mechanisms of music therapy induced brain recovery following TBI.

Executive dysfunction is considered to be the core clinical feature of TBI [20], particularly in moderate-to-severe cases, on which our study focused on, contributing significantly to both acute and chronic disability. Executive function is a broad term referring to higher order cognitive processes that enable individuals to regulate their thoughts and actions during goal-directed behaviour [43]. These processes are widely thought to include inhibition of prepotent responses, attentional control, task switching and working memory updating [44]. Neurally, executive function contributes to the coordination of these processes across a network of brain structures that need to work in concert. A large-scale quantitative meta-analysis of 193 functional neuroimaging studies has revealed that such executive control network comprises dorsolateral prefrontal, anterior cingulate, and parietal cortices [45]. To allow efficient communication, these spatially distributed brain regions are structurally connected via white matter pathways [46]. The three dominant structural components of this executive control network are (i) interhemispheric connections via the corpus callosum, (ii) fronto-parietal pathways, mainly the superior longitudinal fasciculus connecting dorsolateral prefrontal cortex and the parietal lobe, and (iii) the corticostriatal pathways between right dorsolateral prefrontal cortex and striatum [46–48].

The dependence on multiple white matter pathways working in concert makes executive function vulnerable to reduced communication efficiency following DAI in TBI. This was reflected in a recent meta-analysis evaluating the relationships between white matter fractional anisotropy values and executive dysfunction following TBI [49]. Executive dysfunction in TBI was associated with damage to various white matter pathways, but was most significantly correlated with decreased fractional anisotropy values in the superior longitudinal fasciculus, an association tract connecting the frontal, parietal and temporal lobes [49]. Together with commissural tracts, association tracts are most critical for cognition because they transfer information between lobes and hemispheres [50]. They are also most vulnerable to injury in TBI [12–14] and undergo long-term degeneration after the initial injury [15,16], giving rise to the functional and cognitive outcomes in TBI [16,17].

The white matter is a biologically active component of the brain, amenable to treatmentinduced modifications that bring about beneficial behavioural change [24,51]. In TBI, two principles of white matter structure and function are pertinent to acknowledge in the rehabilitation. First, the severity and prognosis are related to the degree of initial axonal injury, that is, if the axonal scaffolding on which myelin can regain its integrity is preserved, the axonal recovery is plausible, and therefore the restoration of function [52]. Second, white matter structure possesses the capacity of plasticity, that is, modification of its structure by experience [53]. Since neurogenesis has no known clinically meaningful effect on (adult) brain recovery, restoration of function relies upon the ability of spared neurons to grow neurites and form new synapses to rebuild and remodel the injured networks [54,55]. This requires neuronal activity, that is, stimulation via rehabilitation [56] that provides a fertile ground for neuroplasticity and has been shown to produce plastic changes in the white matter including activity-dependent myelination [57] and enhanced oligodendrogenesis [58].

Our present results showing that NMT-induced white matter plasticity changes underpinning the improved executive function comprise the right (i) corpus callosum, (ii) frontoparietal connections, and (iii) the right corticostriatal tract parallel strongly with the previous evidence on the structural core of the executive control network [46–48,59] as well as with the evidence on critical white matter damage loci in TBI giving rise to executive dysfunction [49]. Our results are also well in line with previous DTI studies which have reported neuroplasticity changes in the arcuate fasciculus [60,61], superior longitudinal fasciculus [62,63], corpus callosum [63,64], and corticospinal tract [65] induced by musical training in healthy subjects and in the frontal aslant tract induced by music-based rehabilitation in stroke patients [24]. While previous evidence on treatment-related white matter plasticity in TBI is scarce, our results argue that structural connectivity can be enhanced via treatments in TBI and that the treatment-related white matter changes are associated with improved cognitive outcome.

The present study has some potential limitations that need to be considered when evaluating the findings. Due to the study design, it is impossible to infer whether some components of the intervention, for example active or passive engagement during musicmaking or movements while playing musical instruments, were more relevant to clinical improvement than others. Moreover, although the study is the largest RCT utilizing NMT in moderate-to-severe TBI to date, the sample size remains relatively modest (*n* = 25) and may preclude the detection of smaller neuroplasticity effect sizes due to lack of statistical power. The limited sample size also prevents us from making more fine-grained analyses of what are the differences and commonalities in the patterns of white matter changes and executive function recovery between different time points. Therefore, future longitudinal studies with a stratification of patients by time since injury as well as lesion site would be warranted to explicitly test the relationship of music therapy induced cognitive and neuroplasticity changes across the recovery spectrum.
