Disturbances of sensory responsivity are commonly associated with diagnosed autism, and emerging evidence suggests potentially also with ADHD. However, critical questions remain as to whether these differences emerge early in development – suggesting causality – and the neurobiological mechanisms that may drive variation in sensory responsivity. We present analyses of a prospective longitudinal cohort of infants, followed from 10–36 months. Results showed that a family history of autism (FH-autism) was associated with steeper increases in hyper-responsivity. Conversely, a family history of ADHD (FH-ADHD) was associated with increasing hypo-responsivity. A higher intercept (equivalent to estimated level at 10-month baseline) of hypo-responsivity was associated with being male, and having higher later autistic and ADHD traits at age 3 years; a steeper slope of hypo-responsivity was also associated with higher ADHD traits. Alterations in markers of global E/I balance, potentially indicative of enhanced inhibition, were associated with increased hyper-responsivity (and there was evidence to suggest increased hypo-responsivity also).
Our predictions that FH-autism infants would show increased sensory hyper- and hypo-responsivity, and these increases would in turn be associated with increases in emerging autistic traits, was partially supported. FH-autism infants displayed a steeper increase in hyper-responsivity over the first three years of life, however, hyper-responsivity was not associated with later autism traits. These results contrast with meta-analyses reporting that sensory over-responsivity is best able to distinguish autism from other clinical groups 3, and reports that FH-autism infants who receive a diagnosis of autism score higher on hyper-responsivity at 12 months of age 11. What did predict later autistic (and ADHD) traits was increased hypo-responsivity at 10-month baseline. It is of interest that the form of sensory responsivity associated with a family history of autism (hyper-responsivity) is different to that which relates to later autistic traits (hypo-responsivity), and calls into question the centrality of hyper-responsivity as a mechanism for core autism characteristics. Increased responsivity has been associated with anxiety in both autistic 36 and typically developing 6 children, and thus could in-part explain the increased prevalence of mental health difficulties in autistic populations. The association with hypo-responsivity is consistent with reports that hypo-responsivity is more prevalent in autism in early childhood 37, and predicts later autism symptoms in infant-siblings 38. Early hypo-responsivity may have knock-on effects on cognitive development; hypo-responsivity is associated with lower levels of joint attention and language 39, decreased neural response to social stimuli and fewer social approach behaviours 40. Notably, in most previous studies, associations with hyper- and hypo-responsivity are tested in separate models and without examining co-occurring conditions, limiting inference about specificity. One prospective community study that included both types of sensory responsivity in the same model reported both hyper- and hypo-responsivity were associated with later autism outcomes 28, although this work covered a wider developmental window (6 months – 7 years) and did not assess ADHD traits. Current results suggest that increased hypo-responsivity may not be a specific marker for later autistic characteristics, but also relevant for emerging ADHD behaviours (and thus may function as a shared neuroendophenotype).
In addition to the aforementioned associations between baseline hypo-reactivity and ADHD traits, a family history of ADHD was associated with steeper slope of hypo-responsivity (greater increase between 10–36 months), which in turn was associated with higher ADHD traits at age 3. This result extends findings from meta-analyses that report broad alterations in sensory processing in the early years are predictive of ADHD outcomes 4. The lack of association with hyper-responsivity goes against previous reports 14,41, although many of these studies did not account for the co-occurrence of ADHD and autism, or were conducted in older samples. One hypothesis is that ADHD traits are a form of compensation for early hypo-responsivity, such that the child learns to sample the environment more rapidly and randomly to overcome for a decreased response to external inputs. However, it could also be that increases in parent-reported hypo-responsivity are a reflection off increasing differences in attention control that are characteristic of ADHD, which in turn lead to an apparent lack of responsivity as cues are ‘missed’. Finally, we found males were rated as more hypo-responsive than females (as found in general population cohorts 28), emphasising the need to account for sex in analytic models.
In terms of putative neurobiological mechanisms, our pre-registered hypothesis on the role of E/I balance was supported, such that increased aperiodic exponent values, indicative of enhanced inhibitory activity, were associated with greater hyper-responsivity. However, our pre-registered plan had specified analysis of EEG data at 14 months, whereas we found significant effects at 5- and 10-months only. Thus, current results require replication. Results concur with other infant-sibling studies that find brain or cognitive features are more strongly predictive of later autism outcomes when measured earlier in development 22,23, and genetic research which finds critical periods early in development for SCN2A mutations to produce an autistic phenotype42. Crucially, results demonstrate the role of alterations in E/I balance early in the lifespan, potentially before initial effects become confounded by developmental compensation 24 or later emerging moderating factors 43. We also note the association between 5-month aperiodic exponent and hypo-responsivity was at significance (p = .054), and the standardized coefficients of effect were comparable for hyper and hypo-responsivity (B = .26 and .25), suggesting enhanced inhibition in early development may alter sensory processing at a system-wide level 15. Possible mechanisms of effect include disruption in evoked gamma oscillatory activity, which in turn alters the temporal sharpening of cortical sensory responses 44.
One limitation is the reliance on parent ratings for behavioural phenotypes, which could be impacted by rater biases (e.g., sex effects could in part reflect gendered expectations of child behaviour). Future studies with greater diversity of ethnicity and socio-economic status are needed to understand the generalisability of results.
To summarize, increased hypo-responsivity may function as a general marker for later neurodivergence. Early alterations in E/I balance may contribute to differences in childhood sensory responsivity, highlighting a putative neurobiological mechanism that could inform assessment and intervention. Future research should explore how differences in sensory processing may arise from alterations in E/I balance, and in turn how these translate to observable differences in responsivity.