This study investigated modality-specific sensory difficulties and their relative importance in explaining variance in specific QAT in the adult population. High scores on the auditory modality subscale were strongly predictive of traits in all three domains, suggesting that it may be a robust endophenotype that co-segregates with the broad autism phenotype. In contrast, high tactile and proprioceptive scores specifically predicted social and communicative traits, respectively. The absence of associations for olfactory, gustatory, vestibular, and visual modalities must be interpreted in light of the moderate internal reliabilities for the GSQ subscales (Table 2). However, it might suggest that deficits in these modalities are inherited differently or are associated more specifically with autism as a diagnostic category rather than QAT.
The dominance of the auditory modality is consistent with the known presence of sound sensitivity in autistic populations and first-degree relatives [16, 17, 19], but it does not give any mechanistic explanation. The GSQ auditory subscale probes multiple auditory functions, including aversion to specific, loud, or unpredictable sounds, attraction to specific sounds, and difficulties with speech perception [8]. These difficulties are prevalent in autism but are likely to engage disparate mechanisms, such as auditory brainstem abnormalities, sensory gating deficits, disturbed central gain control, or broader networks involved in emotional reactions to sounds. Subjective auditory sensitivity did not correspond to altered thresholds for detection or discrimination [33, 34], but the behavioral threshold for discomfort or startle is often lowered in autistic populations (reviewed in O’Connor, [35]). A meta-analysis of an extensive experimental literature on early auditory evoked activity recorded with electro- or magnetoencephalography found group differences in the earliest auditory responses reflecting processing in the primary and secondary auditory cortices [36]. Some aspects of speech perception, such as perception of speech in noise and attentional orienting to speech sounds, may be impaired in autistic individuals and might involve atypical hemispheric lateralization [37]. Specific questions about mechanisms and heritability can be addressed by combining neurophysiological measurements of responses to controlled stimuli and inclusion of first-degree relatives in the study design.
While this study did not address causal relationships between sensory symptoms and QAT, atypical auditory processing in early development may contribute to atypical development of higher-order functions impaired in autism. Auditory difficulties in adulthood may also have direct consequences on social, communicative, and rigid symptoms, e.g., by engaging behavioral homeostatic mechanisms such as rigid adherence to routines or social avoidance to avoid perceptual overload. Self-reported noise sensitivity to a wide range of environmental sounds is considered a stable personality trait and is a significant predictor of individual adverse reactions to sounds [38, 39]. Noise-sensitivity was also found to be correlated with the introversion dimension on the NEO Personality Inventory, which taps a similar construct to BAPQ’s Aloof subscale [40, 41]. Similarly, extraversion on the Eysenck Personality Questionnaire was reported to be negatively correlated with noise annoyance [42, 43]. Thus, atypical auditory processing might be a transdiagnostic trait that contributes to disability rather than being specific to autism [44].
Atypical tactile processing was related to higher social QAT, which involves decreased social motivation and less enjoyment of social interactions. This is consistent with some studies exploring associations between social introversion and social touch. For example, self-reported aversion to social touch was found to correlate positively with total QAT [45–47], and parent-reported tactile hypersensitivity predicted an autism diagnosis in children [48]. Further, avoidance of social touch was negatively correlated with extraversion on a personality inventory [46]. The items on the tactile GSQ scale were designed to capture a range of deficits common in autism, ranging from atypical pain, temperature, and touch processing, to disliking haircuts, clothes labels, or hugs. While these items load onto the same latent factor, at least with moderate reliability, it seems unlikely that deficits in these domains are mediated by one mechanism. Tactile detection and discrimination thresholds might be altered in some autistic participants and might be related to E/I imbalances, but findings have been mixed [49–51]. An affective touch fMRI paradigm (slow versus fast stroke) found negative correlations between QAT and BOLD-responses in the superior temporal sulcus and orbitofrontal cortex, suggesting a role for CT-afferents and social brain networks [47]. Another potential mechanism is atypical autonomic reactivity, which was found in response to touch in autistic adults with normal tactile thresholds [49].
We found the proprioception subscale to be a stable and robust predictor of communication QAT across all analyses. It was the only modality-trait relationship that appeared to be as dominant as auditory dysfunction. The proprioception subscale of the GSQ comprises items concerning fine motor skills, interoceptive awareness, and perception of peri-personal space (e.g., standing too close to or too far away from someone), and thus probes a broader sensorimotor construct than its name suggests. Some of the items of the pragmatic subscale of the BAPQ either concern language production (e.g. “I find it hard to get my words out smoothly”/””, or “I speak too loudly or softly”) in addition to higher-order pragmatic language skills. Therefore, further studies should address whether the relationship is specific to motor function or reflects a broader association with higher language functions. In studies of infants, oral and fine motor skills have been found to predict later language capabilities [52–56], suggesting that basic motor development is a prerequisite for higher functions to develop. Our study preliminarily suggests that this association persists into adulthood. Consistent with this, a study on adults in a naturalistic conversation setting found that autistic participants demonstrated lower lexical diversity and produced fewer mouth movements [57]. On a higher cognitive level, a neural overlap was found between syntactic processes and tool use in the basal ganglia, as well as a bidirectional cross-domain transfer of learning between these two skills [58], raising the possibility of motor training to improve language development.