Individuals with autism spectrum disorder (ASD) not only show deficits in social communication, but also show atypical sensory processing characterized by sensory hypersensitivity. This feature is also emphasized in the fifth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-5) (APA
). Sensory dysfunction such as this can be assessed using the Sensory Profile, which is a self-report questionnaire that determines the individual’s sensitivity; for example, if he/she shows strong emotional responses to sensory stimuli experienced as a part of daily life (Dunn
; Dunn and Westman
). Tomchek and Dunn (
) reported that children with ASD are more sensitive to sensory information (tactile, 65.1%; taste and olfaction, 56.2%; visual and auditory, 50.9%) than typically developing (TD) peers (13, 9.7, and 7.9%, respectively) according to their scores on a short version of the Sensory Profile (i.e. Short Sensory Profile).
Sensory hypersensitivity is conventionally explained by the hypothesis that an abnormally high sensitivity for detecting sensory signals leads to atypical responsiveness. Blakemore et al. (
) reported that individuals with Asperger’s syndrome can detect small displacements in vibrotactile stimuli with a lower detection threshold than TD control individuals; this feature remained dominant even when the stimuli were delivered at a high frequency (200 Hz). Cascio et al. (
) and Puts et al. (
) demonstrated that the lower detection thresholds were found even when the vibrotactile stimuli were presented at low frequencies (33 and 25 Hz). However, another study showed that the detection thresholds of children with ASD and TD children were not significantly different regardless of stimulus frequency (40 and 250 Hz) (Guclu et al.
). These contrasting results from previous studies may partly be due to the inherent variability in sensory processing of individuals with ASD (Simon and Wallace
). Thus, individual differences in sensory processing must be accounted for and it should be considered how they relate to varied responsiveness to sensory stimuli in patients’ with ASD daily life.
Another hypothesis that can explain hypersensitivity in patients with ASD involves aberrant temporal processing of sensory inputs. Individuals with ASD frequently complain about the flickering nature of fluorescent illumination, which is also thought to induce their repetitive behaviour (Colman et al.
). These findings might indicate that some patients with ASD might have extremely high temporal resolution (exceeding the 60-cycle flicker) of processing sensory stimuli. A typical symptom of hypersensitivity, avoiding wearing clothes, may stem from aberrant temporal processing of texture (Green and Ben-Sasson
). Another study described the superior ability of individuals with ASD to temporally process visual stimuli (Falter et al.
). In that study, vertical bars were consecutively presented to the left and right of a fixation cross on a monitor with different temporal lag times ranging from 8.3 to 99.6 ms. The participants were instructed to determine whether the stimuli were presented simultaneously or not. The authors found that individuals with ASD judged the stimuli to not be simultaneously presented more frequently than controls, indicating that they may have superior temporal resolution. On the other hand, individuals with ASD have also been shown to have lower temporal resolution while processing tactile stimulation (Tommerdahl et al.
; Wada et al.
). Tommerdahl et al. (
) reported lower temporal resolution in the tactile temporal order judgment (TOJ) of the index and middle fingers of one hand in individuals with ASD, although the tactile TOJ of both hands was not significantly different between the ASD and TD groups. Moreover, the temporal resolution of one hand came to be precise compared with TD individuals when conditioning vibrotactile stimuli (frequency, 25 Hz) were presented on another skin site. In contrast, Wada et al. (
) reported that while the temporal resolution of tactile TOJ for both hands was slightly lower in children with ASD than in TD children, the p-value for this comparison was not so significant, given the large individual differences in sensory processing in the ASD group.
The contrasting results for both sensitivity and temporal resolution of sensory processing in individuals with ASD indicate a diversity in sensory processing in this population. The type of sensory processing underlying hyper-reactivity remains unclear and seems to be related to the temporal processing of stimuli from the environment. In this study, we elucidated the relationship between individual differences in temporal resolution of sensory processing and those in the severity of hypersensitivity. We focused on the tactile modality, given the variety of findings related to tactile temporal processing (Puts et al.
; Tommerdahl et al.
; Wada et al.
). We adopted the TOJ task with vibrotactile stimuli to measure the temporal resolution of stimulus processing and compare it between the ASD and TD groups.
Previous studies have demonstrated large individual differences in sensory processing, particularly in detection sensitivity and temporal resolution of processing vibrotactile stimuli, among individuals with ASD. As temporal processing is often considered to be linked to sensory hypersensitivity, we elucidated whether individual differences in temporal resolution are related to the degree of hyper-reactivity in patients with ASD. Our results suggest that individuals in the ASD group who had higher temporal resolution of processing vibrotactile stimuli tended to be more affected by various sensory stimuli experienced as a part of their daily life. However, the detection threshold and sensitivity were almost not related to this atypical responsiveness but were related to the severity of stereotyped behaviour and restricted interests and, partially, to reciprocal social interactions as assessed by the ADOS-2. These data indicate that temporal processing of tactile stimuli may underlie sensory hyper-reactivity in individuals with ASD, while the detection threshold and sensitivity may underlie the severity of other aspects of ASD.
Our data provide first evidence that temporal processing of stimuli, and not detection threshold and sensitivity, is correlated with a self-assessed score of hypersensitivity. In fact, the temporal resolution and detection threshold/sensitivity were not significantly different between the two groups. This kind of behavioural performance points to the existence of a wide continuum in the degree of hyper-reactivity. Wada et al. (
) reported lower temporal resolution in children with ASD; however, the mean age of their sample (mean age, 11.8 years) was lower than that of our sample (mean age, 19.1 years). Dysfunctions according to developmental changes in sensory processing might contribute to the differences attributed to the participant groups. Since the temporal resolution of stimuli in neurotypical individuals reportedly increases from childhood to adolescence (Stevenson et al.
), it is possible that deviations in temporal resolution appear in individuals with ASD until adolescence. Interestingly, we found relationships between temporal resolution and hypersensitivity related to the ‘enhanced responsiveness’ and ‘total sensory responsiveness’ AASP sub-categories (and not the ‘lower responsiveness’ sub-category). These results suggest that the temporal processing of tactile stimuli is predominantly associated with hyper-reactivity when some types of sensory information included in the AASP such as visual, auditory, somatosensory, olfactory, and gustatory are provided in daily life.
Furthermore, we found that the detection threshold and sensitivity to 200-Hz vibrotactile stimuli were positively correlated with the reciprocal social interactions and stereotyped behaviour and restricted interests ADOS-2 sub-scores. Guclu et al. (
) also indicated that elevated tactile sensitivity was related to socioemotional problems experienced in daily life. Moreover, the detection threshold was lower in individuals with Asperger’s syndrome than in their TD peers, predominantly for high frequency vibrotactile stimuli (Blakemore et al.
), although this result could not be replicated in another ASD sample (Guclu et al.
). Thus, while temporal processing of vibrotactile stimuli is linked to the degree of hypersensitivity, detection threshold/sensitivity for high frequency vibrotactile stimuli might be linked to stereotyped behaviour and restricted interests and other social impairments characteristic of ASD. However, as our sample size was small, further studies are required to examine this link between the severities of several aspects of ASD symptoms and detection performances while processing high frequency vibrotactile stimuli.
With regard to the neural basis of atypical responses to sensory stimuli, mainly hyper-reactivity, idiosyncratic somatosensory evoked potentials for tactile stimuli have been reported (Miyazaki et al.
). Cascio et al. (
) also reported that early (120–220 ms) and late (220–270 ms) brain waves elicited by air-puff stimulation might be related to the degrees of hyper-reactivity and hypo-reactivity, respectively. Hyper-reactivity would then be consistent with the somatosensory association cortical response, while hypo-reactivity would be consistent with later brain processes such as allocation of attention or ascribing emotional valence to stimuli. Simon et al. (
) demonstrated that the degree of hypo-reactivity was associated with elevated levels of left alpha and theta power and increased alpha and theta connectivity in resting state electroencephalography in toddlers at high risk (HR) of being diagnosed with ASD in the future. They also found that hypo-reactivity was related to reduced signal complexity at occipital and temporal electrodes. These findings indicate that reduced sensory responsiveness in HR toddlers corresponds to broad changes in neural synchronization, both within and across cortical areas, and a resultant loss of complex neural interactions.
Several studies using mouse models of autism have reported that autistic mice frequently show an excitation-inhibition imbalance (i.e. E/I imbalance) in the central nervous system (Braat and Kooy
; Pizzarelli and Cherubini
; Rubenstein and Merzenich
). One of the major features of the model mice is reduced concentration of gamma-aminobutyric acid (GABA) in the brain that involves the deactivated GABA receptor and subsequent degraded release of the neurotransmitter. The model mice showed defensive behaviour to air-puff stimulation at the whisker more frequently than wild-type mice (He et al.
), in addition to more frequent pathognomonic behaviour while interacting with cage mates and deficits in social communication. Similarly, human post-mortem studies showed reduced concentrations of GABA in the anterior cingulate cortex and fusiform gyrus in patients with ASD (Oblak et al.
). Recent magnetic resonance spectroscopy studies showing in vivo GABA states in the human brain have revealed the relationship between GABA concentrations in the human brain and behavioural performances. Robertson et al. (
) demonstrated that GABA concentration in the visual cortex was lower in individuals with ASD than in their TD peers. Moreover, individuals with ASD also showed decreased suppressive ratio of perceptual switching when different visual stimuli were concurrently presented to each eye (i.e. binocular rivalry). Puts et al. (
) reported reduction in GABA levels also in individuals with Tourette syndrome, with lower GABA concentrations in the somatosensory cortex being associated with more severe tics. Another study (Terhune et al.
) showed that reduced GABA levels might lead to more precise estimates of the temporal duration of visual stimuli. It is possible that the reduction in GABA levels in the primary sensory cortices underlies perceptual states in various sensory modalities, resulting in the inhibition of sensory inputs and involuntary movement, and sometimes in excessive sensory processing. We speculated that the enhanced temporal resolution in individuals with ASD might be caused by an E/I imbalance in their brains, and future studies are needed to further elucidate this hypothesis.
Aberrant sensory processing in patients with ASD is regarded as the basis of their impairments in social cognition and adaptive behaviour (Ben-Sasson et al.
). Green et al. (
) showed that task-irrelevant tactile stimulation complicates the comprehension of the meaning of “sarcasm”, which is needed to interpret communicative intents in non-literal language. In the task, neural activities in the left auditory language areas (angular gyrus) and the occipital cortex degraded by the distractive tactile stimuli with strong activation in the somatosensory cortex. The degraded neural activity would reflect that they shifted their attention away from the task and towards the sensory stimuli. In contrast, the degradation in neural responses disappeared when they were required to shift their attention to the facial expression and tone of voice of the speaker, while the medial prefrontal cortex (mPFC) strongly activated. Thus, strong mPFC activation was assumed to inhibit distractive sensory inputs in order to properly interpret the communicative intents. These findings including our present results indicate that several stages exist regarding the occurrence factor of hyper-reactivity. In one stage, weak inhibitory function on sensory inputs and resulting strong neural activation in the primary sensory cortex would be an important factor. Higher temporal resolution of sensory stimuli may be related with this stage because this feature of sensory processing would result in vast amounts of inflow of sensory information. Difficulty in attentional deprivation from distractive and/or unpleasant feelings of sensory stimuli may play an important role at another stage. Aversive touch was found to facilitate neural activation in the posterior cingulate cortex and the insula in individuals with ASD (Cascio et al.
), and the amplitude of the activity in the insula was positively correlated with the severity of disabilities in social communication measured by the ADOS-2. Thus, we speculate that excessively strong neural responses by sensory inputs in connection with weak inhibitory function exist at the first stage, and then, this over-responsivity would interfere with adaptive social communication and emotional processing.
Our study provides the first report indicating that the temporal processing of vibrotactile stimuli may underlie sensory hypersensitivity, while detection threshold/sensitivity for high frequency vibrotactile stimuli may be linked to the severity of some ASD symptoms. Enhanced sensory processing in patients with ASD may result in a large inflow of sensory signals from the surrounding environment, the several neural substrates contributing to the diversity of sensory processing in these patients. Thus, treatment plans must consider individual sensory sensitivity levels, which may consequently determine the patients’ compliance to treatment.
We would like to thank T. Atsumi for his comments, and A. Tanaka and T. Nawa for technical help. We also thank Y. Gorie, K. Nishimaki, S. Kim, H. Agarie, and M. Suzuki for help with participant recruitment, and Dr. Y. Nakajima for their continuous encouragement. This study was supported by a Grant-in-Aid from Japan Society for the Promotion of Science (JP15K17333, JP17942790), Ministry of Education, Culture, Sports, Science and Technology-JAPAN (JP16H01520, JP17H05966).