Background
Approximately 90% of autistic children [
1,
2] and 5–15% of non-autistic children exhibit elevated sensory features, characterized by enhanced or reduced reactivity to or interest in sensory stimuli [
3‐
5]. (Please note, identity-first language is used in alignment with the majority preference of the autistic community [
6,
7].) These elevated sensory features are associated with decreased motor performance [
8], increased core autism traits [
9,
10], increased anxiety [
11,
12], decreased adaptive behaviors [
13], and decreased quality of life [
14,
15]. Currently, the neurobiological mechanisms of sensory features in autistic and non-autistic populations are unclear [
16]. However, the underexplored brainstem may offer critical insights into the neural basis of sensory features due to its established role in sensory processing [
17,
18] and associations with core autism traits [
19,
20]. Therefore, the purpose of the present study was to examine the relationship between brainstem microstructure and sensory features in autistic and non-autistic children.
The brainstem is an early developing and highly conserved [
21] structure that is comprised of tightly intertwined white matter tracts, many of which have been linked to sensory processing. Brainstem white matter fibers serve as initial conduits of sensory information, relaying signals from primary sensory organs to nuclei within the brainstem, cerebrum, and cerebellum [
17,
22‐
24]. Brainstem white matter tracts further support basic sensory information processing by transmitting sensory signals among nuclei with demonstrated roles in sensory gating [
25,
26], visual salience [
27], multisensory integration [
28], and sensory responsivity [
29,
30]. While much of this work has been done in animal models, similar associations in humans have been established between brainstem white matter and the early stages of sensory processing [
26,
31‐
33]. Moreover, early developing brainstem pathways are known to subserve early-in-life auditory, visual, gustatory, olfactory processing as well as tactile-motor integration (as reviewed by [
18]). However, it remains unclear how brainstem white matter is related to sensory responses in autism. Despite the brainstem’s demonstrated role in the fundamental elements of sensory processing, previous work looking at the neural contributions to sensory response patterns has largely focused on telencephalic structures as key regions of interest [
34‐
39]. Therefore, we still do not know whether brainstem white matter contributions are limited to relaying and processing basic sensory information or extend into producing heightened or reduced sensory responses.
Evaluating sensory features and their relationships to brainstem microstructure in autistic populations is critical as evidence indicates that brainstem white matter may uniquely contribute to autism [
19] and elevated sensory features are highly prevalent in the autistic population. Epidemiological, molecular, and behavioral evidence suggests that brainstem organization may be closely tied to the etiology of autism [
18‐
20,
40]. Indeed, one of the earliest hypotheses regarding the neural basis of autism centered upon the brainstem’s reticular formation [
41]. More recently, several articles have reviewed the evidence of the brainstem’s role in autism and have put forth theories about how the structure, function, and development of brainstem white matter tracts and gray matter nuclei may be involved in autistic traits [
18‐
20]. Additionally, an exploratory analysis from Wolff and colleagues [
42] linked sensory features to brainstem-cerebellar white matter, finding that infants who later received an autism spectrum diagnosis showed inverse sensory-microstructure correlations compared to infants who did not receive a diagnosis. These diagnosis-dependent neural correlates of sensory features in autism are supported by evidence suggesting the presentation of sensory features and their neurobiological bases may be unique in autism compared to non-autism and/or other psychiatric conditions [
43]. For example, evidence suggests that sensory hyporesponsiveness in autistic populations may be unique in both its behavioral presentation and neural basis. Hyporesponsiveness is more prevalent in autistic individuals than in other populations [
36,
44‐
47] and has been associated with altered patterns of neural activity in infants with and without a predisposition for autism [
48]. This evidence coupled with the distinct contributions of the brainstem to autism traits [
19] highlights the need for a direct comparison of brainstem neural correlates in autistic and non-autistic youth. This direct comparison will determine not only how the brainstem is involved in sensory processing but also if its involvement is similar or distinct in autistic and non-autistic populations. Previously, methodological constraints limited the feasibility of collecting high resolution diffusion-weighted imaging (DWI) data (traditionally a time consuming and sensory intensive process) in pediatric populations with sensory features. However, recent advancements in our DWI protocol have allowed us to overcome these limitations, providing high apparent resolution and improved gray-white matter contrast without requiring long acquisition times [
49]. These innovations offer the opportunity to investigate white matter microstructure of brainstem tracts in children with elevated sensory features with a higher degree of precision than ever before.
Using our optimized DWI, the aim of this study was to determine the extent to which brainstem white matter tracts are associated with individual differences in the sensory features of autistic and non-autistic children (6–11 years of age). Even though the brainstem begins to form in the first trimester of pregnancy [
50], there is evidence that the brainstem tracts subserving vision undergo activity-dependent myelination based on sensory stimulation in the first year of life [
51]. Auditory, olfactory, tactile, and gustatory brainstem tracts are likely to similarly undergo post-natal tuning based on sensorimotor experiences [
18,
20]. This experience-based tuning may lead to cascading white matter differences in school-aged autistic children and beyond. Therefore, this age range was selected for feasibility of collecting the MRI parameters and with the idea that differences in early-maturing brainstem circuits may continue to subserve the sensory features commonly reported in autistic children [
18]. Based on literature from animal models and humans involving brainstem white matter in basic sensory processing, we hypothesized that brainstem microstructure as measured by diffusion MRI (free-water-eliminated fractional anisotropy [FWE-FA] and mean diffusivity [FWE-MD]) would be related to the presence of elevated sensory features in both autistic and non-autistic children. While other DTI measures are possible to calculate, we chose FWE-FA and FWE-MD based on FA and MD findings of previous literature [
42] and evidence of reduced artifacts in brain areas surrounded by cerebrospinal fluid (CSF) when using FWE [
52,
53]. While FWE-DTI measures do not directly measure microstructure, they are commonly used as markers sensitive to changes in white matter microstructural features, including axonal morphology and myelination, axon bundle density and fiber orientation distribution, and other intra- and extra- cellular processes. Based on the evidence suggesting unique brainstem involvement in autism [
19] and a diagnosis-dependent relationship between sensory features and white matter microstructure [
42], we further examined the possibility that sensory-brainstem relationships would be unique within each diagnostic group. To test these hypotheses, we performed region of interest (ROI) linear regression predicting FWE-FA and FWE-MD of brainstem tracts from sensory caregiver report and diagnostic group status, while controlling for key variables such as age, sex, and head motion during the DWI scan. A significant main effect for sensory features would support that the FWE-FA and FWE-MD of the brainstem white matter tracts are significantly associated with caregiver-reported sensory features across groups. A significant group-by-sensory interaction would support our hypothesis of unique brainstem-sensory relations in autistic compared to non-autistic children. Follow-up analyses explored these effects within the autistic group as a function of sensory pattern and sensory modality. To contextualize brainstem findings, follow-up, whole-brain voxel-based correlates of sensory features were assessed across both groups and within just the autistic group.
Discussion
This study set out to identify the relationships between sensory features and white matter microstructure of the underexplored brainstem in autistic and non-autistic children. Using a novel DWI protocol that improved the apparent resolution of the brainstem and cerebellum [
93], we precisely delineated brainstem and brainstem-cerebellar white matter tracts and examined their associations with total sensory features and specific sensory responses. Consistent with our hypotheses, results revealed that the microstructural properties of brainstem white matter tracts were associated with sensory features, particularly in autistic children. Together, with previous animal literature [
23‐
28], this finding suggests that brainstem white matter contributions are not limited to relaying and processing basic sensory information, but that they extend into producing heightened or reduced sensory responses in autistic children. A follow-up whole-brain analysis demonstrated proportionally more of the sensory-brain relationships in autism occurred in the brainstem and cerebellar white matter than what we would have expected based on the size of the search area. These brainstem/cerebellar findings were contextualized by additional brain-sensory findings in white matter areas of the visual cortex, inferior parietal cortex, primary motor and sensory cortices, and thalamic radiations, all areas known to be associated with sensorimotor processing. Further, in autistic children, sensory hyporesponsiveness and tactile sensitivities were associated with white matter microstructure in nearly all brainstem tracts. These findings and their implications are discussed below.
Our study findings suggest that the brainstem plays a role in autistic children's behavioral responses to sensory stimuli. These relationships between brainstem white matter microstructure and sensory features were diagnosis-dependent and extend previous exploratory findings [
42] of inverse relationships between sensory features and brain microstructure in autistic children compared to non-autistic children. These results offer intriguing insights into the potential biology underlying microstructural development of the brainstem in autism. In both the current and previous [
42] studies, lower MD in the MCP and SCP were associated with more severe sensory features in autistic children but not in non-autistic children. Yet, developmental trajectories of the MCP and SCP from previous work appear to be similar in autistic and non-autistic children, with both diagnostic groups showing similar decreases in MD with age [
94]. Together, this information suggests a potentially altered mechanism for sensory responsiveness in autism that heavily depends on brainstem white matter. Specifically, while lower MD is commonly interpreted as indicative of more developed (i.e., more dense and more organized) white matter tracts, present findings suggest that lower MD of the brainstem, cerebellum, and other cerebral areas, may relate to more prevalent sensory features in autistic children. This autism-specific relationship may be indicative of increased brainstem involvement in sensory responsiveness in autistic youth. It may also suggest that higher efficiency information transfer among brainstem sensory processing nuclei can lead to more prominent sensory features in autistic youth. However, MD is an indirect measure of microstructural organization and can be influenced by multiple biological factors [
95]. Therefore, further research is needed to determine the precise cytoarchitectural basis of these brainstem-based relationships, using innovative and complementary quantitative MRI strategies [
96,
97] that provide additional information about cellular properties of white matter.
The moderate-sized relationships between hyporesponsiveness, defined as a reduced behavioral response to stimuli in the environment, and multiple brainstem structures have implications for how we conceptualize and support diverse sensory features in autistic children. The distinct brainstem-hyporesponsiveness relationships in the autistic compared to the non-autistic groups suggest that: 1) hyporesponsiveness in non-autistic children may be neurobiologically distinct from hyporesponsiveness in autistic children in ways that current behavioral measures may not distinguish, or 2) hyporesponsiveness in autistic and non-autistic children may be an example of multifinality, in which differing neurobiological etiologies lead to similar behavioral symptoms. In either scenario, the associations among brainstem microstructural features and hyporesponsiveness in autism underscore the reflex-like orienting of hyporesponsiveness [
44] and help to recontextualize the self-reports of autistic individuals [
98,
99] where behavioral responses to sensory stimuli are reported to feel outside of volitional control. Therefore, therapies that use external reward or punishment to target sensory features may be unlikely to be successful as they assume volitional control and are unlikely to target the brainstem-based neural circuitry that may underlie sensory hyporesponsiveness in autistic individuals. Previous research demonstrated that a six-week biofeedback-based training in autistic and non-autistic adolescents induced treatment-specific changes to the superior cerebellar peduncle [
88], a region found to be associated with sensory features in both the present study and Wolff et al
. [
42] Therefore, there is preliminary evidence of brainstem microstructural changes in response to a multi-week intervention. Used in the context of sensory interventions, future studies should track brainstem changes in relation to intervention-related decreases in sensory features.
The present findings also suggest that brainstem white matter may be particularly related to tactile responsivity in autistic individuals, with eight of the 11 brainstem tracts moderately related to responses to touch. Tactile sensitivity has been commonly reported in autistic individuals [
100‐
103], and reduced tactile responsivity at 12 months was found to be an early predictor of a later autism diagnosis [
104]. Furthermore, the inferior olivary nucleus (ION) in the upper medulla aspect of the brainstem is associated with integration of tactile sensations with motor responses and has been previously found to have atypical structure in postmortem brain analysis of autistic individuals [
105‐
107]. The ION receives numerous brainstem and cerebellar inputs (as reviewed in [
18]) and outputs to the cerebellum via portions of the inferior cerebellar peduncle. Therefore, it is possible that the early-developing brainstem is implicated in tactile experiences of autistic school-aged children in ways that involve the ION. However, future research will be needed to confirm and further examine this relationship, particularly given that the present sensory measure cannot disentangle pain, pressure, and vibration. Fortunately, enhanced imaging of the brainstem may enable elucidation of the size, shape, and microstructural properties of specific brainstem nuclei, like the ION, in future in vivo studies of autistic children and adults.
The follow-up whole-brain analyses further contextualized the present sensory-brainstem findings, by showing that sensory features in autistic children were also related to cerebral white matter in brain areas frequently associated with sensory processing, including the occipital cortex (vision), inferior parietal cortex (audition), primary motor and somatosensory cortices (touch and proprioception), and thalamic projections (multisensory relay). One interpretation of these results is that the brainstem findings are reflective of a whole-brain sensory phenomenon, whereby decreased mean diffusivity is related to more sensory features in autistic children. However, our results also suggested that brainstem and cerebellum findings are overrepresented with respect to the size of the search space, suggesting that the brainstem and cerebellar white matter tracts may play a strong role in the sensory experiences of autistic individuals. These findings are compatible with the brainstem’s involvement in prenatal development of the cortex ([
108‐
110]) and the cascading effects on the brain that prenatal brainstem differences combined with ongoing sensorimotor tuning may have [
18,
20]. However, longitudinal studies, ideally from early prenatal development into the first few years of postnatal development, will be needed to determine the exact role of the brainstem and cerebellum in sensory processing and overall brain development. In all, the present findings, combined with theoretical work and studies implicating the brainstem in autism [
18‐
20], suggest that the brainstem and cerebellum may be integral contributors to the sensory experiences of autistic individuals. Therefore, even though the imaging of the brainstem may require special acquisition and processing procedures [
49], including free water elimination, EPI distortion correction, and careful consideration of brainstem masking, these steps are worth taking, as the brainstem and cerebellum are likely key areas to study to better understand the neurobiological basis of the autistic experience.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.