Impaired odor identication ability and olfactory hedonic capacity in children with elevated autistic traits

Background: Atypical olfactory function in autism spectrum disorder has been documented in numerous studies, but little is known about its occurrence in individuals with autistic traits. The aim of the current study was to investigate odor identication ability and olfactory hedonic capacity in children with autistic traits. Methods: The study included 91 children in a high autistic trait group, and 128 children in a low autistic trait group, as determined based on Short Autism Spectrum Quotient scores. The Universal Sniff (U-Sniff) test was used to measure odor identication ability and olfactory hedonic capacity. Olfactory hedonic capacity was also measured using the child’s version of the Chemosensory Pleasure Scale (CPS-C). Results: Children in the high autistic trait group exhibited signicantly impaired odor identication and olfactory hedonic capacity measured by CPS-C than those in the low autistic trait group, but there was no signicant difference in olfactory hedonic capacity measured by U-Sniff between the two groups. Impaired odor identication was signicantly correlated with olfactory hedonic capacity measured by CPS-C. Limitations: Only a self-reporting questionnaire was used to screen children for autistic trait. The combination of a self-reporting scale and diagnosis by a clinical expert would result in more accurate screening of individuals for autistic trait. Conclusions: Children with high autistic trait exhibited poorer odor identication ability and olfactory trait hedonic capacity than children without high autistic trait. These results pertaining to odor identication and olfactory hedonic capacity may be the endophenotypic markers for autism spectrum disorder.


Introduction
Autism spectrum disorder (ASD) encompasses a variable group of neurodevelopmental disorders characterized by social and communication impairments as well as repetitive behaviors (1). Most individuals with ASD exhibit atypical sensory capacity in childhood (2), which has been included in the DSM-5 diagnostic criteria for ASD (1). Most evidence on altered sensory processing is based on visual and auditory processing in people with ASD (3)(4)(5)(6). Aberrant olfactory function in people with ASD has also been demonstrated in numerous studies, and the most prominent manifestation of it is odor identi cation (7)(8)(9)(10)(11)(12). Some heterogeneous results have been reported however, and in some studies odor identi cation was not signi cantly impaired in people with ASD (13)(14)(15). Inconsistent methodology may have contributed to the heterogeneous results that have been reported (16).
The University of Pennsylvania Smell Identi cation Test (UPSIT) (17) and the "Sni n' Sticks" test (18) are most commonly used to measure odor identi cation, but the odors they utilize may be unfamiliar to children (19). In addition, descriptors in the UPSIT and the Sni n' Sticks test are usually presented in writing, which is not feasible to use to measure odor identi cation in children with ASD because of their delayed speech and communication impairment (8). Schriever,Agosin (19) developed the Universal Sniff (U-Sniff) test to measure odor identi cation in children, in which the descriptor of each odor is presented both photographically and in writing. In addition, the U-Sniff odors were selected by children from different countries, and the test has exhibited good reliability and validity in children from different countries (19,20). Therefore, the USniff test was used to measure odor identi cation in the current study, to reduce the additional effects of speech and verbal labeling.
Hedonic capacity, which refers to the ability to experience pleasure, is considered a crucial index for understanding the core impairments in ASD-social and affective de cits (21,22). People with ASD exhibit signi cant impairment in physical and social hedonic capacity (21). The overlapping brain areas associated with olfaction and emotional processing suggest that olfaction may have an unusual impact on emotion perception (23). In previous studies people with ASD have exhibited diminished odor hedonic discrimination and judgement capacity (9,12).
Autistic traits were continuously distributed in the general population (24), which would bene t heritable ASD marker research. Individuals with so-called "autistic trait", also referred as the broader autism phenotype, are a group of people who have autistic traits but their severity is subclinical (24,25).
Therefore, it is reasonable to hypothesize that individuals with autistic trait also have odor identi cation impairment and olfactory anhedonia. To our knowledge no study has examined odor identi cation in individuals with autistic trait. Novacek, Gooding (22) reported that young adults with autistic trait exhibited signi cantly lower social pleasure, but olfactory hedonic capacity in children/adolescents with autistic trait is understudied.
The aim of the current study was to investigate odor identi cation ability and olfactory hedonic capacity in children with autistic trait. We hypothesized that children with autistic trait would exhibit reduced odor identi cation ability and olfactory hedonic capacity.

Participants
The study included 91 children with high autistic trait as determined via Short Autism Spectrum Quotient (AQ-10) scores (26), 35 boys and 56 girls, with a mean age of 13.01 ± 2.93 years (range 9-17 years). The AQ-10 was also used to identify 128 children with low autistic trait scores, 50 boys and 78 girls, who had a mean age of 13.43 ± 2.71 years (range 9-17 years). The children were recruited from several schools via convenient sampling. A total of 415 students were initially administered the AQ-10, and those with scores ≥ 6 were assigned to the high autistic trait (high AQ) group, and those with scores ≤ 3 were assigned to the low autistic trait (low AQ) group (27). Participants were excluded if they had had a bad cold in the past week, or if they had a history of nasal diseases, brain injury, or any neuropsychiatric disorders as determined via a personal interview with a trained interviewer. The study was approved by the Ethics Committee of Southern Medical University, and conducted in accordance with the principles of the Declaration of Helsinki. Informed consent was obtained from teachers and the children's legal guardians.

The U-Sniff test
Odor identi cation ability and olfactory hedonic capacity were assessed using the USniff test (19). The U-Sniff test utilizes 12 pen-like "Sni n' Sticks", which emit apple, banana, butter, coffee, cut grass, sh, ower, lemon, onion, orange, peach, and strawberry odors. It has previously exhibited good reliability and validity (19,20), and in a study conducted in China its test-retest reliability was 0.69 (20).
The test was conducted in a quiet and well-ventilated room. Before odor presentation, the descriptors were showed to the participant and read aloud by the examiner. One odor at a time was birhinally presented to the participant for 3 seconds. The participant made a forced choice from four given descriptors, which were presented in picture and textual forms. The participant was asked to smell the odor again if they were not sure of their choice. In addition, participants were asked to rate the perceived odor pleasantness from 1 ("very much dislike this odor") to 9 ("very much like this odor") based on their subjective feelings. Odor identi cation ability was scored as the sum of correct odor choices. The total odor identi cation score for each participant ranged from 0-12, and higher scores represented better odor identi cation ability. The sum of the odor pleasantness ratings was used to quantify olfactory hedonic capacity, and higher scores represented better olfactory hedonic capacity.

The child's version of the Chemosensory Pleasure Scale
Olfactory hedonic capacity was also assessed via the child's version of the Chemosensory Pleasure Scale (CPS-C), in which the expression of the standard Chemosensory Pleasure Scale (28) items is modi ed to facilitate understanding by children. As with the standard Chemosensory Pleasure Scale, the CPS-C consists of 12 items encompassing three broader factors, "Food", "Imagination", and "Nature", which respectively relate to hedonic capacity while eating, anticipating food, and smelling natural scents.
Participants were asked to respond using a 6-point Likert scale ranging from 1 ("not at all true for me") to 6 ("very true for me"). The sum of scores from the CPS-C items indicated the participant's olfactory hedonic capacity, and higher scores represented better olfactory hedonic capacity. The test-retest reliability of the CPS-C was 0.718.

The Short Autism Spectrum Quotient
Autistic trait was assessed using the Short Autism Spectrum Quotient (AQ-10) (26), which contains the 10 screening items from the 50-item Autism Spectrum Quotient (29) with the highest sensitivity and speci city. It has been con rmed that the AQ-10 is a useful screening instrument for autistic trait and ASD (26,30). There are child, adolescent, and adult versions of the AQ-10, which were used to measure autistic traits in the participants in the current study based on their ages. The participants who aged from 5-8 years and 9-11 years used the child version of AQ-10 and the participants who aged 12-14 years and 15-17 years respectively used the adolescent version and adult version of AQ-10. The participants were asked to choose the degree to which they agreed or disagreed with the statements in the AQ-10. Each item has four alternative choices: strongly agree, slightly agree, slightly disagree, and strongly disagree. In the current study the response to each item was recorded as "1" if the choice was "strongly agree" or "slightly agree", and "0" if it was "slightly disagree" or "strongly disagree". All scores were summed, and higher scores represented more severe autistic trait symptomatology.

Data analysis
Statistical analysis was performed using SPSS 25.0 (SPSS Inc, Chicago, IL, USA). The t-test for independent-samples or the chi-square test were used to assess the signi cance of differences in age or sex between the high AQ group and the low AQ group. Group differences between the high AQ group and the low AQ group assessed using the t-test for independent samples. In addition, the effect size estimates of the Independent t-test were calculated using Cohen d and phi estimate was used to measure the effect size of chi-square test. The Shapiro-Wilk normality test was used to assess the normality of data distributions in each group. Relationships between odor identi cation score and CPS-C total or subscale scores were assessed in each group. Pearson's correlations were calculated if the data were normally distributed, and Spearman's correlations were calculated if the data were not normally distributed.

Discussion
To our knowledge the current study is the rst to directly measure odor identi cation and olfactory hedonic capacity in children with autistic trait. It is also the rst study in which psychophysical measurement and self-reporting scales were both used to assess olfactory hedonic capacity. Children with autistic trait exhibit impaired odor identi cation. Interestingly, in the present study there was a pattern of impaired olfactory hedonic capacity measured by CPS-C, but intact olfactory hedonic capacity measured by U-Sniff. Children in the high AQ group were worse than those in the low HQ group at odor identi cation and olfactory hedonic capacity as measured by the CPSC, but the groups did not differ signi cantly in olfactory hedonic capacity as measured by the U-Sniff test. There was also a signi cant association between odor identi cation and olfactory hedonic capacity measured by the CPS-C in the high AQ group, but not in the low AQ group.
In the current study children in the high AQ group exhibited impaired odor identi cation, which is not consistent with Dudova, Vodicka (15) or Brewer,Brereton (14), in which odor identi cation did not differ signi cantly in children/adolescents with ASD and healthy controls. A possible reason for the inconsistent results is the limitations of the only previously available odor identi cation tools, which are restricted by speech and communication capacity. The results of the current study are consistent with previous studies in which abnormal odor identi cation was detected in children/adolescents with ASD (2, 7, 10), suggesting that attenuated odor identi cation may be an endophenotypic marker of children/adolescents with high autistic trait.
In the present study lower olfactory hedonic capacity was evident using the CPS-C, but not the U-Sniff test. This is consistent with a recent meta-analysis in which laboratory-based methods did not detect impaired olfactory hedonic capacity in people with ASD (31). In a previous report it was suggested that hedonic capacity was composed of trait hedonic capacity and state hedonic capacity (32). Trait hedonic capacity was de ned as "non-current" hedonic capacity assessed via interviewing or self-reporting scales, and state hedonic capacity was de ned as "current" hedonic capacity assessed via laboratorybased measures. In the present study children with high autistic trait exhibited signi cantly impaired olfactory trait hedonic capacity and intact olfactory state hedonic capacity.
Robinson and Clore (33) proposed the accessibility model of emotional self-reporting, in which people mainly rely on experimental knowledge when reporting their current feelings, whereas non-current feelings are based on their semantic emotional knowledge, including episodic memory, beliefs about the particular situation, and their general beliefs about themselves (33,34). "Non-current" hedonic capacity includes prospective hedonic capacity for the future and retrospective hedonic capacity for the past. Long-term memories related to a speci c situation play a key role in retrospective hedonic experience (34). In a previous study people with ASD exhibited impaired long-term memory, especially detailed memory (35). Like people with ASD, people with high autistic trait also exhibit lower neuropsychological performance (i.e., memory, attention, and learning) (36). Thus, in such individuals impaired long-term memoryespecially episodic memory-may in uence reported olfactory retrospective hedonic capacity.
More complex cognitive skills are required for prospective hedonic experience, including working memory, imagination, and re ection (37). Strauss (34) surmised that it was essential to generate a mental representation to elicit a hedonic reaction in the present when measuring prospective hedonic capacity, and the mental representation mainly relied on the ability of working memory and the peaking experience of the situation in the past. Notably however, in people with ASD or autistic trait the mental representation may be too abbreviated or it may be di cult to extract the relative peaking experience in the past because of impairment of working memory and long-term memory (35,(38)(39)(40). Thus, in the present study children in the high AQ group reported lower olfactory prospective hedonic capacity than those in the low AQ group. Therefore, lower olfactory trait hedonic capacity may be an endophenotypic marker of ASD, which may re ect the impaired cognitive function associated with the disorder.
In the current study there was a signi cant association between odor identi cation capacity and olfactory hedonic capacity as measured by the CPS-C in children in the high AQ group, but not those in the low AQ group. This further suggests that in children with high autistic trait attenuated odor identi cation may be affected by impaired emotional and cognitive function, but not current hedonic capacity. Odor identi cation processing is performed at the central level, including the orbitofrontal cortex, hippocampus, insula, and prefrontal cortex (23). This brain area is associated with emotion processing, social control, and working memory, which are characteristically involved in ASD (10).
It has previously been reported that hedonic capacity played a key role in understanding social and affective disorders in people with ASD (21,22), and that odor identi cation ability was signi cantly affected by cognitive function (i.e., semantic memory, executive function, and episodic memory) (41) at the behavioral level. Accordingly, abnormally low odor identi cation ability may be a marker of impaired emotional and cognitive function in people with ASD.

Limitations
The current study had some limitations. Only a self-reporting questionnaire was used to screen children for autistic trait. The combination of a self-reporting scale and diagnosis by a clinical expert would result in more accurate screening of individuals for autistic trait. Additionally, although in the high AQ group there was signi cantly lower olfactory trait hedonic capacity but not signi cantly lower olfactory state hedonic capacity, future studies are needed to explore the neural mechanism of this pattern of olfactory hedonic anhedonia.

Conclusions
In the current study children with high autistic trait exhibited poorer odor identi cation ability and olfactory trait hedonic capacity than children without high autistic trait. Poorer odor identi cation ability was also associated with olfactory hedonic capacity in children with high autistic trait. This knowledge about odor identi cation and olfactory hedonic capacity may suggest that impaired odor identi cation and olfactory hedonic capacity are the endophenotypic markers for ASD. The study was approved by the Ethics Committee of Southern Medical University, and conducted in accordance with the principles of the Declaration of Helsinki. Informed consent was obtained from teachers and the children's legal guardians Figure 1 Odor identi cation scores in the high autism quotient group and the low autism quotient group. Note. AQ, autism quotient. *p < 0.01