**1. Introduction**

Traumatic brain injury (TBI) is a common and devastating neurological disorder, affecting over 50 million people each year worldwide [1], with often poor long-term outcomes [2].

**Citation:** Sihvonen, A.J.; Siponkoski, S.-T.; Martínez-Molina, N.; Laitinen, S.; Holma, M.; Ahlfors, M.; Kuusela, L.; Pekkola, J.; Koskinen, S.; Särkämö, T. Neurological Music Therapy Rebuilds Structural Connectome after Traumatic Brain Injury: Secondary Analysis from a Randomized Controlled Trial. *J. Clin. Med.* **2022**, *11*, 2184. https://doi.org/10.3390/ jcm11082184

Academic Editors: Nada Andelic, Cecilie Røe, Eirik Helseth, Emilie Isager Howe, Marit Vindal Forslund and Torgeir Hellstrom

Received: 11 February 2022 Accepted: 11 April 2022 Published: 14 April 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

The primary neuropathology associated with TBI is structural white matter damage, that is, axonal injury, established already half a century ago by the seminal post-mortem studies [3,4]. In TBI, the white matter damage disrupts effective neural communication and impairs neural networks that link brain structure to function, typically causing deficits in cognitive, social, and emotional functioning [5–7]. Among the most common, persistent, and disabling aspects of cognitive impairment following TBI is executive dysfunction [8], which is often caused by diffuse axonal injury (DAI) resulting in widespread connectivity deficits in the brain [9,10].

Due to the neuropathology of TBI, neuroimaging plays a crucial role in its assessment. In the acute setting, MRI scans are used to guide appropriate management by detecting brain injuries that require neurosurgical interventions or further monitoring. However, routinely acquired MRI might not reveal findings even in patients with symptoms due to the DAI mechanism [11,12]. Therefore, advanced neuroimaging techniques reflecting white matter structures such as diffusion tensor imaging (DTI) have been under active research in TBI. Studies have shown that TBI patients have structural connectivity deficits in multiple white matter tracts, most commonly in long coursing and commissural fibres that are most vulnerable to injury in TBI [12–14]. After the initial injury, the degeneration of white matter tracts persists for years [15,16] and is associated with poor long-term functional and cognitive outcomes in TBI [16,17].

While DTI has been used to improve the diagnostics and classification system of TBI, very little research has thus far been carried out in determining treatment-induced white matter neuroplasticity changes in TBI. Given the dynamic nature of DAI, intervention studies charting the possible discontinued deterioration or recovery of white matter injury over time would be of great interest. Ultimately, this information would help to target clinical interventions for rehabilitation. A recent animal study suggests that cognitive TBI treatments can induce white matter plasticity [18], but to our best knowledge, studies on treatment-induced structural white matter neuroplasticity in TBI patients have not been published.

Cognitive therapies have emerged as efficient treatments to restore cognitive functions and improve functional outcomes in TBI [19,20]. In cognitive neurological rehabilitation, music has emerged as a viable and applicable tool during the past decades, partly owing to its capacity to engage widespread neural networks across bilateral cortical and subcortical areas [21]. Research findings in stroke patients suggest that music-based interventions engage an array of cognitive functions, resulting in cognitive improvement [22,23] and structural and functional neuroplasticity changes [24,25] in the damaged brain. In our recent randomized controlled trial (RCT) on patients with moderate-to-severe TBI, we found that a 3-month neurological music therapy (NMT) intervention enhanced executive function and increased structural grey matter neuroplasticity in prefrontal areas [26] as well as normalized or enhanced functional connectivity in the brain, especially in frontal and parietal regions [27]. Given the extent of the brain regions and pathways stimulated by music, it is possible that NMT may induce also more widespread structural connectivity changes in TBI, but this has not been studied previously.

Here, using longitudinal diffusion MRI (dMRI) data from our previous RCT [26,27], we set out to determine as a secondary outcome of the trial NMT-induced structural white matter connectivity changes and their association with improved executive function in a sample of 25 TBI patients with a 6-month follow-up. To do this, we carried out white matter connectometry analysis utilizing quantitative anisotropy (QA), which has been shown to be superior to conventional single-tensor based or tract-based analysis [28]. Connectometry analysis utilizes permutation testing to identify group differences in white matter tracts across the whole brain and has been used in neurological patients, for example, to uncover white matter tracts subserving word production [29] and verb retrieval [30] in post-stroke aphasia. Based on our previous findings of increased grey matter volume and functional connectivity, which were most evident in right prefrontal areas, after the NMT [26,27], we hypothesized that it would induce structural connectivity changes especially in the right frontal and dorsal pathways.

#### **2. Materials and Methods**

#### *2.1. Subjects and Study Design*

Forty TBI patients from the Helsinki and Uusimaa Region of Finland were recruited through the Brain Injury Clinic of the Helsinki University Central Hospital (HUCH), Validia Rehabilitation Helsinki, and the Department of Neurology of the Lohja Hospital during 2014–2017 to this RCT (NCT01956136). The inclusion criteria were: (1) diagnosed TBI according to the International Statistical Classification of Diseases and Related Health Problems, 10th revision (ICD-10), fulfilling the criteria of at least moderate severity (Glasgow Coma Scale [GCS] score: ≤12 and/or loss of consciousness >30 min and/or post-traumatic amnesia [PTA] ≥24 h and positive findings on CT/MRI); (2) time since injury ≤24 months at the time of recruitment; (3) cognitive symptoms caused by TBI (attention, executive function, memory); (4) no previous neurological or severe psychiatric illnesses or substance abuse; (5) age 16–60 years; (6) native Finnish speaking or bilingual with sufficient communication skills in Finnish; (7) living in the Helsinki-Uusimaa area; and (8) understanding the purpose of the study and being able to give an informed consent. Patients with GCS score 9–12 and/or loss of consciousness 30 min–24 h and/or PTA 1–7 days and abnormal structural imaging on CT/MRI were defined as having a moderate TBI, and patients with GCS score 1–9 and/or loss of consciousness >24 h and/or PTA > 7 days and abnormal structural imaging on CT/MRI were defined as having a severe TBI [31]. Both the extended Glasgow Outcome Scale (GOSE) [32] and the Neurological Outcome Scale for Traumatic Brain Injury (NOS-TBI) [33] were administered to obtain information of the overall symptoms and current functional outcome after TBI. The trial was conducted according to the Declaration of Helsinki and was consistent with good clinical practice and the applicable regulatory requirements. The trial protocol was approved by the Coordinating Ethics Committee of the Hospital District of Helsinki and Uusimaa (reference number 338/13/03/00/2012) and all participants signed an informed consent.

The study was a single-blind crossover RCT with a 6-month follow-up period. During 2014–2017, 4994 patients with TBI were screened for eligibility, 190 met the inclusion criteria, and 40 were randomized to the AB (*n* = 20) and BA (*n* = 20) groups. The randomization was stratified for lesion laterality and performed using a random number generator by a person not involved in patient recruitment or assessments. To ensure steady allocation to both groups across the trial, the randomization was done in batches of two consecutive patients. After the baseline measurements at time point 1 (TP1), which included MRI scans and neuropsychological assessments, the AB group received NMT in addition to standard care for the first 3 months, whereas the BA group received only standard care. At the 3-month crossover point (TP2), follow-up measurements using the same outcome measures were carried out. After this, the BA group received NMT and standard care for 3 months and the AB group received only standard care. At the 6-month completion point (TP3), the measurements were carried out once again. All assessments were carried out by research personnel blinded to the patients' group allocation. Due to the nature of the intervention, patients were not blinded. Standard care comprised any physical therapy, occupational therapy, speech therapy or neuropsychological rehabilitation which the patients received in public (or private) healthcare during the study period. There were no statistically significant differences between the AB and BA groups in the amount of received standard care [26].

Out of the 40 randomized patients, 1 participant dropped out before the TP1 measurements, 2 participants dropped out before TP2, and another 3 participants dropped out before TP3. The dropouts were mainly due to lack of energy and motivation. All dropouts (*n* = 6) occurred in the BA group, which was likely linked to the long waiting period before the intervention. Of these, five took place before the onset of the intervention. Of the remaining 34 patients, 1 was excluded from the analyses due to intensive self-implemented

piano training, which was not part of the trial protocol, and 8 were excluded from the analyses due to lack of MRI data owing to contraindications or technical difficulties during the scanning. Finally, 25 patients (AB: *n* = 16, BA: *n* = 9) completed the MRI acquisition in the three time points and were included in the present study. The flowchart of the included patients with TBI is shown in Figure 1.

**Figure 1.** Flow diagram outlining the trial.

#### *2.2. Intervention*

The NMT intervention is described in detail in our previous publication [26]. Briefly, it consisted of 20 individual therapy sessions (2 times/week, 60 min/session) held by a trained music therapist at Validia Rehabilitation Helsinki. No previous musical experience was required to participate in the NMT. The focus of the NMT was on active musical

production using different instruments (drums, piano) in three training modules involving (i) rhythmical training (playing sequences of musical rhythms and coordinated bimanual movements on a djembe drum and on own body), (ii) structured cognitive-motor training (playing musical exercises on a drum set with varying levels of movement elements and composition of drum pads), and (iii) assisted music playing (learning to play own favourite songs on the piano with the help of figure notes). All modules also included musical improvisation to facilitate more creative and interactive musical expression The difficulty level of the exercises was initially adjusted and then increased in a stepwise manner within and across the NMT sessions, to meet the skill level and progression of the patient.

#### *2.3. Neuropsychological Assessment*

The primary behavioural outcome measure was change in performance on the Frontal Assessment Battery (FAB) [34]. Assessing global executive function and being applicable across all severity levels of TBI, the FAB measures different aspects of frontal lobe functions and consists of six subtests exploring conceptualization (similarities subtest), mental flexibility (lexical fluency), motor programming (Luria's fist-edge-palm test), sensitivity to interference (conflicting instructions), inhibitory control (go-no go task) and environmental autonomy (prehension behaviour). The FAB total percent score (percentage correct) formed the composite score of executive function.

#### *2.4. MRI Data Acquisition and Reconstruction*

All patients were scanned on a 3T Philips Achieva MRI scanner (Philips Medical Systems) with a standard 8-channel head matrix coil at the HUS Helsinki Medical Imaging Center at HUCH. The MRI protocol comprised high-resolution T1-weighted anatomical images and whole-brain diffusion-weighted imaging (DWI) data (TR = 11,106 ms, TE = 60 ms, acquisition matrix = 112 × 112, 70 axial slices, voxel size = 2.0 × 2.0 × 2.0 mm3) with one non-diffusion weighted volume and 32 diffusion weighted volumes (b = 1000 s/mm2).

The DWI data were reconstructed in the Montreal Neurological Institute (MNI) space using q-space diffeomorphic reconstruction (QSDR) [35] that allows the construction of spin distribution functions (SDFs) [36]. The b-table was checked by an automatic quality control routine to ensure its accuracy [37]. Normalization was carried out using the anisotropy map of each participant and a diffusion sampling length ratio of 1.25 was used. The data output was resampled to 2 mm isotropic resolution. Quality of the normalization was inspected using the R2 values denoting goodness-of-fit between the participant's anisotropy map and template as well as inspecting the anatomical localisation of each participant's forceps major and minor to confirm the normalization quality [29]. The restricted diffusion was quantified using restricted diffusion imaging [38] and QA was extracted as the local connectome fingerprint [39] and used in the connectometry analysis. QA-based tractography has been shown to outperform traditional fractional anisotropy-based methods by being more specific to individual's connectivity patterns [39] and less susceptible to the partial volume effect of crossing fibres and free water, as well as to provide better resolution in tractography [40].
