Microstructural Changes in the Brainstem Auditory Pathway in Children with Hearing Loss

Abstract

Objective:

To assess the utility of diffusion tensor imaging (DTI) of the auditory pathway in children with sensorineural hearing loss (SNHL).

Study Design:

Retrospective cohort study.

Setting:

A single academic tertiary children’s hospital.

Patients:

Sixteen pediatric patients with bilateral SNHL of at least moderate severity in the poorer ear (8 male, mean age 5.3±4.9 years). Controls consisted of age and sex-matched children with normal hearing who were imaged for non-otologic, non-neurologic medical concerns and found to have a normal MRI.

Interventions:

Three Tesla magnetic resonance was used for DTI.

Main Outcome Measures:

Quantitative diffusion tensor metrics were extracted from the superior olivary nucleus (SON), inferior colliculus (IC), and ipsilateral fiber tracts between the SON and IC delineated by tractography.

Results:

We identified differences in fractional anisotropy (FA) of the SON between the SNHL cohort and controls (0.377±0.056 vs 0.422±0.052; p=0.009), but not in the IC. There were no differences in the mean diffusivity (MD) values in the IC and SON. Among younger children (≤5 years), MD was decreased in the SNHL cohort compared to controls in the IC (0.918±0.051 vs 1.120±0.142; p<0.001). However, among older children (>5 years), there were no differences in MD (1.124±0.198 vs 0.997±0.103; p=0.119). There were no differences in MD or FA in the white matter fibers of the IC-SON tract.

Conclusions:

Our results suggest abnormal neural tracts along the central auditory pathway among children with SNHL. Longitudinal studies should assess the prognostic value of these MRI-based findings for assessing long-term outcomes and determining intervention efficacy.

INTRODUCTION

Sensorineural hearing loss (SNHL) is the most prevalent congenital sensory deficit, affecting up to 6 in 1000 liveborn infants.^1,2 There is an emphasis on early screening and identification of hearing loss in children, as they are at significant risk for speech delay.³ Once hearing loss has been identified, children undergo extensive workup to identify etiology and to assess candidacy for interventions such as cochlear implantation. Early intervention is of paramount importance in order to maximize speech, language, and other developmental outcomes in children with congenital SNHL.⁴ Factors such as patient characteristics (e.g., age at intervention, etiology and duration of auditory deprivation), the patient’s environment (e.g., home environment, socioeconomic status), and the baseline status of the auditory system (e.g., etiology of hearing loss, neural integrity) are known to influence outcomes in children with SNHL.^5,6 Considerable variability in developmental and audiologic outcomes remains after accounting for these factors. Therefore, there is a need to identify additional prognostic factors that can be clinically derived and standardized.

Brain and internal auditory canal imaging is integral to the diagnostic workup of pediatric hearing loss. MRI is a diagnostic tool that can evaluate the inner ear and can provide insight into development, congenital brain malformations, and the integrity of soft tissues including the cochlea and the cochlear nerve.⁷ There is a growing body of literature exploring the potential for this imaging modality to characterize underlying neural integrity in children with SNHL, which may provide prognostic insight for developmental and audiologic outcomes.^8–10 Although conventional MRI provides information on structures at the macroscopic level, it is not sensitive enough to delineate gray and white matter microstructure. Diffusion tensor imaging (DTI) is an MRI technique that offers quantitative delineation of brain microstructure at specific regions of interest (ROI). In addition, DTI-based fiber tractography allows for the 3-dimensional reconstruction of white matter tracts, providing another tool for evaluating myelin and axonal integrity by quantitating water diffusion directionality.¹¹

Several studies have previously employed DTI to assess whether there is microstructural alteration of the auditory pathway in patients with hearing loss.¹⁰ For instance, this technique has been used to investigate changes in the auditory pathway after cisplatin and radiation therapies in children with medulloblastoma.¹² Furthermore, brain changes identified by DTI has been used to predict long-term auditory and speech outcomes after cochlear implantation^13,14, as well as school performance in children with hearing loss.¹⁵ Prior studies of the auditory pathway have mainly focused on characterizing more central ROIs and fibers that correspond to the auditory radiations.^10,16,17 However, investigations of the brainstem in children with SNHL remains sparse, and no previous study has applied tractography to characterize this area in the context of congenital hearing loss. In this pilot study, we use DTI to investigate the brainstem region of the central auditory pathway in children with SNHL.

MATERIALS AND METHODS

Study Cohort

Medical records and MRI data from 16 children (ages 1 day to 17 years) who received imaging studies at a single academic tertiary children’s hospital from 2011 to 2019 were retrospectively reviewed after approval by the Institutional Review Board. Inclusionary criteria were as follows: an MRI brain and IAC scan that captured the necessary sequences for DTI analysis (detailed in MRI Data Acquisition) with a negative MRI report, a diagnosis of bilateral sensorineural hearing loss of at least moderate degree in the poorer ear as documented via audiologic evaluation; and no cochlear implantation at time of MRI. Exclusionary criteria included lack of necessary DTI sequences as well as corrupted MRI imaging due to metallic artifacts or excessive motion. Several members of the study cohort have been included in a prior study.⁸

The control group included 16 age and sex-matched children with normal hearing, without neurologic or developmental deficits, and with normal-appearing brains on MRI as verified by a board-certified, pediatric neuroradiologist (KWY). As previously described⁸, control subjects obtained brain MRI at 3T for non-hearing loss related indications including syncope, nausea, family history of aneurysm or cancers, cancer screening (e.g., P53 mutation), scalp nevus, isolated facial lesions, naso-orbital dermoid, orbital strabismus, sinus disease or inflammatory nasal obstruction, and familial short stature.

Clinical Data

We performed chart review of all children with SNHL to verify laterality, degree of hearing loss in the poorer ear, and etiology of hearing loss. We obtained audiologic data from the pure-tone audiometric evaluation or the auditory brainstem response evaluation performed closest to the MRI date. We also obtained information regarding audiologic intervention route, including cochlear implantation and other hearing device use.

MRI Data Acquisition

MRI scans were obtained at 3T (Discovery 750W; GE Healthcare, Milwaukee, Wisconsin) with an 8-channel head coil. Both high-resolution T1-weighted (3D SPGR, TR = 7.76 ms, TE = 3.47 ms, FOV = 240 × 240 mm², acquisition matrix = 512 × 512, voxel size = 0.4688 × 0.4688 × 1 mm³, orientation = axial) and diffusion-weighted images were acquired as part of the pediatric brain MRI protocol. Diffusion data were collected with a twice-refocused GRAPPA DT-EPI sequence (TR = 4000–6000 ms depending on slice coverage, TE = 76.59 ms, FOV = 240 × 240 mm², acquisition matrix = 256 × 256, voxel size = 0.9375 × 0.9375 × 3 mm³) using a b-value of 1000 s/mm² sampling along 25 isotropically distributed diffusion directions. One additional volume was acquired at b = 0 at the beginning of each scan.

MRI Data Preprocessing

Image preprocessing was carried out as previously described.^9,10 In brief, image files were first converted from Dicom to NIFTI using dcm2niix¹¹, before dMRI data were preprocessed using the open-source software mrDiffusion (github.com/vistalab/vistasoft/mrdiffusion) implemented in Matlab R2021a (Mathworks, Natick, MA, USA). The b0 image was registered to the patient’s T1-weighted AC-PC aligned image. The combined transform that resulted from motion correction and alignment to the T1-weighted image was applied to the raw data (as well as the diffusion gradient tables), and the transformed images were resampled to 2 × 2 × 2 mm isotropic voxels. Diffusion gradient directions were then adjusted to fit the resampled diffusion data. Maps of fractional anisotropy (FA) and mean diffusivity (MD) were generated using the standard least squares method. Finally, the color FA maps (RGB) were visualized for quality control. Subjects with low RGB were excluded from analysis (n = 4 children).

Relative head motion was assessed by quantifying the number of volumes with translational motion of 1 voxel or more, relative to the prior volume. Subjects who deviated more than 2 standard deviations from the mean number of motion-affected volumes (M = 1.48, SD = 1.57) across the entire group were excluded from analyses (n = 4 subjects).

ROI placement, White Matter Tract Identification

Regions of Interest (ROIs) was performed on anatomic locations of inferior colliculus (IC) and superior olivary nucleus (SON) (FSLeyes, Wellcome Centre for Integrative Neuroimaging, University of Oxford; Figure 1). Other regions of the brainstem auditory pathway (i.e. trapezoid body, cochlear nucleus) were not assessed because of the inherent neuroanatomical difficulty with consistent and accurate ROI placement across subjects. Accuracy of the ROI placements were confirmed by consensus between a senior neuroradiologist and 2 pediatric otolaryngologists (KWY, AGC, INA). For white matter tract identification, ipsilateral tracts were generated between the IC and SON using a deterministic tractography technique (step size 1mm, angle > 30, FA > 0.15, length > 20mm). Scripts for this process were implemented in Matlab R2021a. Fiber renderings were then visually inspected for quality control. In the refinement process, fiber projections were discarded if they did not conform to known anatomical configurations of the white matter pathway between the IC and SON (IC-SON). While this process minimizes the capturing of stray fibers, deterministic tractography may still record false positive fibers for both the SNHL cohort and controls.¹⁸ Finally, diffusion values for bilateral ROIs and tracts were generated across the pathway, then averaged to calculate a mean FA and MD value for each subject.

Figure 1.

Representative regions of interest in a control subject

A: Sagittal, coronal, and axial images of the inferior colliculus (left-to-right)

B: Sagittal, coronal, and axial images of the superior olivary nucleus (left-to-right)

Statistical Analysis

All statistical analyses were conducted using the R statistical analysis program (R version 4.2, Vienna, Austria). The Shapiro-Wilk test was used to assess whether mean tract-diffusion indices were normally distributed. Paired t tests were used to examine whether mean tract-diffusion measures of the IC, SON, and IC-SON differed between the SNHL and control groups. Descriptive statistics for baseline characteristics were conducted. For continuous baseline characteristics, the Shapiro-Wilk Test was first used to assess normality of the data prior to using a paired t-test to compare the mean values between the two groups. For categorical baseline characteristics, Fisher’s exact test was used. For all tests in this study, results were considered statistically significant at p < 0.05.

RESULTS

Clinical Characteristics

Sixteen children with bilateral SNHL (8 males; mean age 5.3 ± 4.9 years) were compared against 16 age and sex-matched controls. The most common known hearing loss etiologies were connexin 26 mutations, bacterial meningitis, and cytomegalovirus. Fourteen children with SNHL (88%) had profound hearing loss in at least one ear. All other patient characteristics are displayed in Table 1.

Table 1.

Baseline Characteristics of Patients

Characteristics	Normal Hearing Control (N=16)	Bilateral Sensorineural Hearing Loss (N=16)
Age at Imaging (in years) a
Mean ± Standard Deviation	5.3 ± 4.9	5.3 ± 4.9
Sex – no. (%) b
M	8 (50.0)	8 (50.0)
F	8 (50.0)	8 (50.0)
Interval between Imaging and audiologic testing (in days) a
Mean ± Standard Deviation	.....	35 ± 24
Degree of Hearing Loss in Poorer Ear
Mild (20–40 dB)	.....	0 (0)
Moderate (41–55 dB)	.....	2 (12.5)
Moderately-severe (56–70 dB)		0 (0)
Severe (71–90 dB)	.....	0 (0)
Profound (91+ dB)	.....	14 (87.5)
Hearing Loss Etiology – no. (%)
Connexin 26	.....	2 (12.5)
Cytomegalovirus	.....	2 (12.5)
Meningitis	.....	2 (12.5)
Other/Unknown	.....	10 (62.5)
Audiologic Intervention
Bilateral cochlear implants	.....	9 (56.2)
Bilateral hearing aids	.....	3 (18.8)
Unilateral cochlear implant +/−   contralateral hearing aid	.....	4 (25.0)

^a.T-value = 0.251, P-value = 0.806

^b.Fisher’s exact test P-value = 0.999

Diffusion

We assessed the major brainstem nuclei of the central auditory pathway, the IC and the SON. Diffusion metrics for the IC, SON, and IC-SON fiber tract are shown in Figure 2. FA values were significantly decreased in the SNHL cohort compared to controls in the SON (0.377±0.056 vs 0.422±0.052; p=0.009), but not in the IC. There were no significant differences in MD values in IC and SON between the two groups. These results suggest that white matter integrity is preferentially compromised in the SON of children with SNHL.

Figure 2.

Diffusion metrics for inferior colliculus, superior olivary nucleus

FA: fractional anisotropy

MD: mean diffusivity

SNHL: sensorineural hearing loss cohort

Next, we conducted sub-analysis by age grouping (younger: ≤5 years, older: >5 years). Among younger children, MD values in the IC were significantly decreased in the SNHL cohort compared to controls (0.918±0.051 vs 1.120±0.142; p<0.001). However, among older children, there were no significant differences in MD values in the IC (1.124±0.198 vs 0.997±0.103; p = 0.119). Among younger and older children, there were no significant differences in FA values in the IC and SON. Fiber projections between the IC and SON were determined from tractography and we found no significant changes in FA or MD values (Figure 3). Diffusion metrics for this white matter tract are shown in Figure 4. There was no significant difference in MD or FA between the SNHL and control groups.

Figure 3.

Representative Fiber Tracts from Inferior Colliculus to Superior Olivary Nucleus (Sagittal Brain MRI)

Figure 4.

Diffusion metrics for white matter fiber tracts between inferior colliculus-superior olivary nucleus

FA: fractional anisotropy

MD: mean diffusivity

IC-SON: inferior colliculus – superior olivary nucleus

DISCUSSION

We utilized DTI to characterize microstructural alterations of the central auditory pathway in children with bilateral SNHL. To our knowledge, this is the first study to use tractography to delineate auditory fiber projections in the brainstem in the context of hearing loss. The findings not only represent imaging correlates of SNHL, but also demonstrate the potential utility of DTI for prognosticating long-term outcomes and determining intervention efficacy in these children.

Imaging plays a central role in the work up of pediatric hearing loss.^12,13 DTI is an MRI-based modality that assesses microstructural white matter changes in the central auditory pathway of the brain. In the literature, this pathway is of particular interest as it serves an integral role in auditory perception, as well as speech and language development.^14–16 Prior studies of DTI applied to the central auditory pathway in patients with SNHL have focused on the auditory radiations and the auditory cortex including the Heschl’s gyrus.¹⁰ Given that few studies have focused on brainstem regions, we aimed to address this gap in the literature.

To our knowledge, only one prior study assessed the brainstem nuclei of the central auditory pathway in children with bilateral SNHL using DTI. In that study, Huang et al. analyzed 24 children with bilateral profound congenital SNHL and reported diffusion metrics from six ROIs along the auditory pathway including the trapezoid body, SON, IC, medial geniculate body, auditory radiation, and Heschl’s gyrus.¹⁴ The authors reported decreased FA values without differences in MD in all ROIs among children with SNHL in comparison to controls. We also observed similar trends in the SON amongst the overall SNHL cohort. The decrease in FA, an index that indicates the directionality of water diffusion, suggests that myelin and axonal integrity of the SON is compromised in children with SNHL^19,20 In contrast with Huang and colleagues’ study, we did not find any differences in FA of the IC. The IC serves a crucial role as the integration center of the auditory pathway, and receives considerable ipsilateral and contralateral input. Given that orientation of crossing fibers can significantly attenuate FA, our results may reflect the incidental capturing of these fibers during ROI placement. Therefore, the SON may be a more reliable area to assess for true differences in white matter integrity of the auditory pathway between subjects.

The lack of differences in MD in both ROIs is not as well understood, though in theory, MD is negatively correlated with white matter integrity as it is often increased in brain areas where structural integrity is compromised. However, FA is the more consistent marker for white matter integrity.²¹ The results of our sub-analysis by age group revealed that among younger children (≤5 years), MD of the IC was significantly decreased in the SNHL cohort compared to controls, while no differences were observed among older children (>5 years). This may reflect the heightened vulnerability of the maturing auditory pathway in early childhood – a crucial period for myelination and synaptic pruning in the developing brain.²² Therefore as a promising neural correlate for auditory development, future studies should characterize these metrics in children with varying degrees of future progress.

Prior studies of tractography on the auditory pathway have mainly focused on characterizing fibers of the auditory radiations.^16,17 The novelty of our study stems from our application of tractography within the brainstem regions of the auditory pathway. This methodology involves extraction of fiber projections between the seeded brainstem regions rather than site-specific, localized ROI-based analysis. Interestingly, despite differences in FA of the SON in the SNHL cohort compared to controls, we did not find diffusion tensor differences of the IC-to-SON fiber projections between SNHL patients and controls. The implication of these results is not well understood. Previous animal studies have reported that projections between the IC and SON are unaffected upon cochlear removal or congenital deafness^23–25, with neural plasticity possibly playing a role.²⁶ Future studies are needed to replicate these findings.

We contribute to the current body of literature by investigating microstructural auditory pathway changes of the brainstem in children with SNHL. Future studies that correlate diffusion metrics to longitudinal hearing outcomes would help elucidate the use of DTI as a prognostic biomarker. Previous studies have attempted to correlate DTI to categories of auditory performance (CAP) scores^13,14,27,28, one measure that has been suggested as a proxy for cochlear implant (CI) outcomes.²⁹ However this is an area of the literature in need of high-quality prospective studies, as well as the development of a standardized metric to assess CI outcomes. Given that the acquisition, pre-processing, and analysis of MRI data are highly protocolized, we anticipate that these techniques will be widely adopted across institutions to study hearing loss and rehabilitation.

The present study has several limitations. First, our cohort consisted of children with SNHL of varying degree, etiology, and onset which may affect the generalizability of these findings. However, we are reassured by the trends in diffusion metrics amongst our children with SNHL that are consistent with those of previous studies. In addition, each subject in our study was imaged at a single time point. Due to the cross-sectional nature of the study, we assumed that diffusion metrics from the single time point were representative for each patient, and we could not assess longitudinal changes in diffusion. Furthermore, reconstruction of contralateral tracts was difficult due to the orientation of crossing fibers³⁰, and therefore we limited our fiber tracking to ipsilateral IC-SON fibers. Future studies are warranted to characterize and understand the impact of crossing fibers, such as those from the contralateral cochlear nucleus and IC, as well as the contralateral SON and IC. Finally, our limited sample size underscores the need for future studies replicate these findings with a larger number of children with SNHL.

CONCLUSION

In this study, we report significant white matter microstructural differences in the SON among children with bilateral SNHL. In addition, we reconstructed and quantified diffusion along the fiber tracts between the IC and SON. This pilot study highlights the plausibility and utility of DTI and tractography to characterize brainstem regions of the auditory pathway in the context of pediatric SNHL. Longitudinal studies are warranted to correlate these findings with long-term outcomes and intervention efficacy.