Arm Movement Abnormalities May Aid Autism Diagnosis

Batya Swift Yasgur, MA, LSW

January 29, 2018

Subtle abnormalities in arm movements may provide objective information that may assist in more accurately diagnosing autism spectrum disorder (ASD), new research shows.

Researchers used sophisticated high-definition electromagnetic sensors to analyze continuous arm movements in children and adults, both with and without autism, who were reaching out to touch a screen. Larger flucturations in motion were associated with greater risk for ASD and also with more severe forms of the disorder.

"In collaboration with researchers at Rutgers University, we measured the motion of arms with much deeper precision than what can be detected with the naked eye," lead investigator Jorge José, PhD, James H. Rudy Distinguished Professor of Physics, Indiana University, Bloomington, told Medscape Medical News.

"No two people are the same; each of us moves in a different way, and if you look at the micromovements microscopically, you find very clear differences between autistic children and each other, and even more importantly, between autistic children and controls, which can then be used for diagnosis," he said.

The study was published online January 12 in Scientific Reports.

Noise and Signals

Diagnostic criteria for neurodevelopmental disorders (NDDs) and ASD are largely "based on the symptoms directly observable by clinicians or parents, with few biological foundations," the authors write.

Symptom-based criteria are "not reliable enough to provide insights into underlying neurobiological mechanisms," they add.

Recent research has "mainly focused on studying and understanding brain circuits potentially connected to neurodevelopment and NDD," but it is "of equal importance to study the corresponding behavioral outputs as well," the authors suggest.

In particular, the "information contained in behavioral outputs, especially in movements, seems to have been mostly overlooked."

The authors explain that even seemingly simple movements are "actually produced via rather complex integrations in the sensorimotor system" and "arise through a complex learning process directly related at each stage in neurodevelopment."

One of the basic motions that a typical adult engages in multiple times a day is reaching for an object. Such "goal-directed reaching" is learned during infancy and is stabilized at developmental periods during childhood.

The nervous system has "different sources of stochastic noise at different temporal and spatial length scales for every motor process involved." This noise is naturally produced by random neuron firings in the brain.

Previous ASD reaching studies have found "increased trial-to-trial spatial variability and increased speed magnitude variability, compared to TD [typically developing] individuals, suggesting increased level of noise in the ASD sensorimotor systems," the authors note.

Instead of studying trial-to-trial variability, the researchers set out to "examine the continuous speed profiles within each motor cycle continuously at millisecond timescales" using "novel and robust quantitative measures."

"In the past, clinicians or people researching movement noticed some movement irregularities in people with autism but thought it was just noise — meaning, a velocity that has many peaks — so they averaged the noise to get smooth movement trajectories," Dr José said.

"We hypothesized that, in fact, there may be some important physiological signals in the noise because the CNS [central nervous system] sends orders to our limbs to move, so we decided to do measurements that looked at the relationship between signal and noise," he recounted.

Imperceptible Movements

The researchers evaluated 71 participants, including 30 patients with ASD (aged 7 to 20 years), 15 TD adults (aged 19 to 31 years), six TD children (aged 3 to 5 years), and 20 parents of ASD children who did not themselves have ASD.

Of the ASD participants, 18 completed four psychometric tests: IQ, the Vineland Adaptive Behavior Composite assessment, the Autism Diagnostic Interview Revised (ADI-R), and the Autism Diagnostic Observation Schedule (ADOS).

The remaining 12 ASD participants were diagnosed by certified clinicians and were classified as being high- or low-functioning.

All participants were asked to point to a target at the center of a touch screen. They then retracted their hand (right hand, dominant hand), after which the target disappeared and then reappeared, at which time the next reaching gesture commenced.

The participants' hand motions were continuously captured by high-speed, high-resolution electromagnetic sensors (240 Hz). Changes were tracked at every point of the arm movement.

The researchers applied a triangular smoothing algorithm to the velocity profiles calculated from the recorded position data in each direction so as to separate electromagnetic sensor noise from actual signal information.

After the assessment, each participant received a score that was based on the level of hidden speed fluctuations in their movement. A lower score indicated a greater risk for autism.

The average speed profiles were similar for all participants. However, the continuous speed profiles, in particular, the fluctuation with each motion cycle at millisecond time scales, differed significantly for participants with ASD in comparison with TD adults.

"These fluctuations were imperceptible and not visible to unaided eyes during the experiment," Dr José emphasized.

Strikingly, the movement-based diagnoses were positively correlated with the psychiatric assessments in the presence and severity of ASD (P < .01 for the Vineland Composite score and P < .05 for IQ).

There was a modest negative correlation with ASD severity level in the spectrum and ADI-R scores, with P values slightly higher than .05.

The ASD participants who were classified as high-functioning were found to have fewer fluctuations than those classified as low-functioning.

Moreover, 17 of 20 parents (10 mothers and 7 fathers) of ASD children, who did not themselves have ASD, also displayed higher levels of noise.

"It came as quite a surprise to us to find that a single motion could provide so much information," lead author Di Wu, a PhD candidate at Indiana University, told Medscape Medical News.

"What was even more surprising is that this method not only separated children with autism from typically developing children but also gave a qualitative correlation between the way autistic children move and the other psychometric scores, such as IQ," she said.

Dr José called the correlation "incredible" and "the most important finding in our paper."

Early Intervention

Commenting on the study for Medscape Medical News, Susan Hyman, MD, professor of pediatrics and Chief, chief, Division of Developmental and Behavioral, Pediatrics, University of Rochester Medical Center, New York, who was not involved with the study, said that the "search for novel ways to physiologically measure the neurobehavioral symptoms of ASD is an important undertaking as we search for the symptoms that can be addressed with specific interventions that might impact outcome."

She commended the investigators for "cleverly using technology to examine the functional behaviors observed in people with ASD and establish the potential to trace backward into the brain the physiologic and biologic implications."

Nevertheless, she expressed some concerns.

"The investigators have interesting preliminary data; however, the subject selection complicates interpretation" because "the current subjects are not matched by age or IQ, and the authors do not include data on executive functions, medications used, and measures of attention and/or impulsivity."

Additionally, "comparing individuals with ASD across a wide age span on the same test may introduce maturational features into the data."

She noted that medications used by some patients with ASD "may impact the speed of movement."

Also commenting on the study for Medscape Medical News, Lawrence Reiter, PhD, professor, Department of Neurology and Department of Anatomy and Neurobiology, University of Tennessee Health Science Center, Memphis, who was not involved with the study, called it "important, since there are currently no good testable biomarkers for autism."

He noted that, "although some people now use eye tracking as a proxy, I think this study gets to the key motor features indicating possible ASD.

"This and other subtle biomarkers like eye tracking could be used in concert early during childhood development to identify patients who are at higher risk for ASD later in life," he added.

Dr José agreed, noting that early diagnosis can facilitate early intervention. He added that if a parent is tested and has the same autisticlike movement characteristics, it increases the probability the children have ASD.

Dr José foresees a time when this technology can be available as a wearable device or downloaded as an app to a cellphone, thereby providing more extended data as to the movement of the child.

Dr Jos é is the recipient of a Presidential International Fellowship Initiative from the Chinese Academy of Sciences. Di Wu and Dr Jos é received partial support from a National Science Foundation (NSF) grant. Coauthor Elizabeth Torres, PhD, received support from an NSF Cyber Enabled Discovery and Innovation Type I (Idea) grant and a Development Career Award from the Nancy Lurie Marks Family Foundation. Dr Hyman and Dr Reiter have disclosed no relevant financial relationships.

Sci Rep. Published online January 12, 2018. Full text

For more Medscape Psychiatry news, join us on Facebook and Twitter.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: