On February 10, SFARI welcomed attendees to New York City for an in-person meeting focused on measuring sensory system function in humans with autism spectrum disorder (ASD) and other neurodevelopmental disorders (NDD). Nine researchers presented their work to over 30 attendees on different aspects of sensory function: how to measure it, biomarkers for it, different levels of sensory processing and specific domains that are altered in ASD.

Since the publication of DSM-5 in 2013, “hyper- or hypo-reactivity to sensory input or unusual interest in sensory aspects of the environment” has been among the diagnostic criteria for autism. People with ASD can show signs of hypersensitivity, such as when a seemingly innocuous stimulus is aversive to them, as well as hyposensitivity, such as when a salient stimulus does not evoke a response. These symptoms often make it difficult to engage in day-to-day activities.

In her opening remarks, SFARI executive vice president Kelsey Martin said that getting a handle on sensory system function in ASD could help parse the heterogenous ASD phenotypes, provide outcomes and endpoints for clinical trials and inform animal models for back-translational work.

“We hope to start a conversation about these intertwined issues,” Martin said. “Eventually an intervention that addresses sensory function may alleviate some of the challenges people with autism face.”

Assessment of sensory function has a place both in clinical and pre-clinical work. According to meeting co-organizer Paul Wang of Clinical Research Associates (a subsidiary of the Simons Foundation), clinical trials are beginning to incorporate measures of sensory function. Fellow co-organizers David Ginty of Harvard University and Lauren Orefice of Harvard Medical School and Massachusetts General Hospital outlined how sensory function presents a rich venue for animal model studies with translational potential: Ginty and Orefice have found touch hypersensitivity in mice carrying genetic mutations related to ASD, even when the mutations are restricted to the peripheral nervous system¹. Notably, these mice also develop anxiety-like behavior and aberrant social behavior, which can be rescued through the administration of a peripherally restricted GABA_A receptor agonist². Ginty and Orefice are now advising an industry-led effort to develop GABA-A receptor agonists that do not cross the blood-brain barrier as a therapeutic strategy to treat select ASD-related behaviors. “These findings illustrate the important role of rodent studies in understanding the circuit mechanisms underlying behavior, which can have translational impact,” said Brigitta Gundersen, a meeting co-organizer. However, bringing these therapeutics from bench to bedside will require developing better behavioral measures to assess sensory differences in humans with ASD, which was the focus of this workshop.

The meeting goals included discussions about the different measures of sensory system function; the types of data obtained from different assessments of sensory system function; how robust measures of sensory function are as a function of age, cognition (IQ) and verbal ability; what the appropriate time period for measuring changes in sensory symptoms would be; how changes in sensory function could relate to clinically meaningful changes; and identifying sensory measures that can be assessed at scale to leverage large, well-characterized cohorts such as SPARK and Simons Searchlight.

Paige Siper of the Icahn School of Medicine at Mount Sinai described an assessment that she co-developed called the Sensory Assessment for Neurodevelopmental Disorders (SAND)³. Taking 20 minutes to administer, the SAND combines clinician observation with a caregiver interview to quantify sensory hyper-reactivity (adverse response to a stimulus), hypo-reactivity (indifference or unresponsiveness to a stimulus) and sensory seeking (excessive interest in a stimulus) across visual, tactile and auditory domains. It has been validated in individuals 2 to 12 years old, including those with low mental age and those who are non-verbal and show high test-retest reliability.

With the SAND, Siper and colleagues find atypical sensory reactivity across different monogenic neurodevelopmental syndromes, with some notable differences: for example, hyporeactivity for visual, auditory and touch stimuli is more prominent in Phelan-McDermid Syndrome (PMS) than in idiopathic ASD, which shows more hyperreactivity and sensory seeking⁴. SAND scores also correlate with commonly-used behavioral measures such as the Vineland Adaptive Behavior Scale, which suggests links between atypical sensory responses and other behaviors. The SAND is currently being used in early-stage clinical trials for monogenic ASD-related disorders.

Several researchers are exploring physiological characteristics to identify objective measurements of sensory function. Electroencephalogram (EEG) signals are a potentially rich source of biomarkers for sensory function because they reflect temporal dynamics of neural information processing, are low cost and can be acquired from almost all populations. Such biomarkers could be useful for clinical trials in humans as well as for refining animal models.

There are many potential biomarkers that measure different aspects of sensory function. April Levin of Boston Children’s Hospital highlighted types of brain activity that could modulate responses to stimuli in ways that could explain the “sensory paradox” observed in ASD, in which individuals sometimes exhibit both hyper- and hypo-responsivity. One involves endogenous brain activity that can modulate perception of a stimulus, such as transient beta activity⁵. Another relates to the dynamic range available to incoming sensory input, which seems narrower in ASD, and could be due to alterations in feed-forward inhibition. A third involves the timing of incoming signals, which could evoke habituation or facilitation that is dependent on cortical interneurons. The fourth involves how the brain integrates simultaneous stimuli, whether combining them to get a gestalt, or focusing on details of individual stimuli.

Guided by the sensory hypersensitivity commonly found in fragile X and ASD, Lauren Ethridge of the University of Oklahoma presented her work on EEG biomarkers of auditory function in these conditions. One paradigm uses an auditory “chirp” stimulus that varies in frequency to evoke neural oscillations that reflect both auditory cortex function and network organization of the brain as a whole. The paradigm is well-tolerated, and does not require specific behavioral tasks or attention. An increase in gamma power and a decrease in synchronization is observed in fragile X subjects relative to controls⁶. Similar effects are observed in mouse models of fragile X, and experiments suggest that the degraded cortical phase-locking originates in subcortical structures⁷.

In the visual domain, Sophie Molholm of Albert Einstein College of Medicine described experiments that collect visual evoked potentials (VEP) via EEG to get a reliable readout of visual system function. Responses to a train of visual stimuli produce a signal that differs in those with ASD: though the initial response resembles that of controls, the late response diverges, which suggests a difference in top-down anticipation of a stimulus as opposed to bottom-up sensory processing⁸. They used similar neural measures to report differences in Rett syndrome using auditory stimuli⁹.

In his talk, Nicolaas Puts of King’s College London hypothesized that sensory difficulties could contribute to a sense of unpredictability in the environment, which in turn could promote anxiety and rigid behavior. Puts presented research on early-stage sensory processing based on a tactile paradigm to explore touch discrimination between fingers¹⁰. This has identified differences among people with ASD and attention deficit hyperactivity disorder (ADHD); specifically, people with ADHD and people with a co-diagnosis of ADHD and ASD show higher detection and discrimination thresholds¹¹. These tactile differences were higher in those scoring higher for sensory reactivity and correlated with symptom severity. These findings suggest that measures of sensory perception could serve as biomarkers of sensory responsivity. In addition, elevations of glutamate and glutamine in somatosensory cortex, as measured with magnetic resonance spectroscopy (MRS), correlate with these tactile characteristics and sensory symptoms in ASD, which suggests these symptoms are driven by increased excitation in the brain¹².

Shulamite Green of the University of California, Los Angeles highlighted how the same behavioral response to a stimulus may reflect different underlying processes: for example, a lack of an aversive response to an unpleasant stimulus may reflect typical sensory hypo-reactivity, or it may come from strong regulatory processes that suppress a behavioral response despite the presence of biological and/or emotional reactivity. Using functional magnetic resonance imaging (fMRI), Green has found evidence for both kinds of processes in ASD: imaging shows stronger activation in sensory cortices and amygdala in some youth with sensory over-reactivity, as well as increased connectivity between the prefrontal cortex and amygdala¹³ in other youth who show fewer behavioral responses. Green has also been exploring biomarkers such as skin conductance and heart rate as readouts of sensory function and has so far found that heart rate elevates in response to mildly aversive stimuli in those with ASD who show sensory over-responsivity¹⁴.

Pawan Sinha of the Massachusetts Institute of Technology has proposed that the multi-faceted symptoms of ASD, including sensory anomalies, could be due to an underlying impairment in temporal prediction by the brain¹⁵. Experiments probing habituation, which may rely on prediction, have found it to be reduced in ASD in both the visual and auditory domains, as measured by EEG¹⁶, MEG and skin conductance¹⁷. Habituation also correlates with sensory symptoms, and Sinha suggested that reduced habituation in ASD might contribute to the obsessive interests and repetitive behaviors observed in ASD.

Evdokia Anagnostou of Holland Bloorview Kids Rehabilitation Hospital Toronto guided a discussion of the day’s proceedings. She and others reinforced the idea that the field needs consensus on how to classify sensory behavior, and on definitions of commonly used terms, such as sensitivity, reactivity and responsivity. For example, “sensitivity” relates to sensory thresholds for detection; “reactivity” relates to autonomic bodily responses to stimuli, such as heart rate and skin conductance; and “responsivity” encompasses top-down processes such as attention, environmental context and anticipation, which combine to provoke a response or not.

Anagnostou also emphasized there may be a difference between the basic biology of sensory processing and what causes distress to people with ASD. While dissecting differences in sensory processing may be important for stratification of ASD for clinical trials, these differences might not be related to the sensory discomfort that interventions seek to mitigate.

Several attendees spoke of the multiple stages of sensory processing, and the uncertainty about where in the pathway the sensory anomalies in ASD originate — whether in the earliest steps in sensory detection, or in later stages in which the brain formulates a motor response. Others noted that the heterogeneous measures of sensory function currently reported for ASD may in part reflect the many different paradigms currently in use by different labs. Also, current measures of sensory function in ASD ignore olfaction and gustation; finding measures for these sensory modalities could fill out the picture of sensory differences in ASD, and help with back-translation to animal models, as olfaction is very important to mice.

In the future, researchers could work toward assembling a sensory “toolbox” that consists of methods to measure aspects of sensory function consistently across labs.

1. Orefice L.L. et al. Cell 166, 299-313 (2016) PubMed
2. Orefice L.L. et al. Cell 178, 867-886 (2019) PubMed
3. Siper P.M. et al. Autism Res. 10, 1133-1140 (2017) PubMed
4. Tavassoli T. et al. Genes (Basel) 12, 977 (2021) PubMed
5. Law R.G. et al. Cereb. Cortex 32, 668-688 (2022) PubMed
6. Ethridge L.E. et al. Front Intgr. Neurosci. 13, 60 (2019) PubMed
7. Holley A.J. et al. Neurobiol. Dis. 170, 105767 (2022) PubMed
8. Beker S. et al. J. Neurophysiol. 126, 1783-1798 (2021) PubMed
9. Sysoeva O.V. et al. Transl. Psychiatry 10, 188 (2020) PubMed
10. Puts N.A.J. et al. J. Neurosci. Methods 18, 239-247 (2013) PubMed
11. He J.L. et al. Commun. Biol. 4, 97 (2021) PubMed
12. He J.L. et al. Transl. Psychiatry 11, 411 (2021) PubMed
13. Green S.A. et al. JAMA Psychiatry 72, 778-786 (2015) PubMed
14. Jung J. et al. J. Child Psychol. Psychiatry 62, 1183-1194 (2021) PubMed
15. Sinha P. et al. Proc. Natl. Acad. Sci. U.S.A. 111, 15220-15225 (2014) PubMed
16. Jamal W. et al. J. Autism Dev. Disord. 51, 3153-3164 (2021) PubMed
17. Gandhi T.K. et al. J. Autism Dev. Disord. 51, 2218-2228 (2021) PubMed