乌鸦传媒

Abnormalities in the olfactory response to pleasant and unpleasant odors may be related to an underlying deficit in sensorimotor control for children with autism spectrum disorder (ASD), researchers found.

In a study involving 18 children with ASD and 18 typically developing (TD) children, those with ASD failed to modulate the magnitude of a sniff -- that is, the strength of the nasal inhalation, or "sniff response" -- upon exposure to pleasant versus unpleasant odors, according to doctoral student , of the Weizmann Institute of Science, and colleagues.

In contrast, TD children had an increased sniff response for pleasant odors relative to unpleasant odors, with the sniff strength lower by about 30% for unpleasant versus pleasant aromas (mean 0.665 versus 0.918 normalized flow units, P<0.05 after Bonferroni correction), the researchers reported online in

Importantly, among children with ASD, greater abnormalities in the sniff response were associated with greater social impairment (r=-0.72 for social affect, P<0.001) and overall autism severity (r=-0.75, P<0.0005), as measured by the Autism Diagnosis Observation Schedule (ADOS). But there was no relationship between the sniff response and performance on a battery of standardized motor assessments (P≥0.18), suggesting that the findings were not related to a generalized motor impairment.

The investigators also used the results to develop a computer program that could classify individual children as ASD or TD based on the characteristics of their sniff response alone. When tested on the children individually, the classifier correctly identified 17 of 18 TD children and 12 of 18 ASD children (81% accuracy, P<0.001).

Rozenkrantz and colleagues suggested that quantification of the sniff response could "provide for a novel early non-verbal non-task-dependent ASD marker."

The sniff response was recorded using a custom-designed, double-barreled, pediatric nasal cannula, which delivered either pleasant (roses or shampoo) or unpleasant (sour milk or rotten fish) odors while recording nasal airflow. Participants -- mostly boys, mean age 7 -- watched a cartoon during the 10-minute task.

Prior studies of olfaction in ASD typically required verbal or written instructions, as well as verbal or written responses. "The sniff response is largely devoid of these limitations, as it consists of a language-free task-free automatic odor-specific response," Rozenkrantz and colleagues wrote.

As such, they asserted, examining the sniff response eliminates the potential confounders of task comprehension and motivation.

Another benefit of the measure comes from the speed at which the researchers were able to obtain their measurements. "As a clinician, you want something fast that tells you there is a problem here," , associate professor of psychology at Rutgers University, who was not involved with the study, told 乌鸦传媒. "You have a limited number of minutes with the child in your clinical practice to make a decision as to what course of treatment you are going to follow and to actually provide a set of recommendations to the parent," she continued.

But Torres was wary of placing too much faith in the relationship between an objective measure, such as the sniff response, and the observational measures of autism severity and social impairment.

"[The ADOS] is a diagnostic tool that is very useful to identify the problem, but it should not be used to guide our research," Torres explained. "It is much too noisy to be correlated with anything that comes from a physical measurement of something that the body is experiencing," she said.

The authors described their findings in the context of internal models of action (IMAs), or mappings within the brain that relate a prediction of the sensory consequences of an action to the actual sensory feedback experienced. Prior research has suggested that IMAs may be impaired in children with ASD, and that this may lead to social dysfunction.

But the relationship between the sniff response and IMAs is not well defined in the paper, and Torres was concerned with the lack of discussion regarding feedforward/feedback control, or how the brain estimates the sensory consequences of the upcoming action. This, she explained, "is at the heart of internal models."

Torres suggested that abnormalities in the reward system, or the brain network that qualifies a positive or negative experience, may contribute to the lack of sniff modulation in children with ASD.

However, the potential contribution of the reward system was not addressed in the paper. Rather, like many recent studies of ASD, the authors "aim for establishing a link with social interaction issues," Torres noted. The use of such social and diagnostic measures has become common across the field, and Torres explained that, "instead of science leading the clinical practice, the clinical practice is leading the science."

Comment