In this taxonomy, we propose a hierarchical terminology with levels. We use the term levels to refer to “levels of analysis” as used in the social sciences to place research targets in a specific scale (e.g. micro-, and macro-levels of analysis) or domain. Our taxonomy operationalizes this approach by categorizing sensory-relevant constructs into five hierarchical levels that broadly reflect neural activity (sensory-related neural excitability), perception (perceptual sensitivity), stimulus appraisal (physiological- and affective reactivity to sensory input), and behaviour (behavioural responsivity to sensory input).
We wish to note that others have attempted to do something similar in the past. However, these attempts were either not specifically focused on the sensory differences of autism [
27,
40,
41] or were not specifically focused on precise terminology-use [
42]. It is also perhaps worth noting that we do not consider the taxonomy as a framework or a model; we are not making any specific predictions of how sensory differences in autism emerge or are even associated with one another. However, we do believe that this taxonomy could be used to facilitate the development of relevant frameworks by clearly differentiating constructs that are currently being construed as related or synonymous, allowing their hypothesized relations to be more clearly stated. Thus, by helping to standardize the terminology used within autism research, the taxonomy promotes the development of better hypotheses, which we hope will result in work that will help improve our understanding of the mechanisms that underlie the sensory differences of autism.
Below, we describe each level and make suggestions for operationalization and measurement. For every level other than the first, we also explain why the level should be considered as distinct from adjacent levels. For convenience, we have provided a table so readers can get quick definitions of each level, as well as a list of measures readily available to assess each level. Table
1 should be seen as a supplement to the descriptions provided in the main text.
Table 1
Description, examples, and measurement of the levels of the five-level taxonomy of sensory-relevant constructs
Sensory-related neural excitability | The size, variability, latency, or degree of change, in a measurable brain response following sensory stimulation | The difference in ERP following the presentation of a 1000 Hz pure tone played from 0 to 70 dB in 5 dB steps | Sensory stimulation during the application of neuroimaging methods such as fMRI, fNIRS, EEG and MEG. Peripheral evoked potentials (e.g. auditory evoked potentials or peripheral nerve conduction methods may also be of relevance) | Most studies which have assessed sensory-related neural excitability have used a combination of custom paradigms and commercial equipment |
Perceptual sensitivity | How well an individual is able to detect and discriminate between stimuli | Detection Hearing the sirens of a passing fire truck | Psychophysical paradigms (e.g. 2AFC, 2IFC or MCS approaches in which participants are required to report whether they do or do not detect (or in which interval they do or do not detect) a stimulus of a given intensity across a range of trials | Examples of “off the shelf” tools for assessing perceptual sensitivity include: German Research Network on Neuropathic Pain [ 43] Audiometers which include pure-tone testing (e.g. see [ 45]) The University of Pennsylvania Smell Identification test [ 46] The Sniffin’ sticks test from Burghardt [ 47] Certain measures from the Sensory Integration and Praxis Tests (SIPT; [ 48]) Certain measures from the Evaluation in Ayres Sensory Integration (EASI; [ 49]) The NIH Somatosensory Toolbox [ 50] Vibrotactile psychophysical assessments conducted using Cortical Metrics stimulators (e.g. [ 51]) Psychophysical experiments can also be custom programmed in free-to-use, purpose-built software such as PsychoPy [ 52]. However, additional peripheral equipment may be required Questionnaires specifically designed to assess perceptual sensitivity: Sensory Perception Quotient (SPQ; [ 53]) |
| | Discrimination Accurately differentiating between a light touch and forceful physical contact | Psychophysical paradigms (e.g. 2AFC, 2IFC or MCS approaches in which participants are required to report whether they are able to discriminate between two simultaneously or sequentially presented stimuli | |
| | Temporal judgement | Psychophysical paradigms (e.g. 2AFC, 2IFC or MCS approaches in which participants are required to judge which of two stimuli were presented first [i.e. order judgement] or which of two stimuli were presented for more or less time [i.e. duration discrimination]) | |
| | Other | Perceptual sensitivity has other domains beyond detection, discrimination, and temporal judgement. There are also many other psychophysical paradigms that can be used to assess the low-level sensory processes that give rise to perception such as adaptation and habituation. Description and explanation of these additional domains and paradigms is beyond the scope of the article | |
Physiological reactivity to sensory input | Refers to how much an individual displays changes in relevant bodily processes in reaction to sensory input | Pupil dilation in response to sensory stimulation | Physiological measurements of near-automatic responses to sensory input (e.g. pupil dilation, galvanic skin responses, cortisol levels and skin conductance) | Most studies which have assessed physiological reactivity to sensory input have used custom paradigms coupled with commercial equipment |
Affective reactivity to sensory input | Refers to how an individual appraises and reacts to sensory input | Feeling uncomfortable and startled after hearing the sirens of the fire truck driving by Experiencing distress in reaction to certain textures or smells of a given meal | Questions which specifically ask autistic individuals how they feel about certain sensory stimuli without making reference to how they choose to respond are ideal. Similar questions directed at parents and clinicians may also be used (see main text for a discussion about limitations) Performance-based measures administered by clinicians could also be used to observe how children react on an affective level to sensory stimuli | Some of the available questionnaires with items probing affective reactivity to sensory input include: Sensory Profile and its various iterations [ 54, 55] Sensory Experience Questionnaire (SEQ; [ 56]) Sensory Processing Measure (SPM; [ 57]) Glasgow Sensory Questionnaire (GSQ; [ 58]) Sensory Sensitivity Questionnaire (SSQ; [ 59]) Clinician administered performance-based measures include: Sensory Processing Scale Assessment (SP3D; [ 60]) The Sensory Challenge Protocol (see [ 61] for details) Tailor made tasks to assess participant-reported pleasantness ratings in reaction to tactile stimuli, administered by trained researchers or clinicians, have also been used. For some examples, see [ 62] and [ 63] |
Behavioural responsivity to sensory
input | Refers to how an individual responds (or chooses not to respond) to sensory input they may find discomforting or pleasurable | Avoidance of loud and bright environments (sometimes referred to as “sensory avoidance”) Seeking of certain smells or textures one finds comforting (sometimes referred to as “sensory seeking”) | Questions which specifically ask autistic individuals or their caregivers how they or the autistic individual in reference responds to certain sensory stimuli. Ideally, such questions should help determine whether behavioural responses are due to a) affective discomfort or b) restricted and repetitive interest, as the latter may not be a sensory-specific behaviour per se. Administered tasks or stimuli followed by clinical observation may also be used | Given that currently available questionnaires contain items that assess both Affective reactivity to sensory input and behavioural responsivity to sensory input, we refer readers to the questionnaire in the cell above for currently available measures of this level Clinician administered performance-based measures include: Sensory Assessment for Neurodevelopmental Disorders (SAND; [ 64]) |
We broadly define sensory-related neural excitability as how an individual’s central and peripheral neural structures will activate in response to sensory input. This level can be used to place studies that have investigated neural activity in reaction to sensory input in both animal models of autism and autistic individuals. This level can also be differentiated to include unimodal and multimodal sensory-related neural excitability. This differentiation allows us to capture the studies which have predominantly assessed sensory-related neural excitability in unimodal sensory areas (where excitability is more directly linked to the immediate sensory cortical response to a stimulus) and studies assessing more multimodal sensory-adjacent processes such stimulus valuation and salience.
Investigations of sensory-related neural excitability typically use neuroimaging methods which quantify activation of the brain following peripheral sensory stimulation (i.e. stimulus-evoked responses). Neuroimaging methods vary in their spatial and temporal resolution. For example, in animal studies, techniques such as in vivo population calcium imaging and intracranial electrocorticography (or electroencephalography) have been used to measure firing patterns and postsynaptic potentials in animal models of autism. In human studies, non-invasive techniques such as functional magnetic resonance imaging (fMRI), electroencephalography (EEG) and magnetoencephalography (MEG) are more common.
Within the taxonomy, increased neural activation compared to ‘typical’ activity would be considered as evidence of
sensory-related neural hyperexcitability, whereas decreased neural activation would be considered as evidence of
sensory-related neural hypoexcitability. A review of studies having assessed
neural sensitivity to sensory input is well beyond the scope of this article, though we point readers to a relevant review article by Schauder and Bennetto [
42]. In brief, evidence of
sensory-related neural hyperexcitability in animal models have been mixed, with some studies having shown evidence of increased pyramidal firing rate in specific cortical regions in vivo [
65‐
67], decreased pyramidal firing [
68‐
73] and unchanged firing [
74‐
78]. Findings have also been mixed in autistic people. There is both evidence of
sensory-related neural hyper- and hypoexcitability, with results varying depending on methodology (e.g. type of sensory stimulation, location of recording, method of recording neural activity), participant sample, and how “excitability” was operationalized (see review by Takarae and Sweeney [
79]). Variability in findings between studies could also be in part due to the E-I balance theory of autism being oversimplified (see articles by O’Donnell and colleagues [
77] and Sohal and Rubenstein [
80], which argue for a more multidimensional approach to studying E-I balance).
Perceptual sensitivity
We define “
perceptual sensitivity” as how well an individual can detect and discriminate between the characteristics of low-level sensory information (e.g. luminance, contrast and frequency in the visual domain). Enhanced perceptual functioning and savantism is thought by many to be part the product of
perceptual hypersensitivity (we point readers to relevant chapters/articles by Mottron and colleagues [
12], and Baron-Cohen and colleagues [
15]).
Perceptual sensitivity to low-level sensory stimulus information is typically assessed through psychophysics. Psychophysics refers to a class of methods used to objectively study the perceptual system [
81]. Psychophysical methods have their strengths in the fact that they are often reliable (i.e. a person’s perceptual threshold on one day is predictive of their perceptual threshold on another) and the outcome measures (e.g. perceptual thresholds and psychometric functions) they produce often have links to known neural processes. For example, poor amplitude and frequency discrimination in the tactile domain, which may manifest as elevated discrimination thresholds, are thought to at least partially reflect alterations in GABAergic lateral inhibition [
82]. Given the link between psychophysically-derived outcomes measures and known biological processes, psychophysics makes it possible to draw inferences from performance to function (or dysfunction) of specific neurophysiological processes [
83].
Studies using the method of constant stimuli have identified steeper psychometric functions in individuals on the autism spectrum [
84,
85], suggesting a less dynamic range of perception. Similarly, the application of two-alternative forced choice and two-interval forced choice paradigms have identified both lower and higher perceptual thresholds in tactile [
31,
39,
86‐
92], visual [
93], olfactory [
94,
95] and auditory [
14,
38,
96‐
99] domains. Lower thresholds suggest
perceptual hypersensitivity whereas higher thresholds suggest
perceptual hyposensitivity.
It is also possible to assess
perceptual sensitivity using a questionnaire approach. However, we wish to caution readers that, to the best of our knowledge, the available questionnaires have not been validated against psychophysically determined perceptual outcomes (e.g. thresholds). Generally speaking, questionnaire-based measures of sensory difficulties have not been good indicators of actual psychophysical thresholds (e.g. see [
100]). Nonetheless, the Sensory Perception Quotient (SPQ) developed by Tavassoli et al. [
53] was developed to specifically assess
perceptual sensitivity. The SPQ contains items that ask individuals to respond to statements such as “I would be the last person to detect if something was burning” and “I would notice if someone added 5 grains of salt to my cup of water”. These kinds of probing statements are like psychophysical methods in that they attempt to identify whether an individual has perceptual thresholds that are non-typical. To the best of our knowledge, the SPQ is the only questionnaire-based measure that aims to specifically probe
perceptual sensitivity.
Our taxonomy differentiates between
sensory-related neural excitability (the level above) and
perceptual sensitivity for conceptual reasons. While it could be argued that altered
perceptual sensitivity necessitates alterations of
sensory-related neural excitability, it is at least theoretically possible for differences in perception to be explained by alterations of neural processes that do not necessarily translate into changes in
sensory-related neural excitability (i.e. changes in neural input–output functions could result in only supra-threshold sensory changes that do not affect the point at which a stimulus is categorically detected). Thus, while altered
perceptual sensitivity might indeed be the product of altered
sensory-related neural excitability (as first posited by Rubenstein and Merzenich [
101]), it is best to assume that these constructs are unrelated until proven otherwise.
Physiological reactivity to sensory input
We have purposely placed
physiological reactivity to sensory input after
perceptual sensitivity, but before
affective reactivity to sensory input. The general distinction between
perceptual sensitivity and a behavioural reaction is logical. This is because the perception of a stimulus does not necessitate a specific or overt behavioural reaction (i.e. the perception of the same stimulus in different contexts may result in different behavioural reactions). In a practical sense, this is to say that differences in
perceptual sensitivity do not fully explain differences in
physiological- and/or
affective reactivity to sensory input. The more specific distinction between
physiological- and
affective reactivity to sensory input might come as a surprise. However, the relationship between physiological reactions and psychological operations are not one-to-one (i.e. inferences about affect cannot solely be made based on physiological signals). This justifies the separation of
physiological and
affective reactivity to sensory input. See Cacioppo and Tassinary [
102] for a discussion of the limitations with inferring psychological significance from physiological signals.
Physiological reactivity to sensory input can be assessed through physiological responses using psychophysiological approaches. Physiological responses are defined as changes to the parameters of a physiological measure in response to sensory input [
103], and are thought to reflect changes in either the autonomic nervous system (ANS) or limbic–hypothalamic–pituitary–adrenal axis (LHPA). The ANS is comprised of the sympathetic, parasympathetic, and enteric nervous systems (the latter is not discussed below due to lack of immediate relevance). While the sympathetic nervous system is responsible for fight and flight responses, the parasympathetic nervous system is responsible for recovery and restoration. The LHPA axis regulates the body’s responses to stress and promotes restoration towards homeostasis following a stressor [
104]. Most studies assessing
physiological reactivity to sensory input in autism are focused on changes of the sympathetic nervous system of the ANS rather than the LHPA system, presumably because studies often aim to establish cause and effect within short time scales and ANS responses, such as changes in heart rate, heart rate (HR) variability (HRV), blood pressure (BP) and electrodermal activity (EDA), are often immediate. Unlike ANS-based responses, LHPA-based responses, such as changes in cortisol, are much slower (it is also worth noting that processes indexed by ANS- and LHPA-based responses are psychophysiologically distinct and may need to be considered as separate sub-levels within
physiological reactivity to sensory input).
There have been many studies which have identified differences in physiological responses to sensory input in autism and we point readers to the systematic review of studies of
physiological reactivity to sensory stimuli in autism by Lydon and colleagues [
105], where the virtues and limitations of measuring physiological responses in autism are also discussed. In brief, existing studies of
physiological reactivity to sensory stimuli in autism have typically compared physiological reactivity before, during and after the presentation of a sensory stimulus (or sensory stimuli). These study designs allow for the comparison of baseline physiological reactivity, changes in physiological reactivity following stimulus presentation, and habituation, which, in this context, can be broadly described as a reduction in the intensity of a physiological reaction (or physiological reactions) following repeated presentation of a sensory stimulus or stimuli. In general, the findings regarding physiological reactivity to sensory input in autism remain mixed and differ depending on the type of sensory stimuli and measures of physiological reactivity. Nonetheless, as highlighted by Lydon and colleagues in their review, the majority of studies do suggest a difference in
physiological reactivity to sensory input between autistics and non-autistic controls. For instance, with regard to EDA (which has been the most common measure of physiological reactivity used in studies to date), many studies have shown group differences in changes in EDA in reaction to basic sensory stimuli [
106‐
110]. Interestingly, some studies have also shown group differences in baseline EDA, with autistic individuals having higher baseline EDA than their typically developing peers [
107,
111]. These findings are supported by the similar finding of heightened respiratory sinus arrythmia (heart rate variability in synchrony with respiration) in autism compared to controls by Schaaf and colleagues [
112]. Still, there are also studies which have not found differences in physiological reactivity between autism and controls either at baseline [
113‐
115] or in reaction to sensory input [
113,
116,
117]. One study has shown an association between
physiological reactivity to sensory input and sensory-related measures such as the Sensory Profile [
118]. However, there is also a study which found no association between physiological reactivity to sensory input and measures on the Short Sensory Profile [
110].
Affective reactivity to sensory input
Our taxonomy explicitly differentiates the point and dynamic range at which an individual perceives, or can perceive, sensory information (i.e. an individual’s
perceptual sensitivity) from their subjective appraisal of sensory stimuli as pleasant or unpleasant (i.e. “
affective reactivity to sensory input”). Like the DSM-5, the “hyper” and “hypo” prefix can also be used to describe affective reactions to sensory input. An individual who is more likely to experience strong emotional responses when perceiving sensory stimuli would be considered as someone who demonstrates
affective hyperreactivity to sensory input, while a person who is less likely to experience affective discomfort upon perceiving sensory input would be someone who demonstrates
affective hyporeactivity to sensory input. As previously discussed, it is often assumed that individuals who are more likely to experience sensory stimuli as distressing or emotionally overwhelming (i.e.
affective hyperreactivity to sensory input) do so out of having
perceptual hypersensitivity. Similarly, individuals who are more likely to be indifferent to stimuli that others would otherwise find aversive are sometimes thought to have
perceptual hyposensitivity [
119]. However, these assumptions have rarely been tested, and where associations have been identified between
perceptual sensitivity and
affective reactivity to sensory input, the associations are often weak or non-existent [
39,
120,
121], suggesting that while perceptual sensitivity and affective reactivity to sensory input may be related, they may in fact be separate constructs.
While affective reactivity to sensory input can be readily assessed using many of the available popular questionnaires currently applied to quantify the sensory differences of autism, we wish to highlight some important considerations when using questionnaire responses to infer interindividual differences in affective reactivity to sensory input.
First, as stated in Table
1, questionnaires containing items that speak to
affective reactivity to sensory input may also contain items that speak to
perceptual sensitivity and
behavioural responsivity to sensory input (described next). The lack of separation between constructs or “levels” within questionnaires can be problematic if one wishes to investigate the association between levels of sensory differences. For example, if a questionnaire that is generally thought to assess interindividual differences in
affective reactivity to sensory input contains items that speak to
perceptual sensitivity, associations between the questionnaire and psychophysically determined measures of
perceptual sensitivity may be driven by the presence of items related to
perceptual sensitivity in the questionnaire primarily used to assess
affective reactivity to sensory input.
Second, there is reason to be sceptical of whether respondents to questionnaires assessing sensory differences in autism are distinguishing between distinct constructs when completing questionnaires. For instance, correlations between the SPQ, a questionnaire-based measure of
perceptual sensitivity, and self-report measures of what we have described as
affective reactivity to sensory input, such as the Sensory Over-Responsivity Inventory (SensOR), are generally high [
53]. While this might suggest a correlation between
perceptual sensitivity and
affective reactivity to sensory input, the strength of the correlation between the SPQ and the SensOR (
r = − 0.46 in an autistic population) have almost comparable effect sizes to correlations between two different self-report questionnaires of
affective reactivity to sensory input (e.g. between the GSQ and AASP, which has been shown to have a Pearson’s
r value of 0.640 in an autistic population, [
122]). We highlight this point to caution readers to the possibility that questionnaires targeting distinct constructs may not actually be separate in the mind of respondents (which in turn might mean that scores using these questionnaire-based approaches may not be appropriate for the specific purpose of testing relationships between levels of the hierarchical taxonomy). Still, it is also possible that the constructs are in fact highly related and demonstrate limited discriminant validity when measured using questionnaire methods.
Third, we wish to highlight an important limitation of measures of
affective reactivity to sensory input which are based on parent-report and even clinical observation. Inferences about an individual’s mental state by others are generally made based on observations of behaviour which can arise due to reasons that are unrelated to sensory differences (e.g. restricted, and repetitive behaviours that are not sensory in nature
5).
Affective reactivity to sensory input, then, is perhaps best assessed through directly asking autistic individuals how they feel about certain sensory stimuli. For example, the GSQ contains many items which directly probe individuals about their affective reactions to sensory input (e.g. item 6: “Do you find certain noises/pitches of sound annoying?” and item 23: “Do you hate the feel or texture of certain foods in your mouth?”). While self-report by autistic participants is ideal, we recognize self-report may not always be possible (e.g. in younger and/or minimally verbal autistic children).
Behavioural responsivity to sensory input
The final and highest level of the taxonomy is “
behavioural responsivity to sensory input”. This level of the taxonomy subsumes the DSM-5 [
123] description of sensory differences (“hyper- or hyporeactivity to sensory input or unusual interests in sensory aspects of the environment). That is,
behavioural responsivity to sensory input can be used to describe both observable behavioural responses to sensory discomfort and expected but unobserved behavioural responses (e.g. apparent indifference to pain/temperature).
Behavioural responsivity to sensory input can also be used to describe behaviours related to the seeking and avoidance of sensory input. Like the DSM-5 and
affective reactivity to sensory input, the “hyper” and “hypo” prefix could also be used to describe behavioural responses to sensory input.
The overall differentiation between
perceptual sensitivity, responses related to stimulus appraisal (i.e.
physiological-, and
affective reactivity to sensory input) and
behavioural responsivity to sensory input is not new. See section entitled “
The current taxonomy in relation to previous models and frameworks”. The key difference between
behavioural responsivity to sensory input and the other levels of the taxonomy is that this level exclusively refers to observable behaviours while the others refer to unobservable internal states (e.g. the perception of sensory input without a behavioural response or the subjective appraisal and experience of sensory input). This makes it unclear whether previous findings of group differences on measures such as the SSP solely reflect group differences on
behavioural responsivity to sensory input”. Differentiating between
perceptual sensitivity and
behavioural responsivity to sensory input prevents the assumption that certain behavioural responses are due to differences in
perceptual sensitivity (e.g. it could be assumed that the indifference towards temperature is due to having higher thermal perceptual thresholds). Similarly, differentiating between stimulus appraisal (i.e.
physiological- and/or
affective reactivity to sensory input) and
behavioural responsivity to sensory input prevents the assumption that certain
behavioural responses to sensory inputs are due to differences in stimulus appraisal, be that increased or decreased
physiological- and/or
affective reactivity to sensory input.
Sensory-related avoidance and seeking behaviours have been typically assessed using questionnaires that assess behaviour responses. An example of an item probing sensory avoidance behaviour is item 8 of the SSP, which states “Avoids certain tastes or food smells that are typically part of children’s diets”. Similarly, the SSP also contains items probing sensory seeking behaviours (e.g. item 15—“Enjoys strange noises/seeks to make noise for noise’s sake). A limitation of some of these kinds of questionnaire-based approaches to assessing behavioural responsivity to sensory input is that the items used to quantify seeking and avoidance behaviour can be overly ambiguous. For example, in item 8 of the SSP described immediately above, it is unclear whether an autistic individual who frequently avoids certain tastes or food smells are doing so because of a sensory difference. Indeed, some behaviours that could be considered as evidence of sensory avoidance (or even seeking) could also be explained by the other symptoms of autism such as restricted and repetitive behaviours, intolerance of uncertainty and/or insistence on sameness, or simply challenges creating adaptive motor actions. Adjustment of existing questionnaires (or the development of new ones) should aim to develop items that non-ambiguously assess behavioural responses that are specifically due to an individual’s sensory differences.
Adoption and development of more non-questionnaire-based measures to assess
behavioural responsivity to sensory input could also be an avenue for future investigation. Further use or development of measures like the Sensory Assessment for Neurodevelopmental Disorders (SAND; [
64]), Sensory Processing Scale Assessment (SP-3D:A; [
124]), Sensory Processing Assessment (SPA; [
60]) and Tactile Defensiveness and Discrimination Test-Revised (TDDT-R; [
125]), which assess clinician-rated behaviours that a child performs in response to standardized stimulus presentations (e.g. the scales might assess whether a child puts their hands over their ears when a sound is presented), may be useful for specifically assessing
behavioural responsivity to sensory input. The development of more accessible and scalable tools that can be delivered on smart phones or smart devices may also be a fruitful avenue. For example, smart phones now have the capacity to assess one’s exposure to sound throughout a given period. Adapted appropriately, this technology (or technology like this) could be applied to implicitly assess whether autistic individuals are more likely to avoid loud spaces compared to their neurotypical counterparts, providing a measure of s
ensory-related behavioural responsivity. Such smart devices could also be used in conjunction with methods that are currently used to measure physiological responses, as well as experience sampling methods (i.e. ecological momentary assessments [
126]).
Studies assessing
affective reactivity to sensory input in autism to date have focused on using questionnaire-based measures such as the SSP, the SP and the SPM. These studies have typically found that individuals on the spectrum tend to display both
behavioural hyperresponsivity to sensory and hyporeactivity to sensory input [
127‐
130]
, although, as mentioned, these findings may be hampered by the fact that questionnaire-based measures do not sufficiently differentiate between different levels of sensory differences. Using questionnaire-based approaches, many studies have also reported heightened
sensory avoidance and
sensory seeking in autistic individuals [
131]. The use of clustering approaches on scores on the questionnaire-based approaches to identify sensory subtypes of autism has also become increasingly popular. As described by Lane [
132], sensory subtypes have been used in attempt to identify homogeneous sub-groups of autistic individuals with similar sensory features. Identifying sensory subtypes can help with our understanding of the basis of sensory differences in autism, as well as with the provision of targeted treatment. At the time of writing, sensory subtypes have only been explored in toddlers, children and adolescents, and only using questionnaire-based measures.