Discover SPARK

Lessons from a Study of Severe Autism

Marina Sarris

Date Revised: July 21, 2022

A sister project to SPARK is expanding what we know about severe autism, challenging behaviors, and mental health.

The Autism Inpatient Collection, or AIC, studies autistic youth who were admitted to special psychiatric units in several U.S. hospitals. Many of these children and teens have few spoken words, intellectual disability, or challenging behaviors that interfere with everyday life.

The AIC has enrolled more than 1,400 participants ages 4 to 20 since the study launched in 2014. That is significant because youth with severe autism often face barriers to participating in research. They are studied less often than others on the autism spectrum, for many reasons, according to child psychiatrist Matthew Siegel, M.D. Siegel researches severe autism for the AIC, and he is the founder of the Autism and Developmental Disorders Inpatient Research Collaborative.

Although hospital stays are relatively uncommon, children with autism are six times more likely to be hospitalized for psychiatric problems than other children.1 Siegel discusses psychiatric hospital stays for people with autism in this recorded SPARK webinar.

What Do We Know about Autism and Mental Health?

Researchers studying these hospitalized patients have learned several things about autism and psychiatric conditions. Some of their findings include:

  • Youth who spoke fluently experienced more depression, and more defiant and argumentative behavior, than those who did not.2
  • 22 percent of the youth who spoke fluently had talked often about death or suicide.3
  • About one in four had been the victim of physical, emotional, or sexual abuse, according to their parents. A small percentage of those youth also had post-traumatic stress disorder, also called PTSD.4
  • Several factors increase the chances that an autistic child or teen will be hospitalized for psychiatric reasons. Having a mood disorder – such as depression or bipolar disorder – is number one. The second most common factor is having a sleep disorder.5

“Having sleep problems was a powerful and independent risk factor for psychiatric hospitalization,” explains Siegel, who is vice president of medical affairs for the Developmental Disorders Services at Maine Behavioral Healthcare.

Sleep problems are common in children with autism. And these conditions can be treated, which could possibly reduce the chance that someone will be hospitalized. “It’s something we should really be paying attention to,” Siegel says.

Unraveling Challenging Behaviors in Autism

Another AIC study appeared to challenge a “prevailing assumption” about the relationship between speech problems and challenging behaviors in autism.6 Families and professionals alike often assume that poor communication skills cause frustration, which leads to behaviors such as hitting oneself (self-injury), striking others (aggression), and tantrums. But researchers found that someone’s ability to cope and adapt was the biggest influence on those behaviors, not the ability to speak.

Another study shows the power of conducting behavioral research in a hospital, where participants can be closely monitored,7 Siegel says.

In the new study, which was led by Jesse Northrup, Ph.D., and Carla Mazefsky, Ph.D., hospital staff observed more than 50 participants for hundreds of hours and took continuous data about their emotions and behavior. Staff documented episodes of aggression, self-injury, and emotion dysregulation, “which is our fancy word for, basically, tantrums,” Siegel says. Staff made note of visible signs that the youth were upset, such as facial expressions, crying, or yelling.

Most people think that a person will become visibly upset before they hurt themselves or others, he says. Although that does happen, the researchers found something else: many youth had episodes of aggression and self-injury before they showed visible signs of being upset.

“That was a surprise to us,” Siegel says. “Perhaps we have to think about the relationship between emotions and behavior differently than we typically do.”

It is possible that the youth did become upset first, but it was simply not visible, he says. “Is there emotion dysregulation before the aggression or self-injury, but it’s internal and unobservable? That’s a remaining question.”

The finding does confirm reports from some parents and caregivers, who say they sometimes cannot predict when their children are about to hurt themselves or others. Knowing when a child or teen is likely to have a serious behavior could help caregivers prevent injury to the youth or other people.

“Some parents have said, ‘I don’t know when it’s coming and therefore I always have to be prepared for the worst,'” he says.

The Genetics of Autism

Like the SPARK study, which AIC members can also join, the autism inpatient study analyzes the DNA of participants who chose to provide a sample. Researchers are hoping to uncover more genes that contribute to autism, so they can better understand how autism affects people’s development and health. The SPARK and AIC studies are funded by the Simons Foundation Autism Research Initiative.

Siegel says that DNA data analysis from the AIC participants should be completed soon. One question researchers hope to answer is whether people with severe autism are more likely to have variations in autism genes than others on the spectrum.

“Does this AIC group, which has a high percentage of people with more severe autism, have a higher rate of detectable genetic differences? And secondly, do those differences cluster in interesting ways?” Siegel asks.

Previous research suggests that people with autism and intellectual disability are more likely to have variations in autism-related genes.

Because this study took place in hospitals, it made it easier for researchers to collect DNA samples from participants for analysis.

Barriers to Participating in Autism Research

Removing barriers to research participation is one of the goals of the inpatient study. Youth with severe autism may find it difficult to give a blood or saliva sample, or otherwise participate in many studies. Some autism studies even exclude youth who have limited language or significant developmental delays. That is because the tests used for the research require that participants understand a certain amount of language.8

Jill Escher, a SPARK participant, knows about these challenges as the mother of two children with autism. “Most so-called autism studies that come across my desk are completely irrelevant to people with classic autism like my children. And further, they would not be able to participate in any study that required verbal ability, following directions, sitting in scanners, etc. My son would need several large men to accompany him to hold him for any blood draws and to ensure he didn’t damage any equipment,” says Escher, president of the National Council on Severe Autism, an organization that advocates for services and research.

Research Trends in Autism

Researchers hope that the AIC study will add to our knowledge about people on the severe end of the spectrum, and how to help them.

In a 2018 article from his group, Siegel reported that research on treatments for severe autism had declined over the years, while research on milder forms of autism had grown. He was not alone in that assessment. Other autism experts have said that older children who speak few, or no, words were neglected by research.8

Today, Siegel says, he does not know if there has been any change in the proportion of research devoted to severe autism. “Certainly there appears to have been no increase, so it is likely the historical trend continues,” he says.

Recently, the Lancet commission, a group of autism experts, families, and self-advocates, recommended using the term “profound autism” for people who need around the clock care and supervision. The recommendation is controversial among those who believe it would divide the autism community.

Siegel hopes that the attention on severe autism will ultimately help. “I am hopeful that, with the Lancet commission identifying and putting forward the term profound autism, there will be more research with this sub-population,” he says.

Interested in joining SPARK? Here’s what you should know.


Current AIC study sites:

Previous AIC study sites:

  • Bradley Hospital, Rhode Island
  • Children’s Hospital Colorado
  • Sheppard Pratt Health System, Maryland


  1. Croen L.A. et al. Pediatrics 118, e1203-11 (2006) PubMed
  2. Lerner M.D. et al. J. Autism Dev. Disord. 48, 3689-3701 (2018) PubMed
  3. Horowitz L.M. et al. J. Autism Dev. Disord. 48, 3702-3710 (2018) PubMed
  4. Brenner J. et al. J. Autism Dev. Disord. 48, 3727-3735 (2018) PubMed
  5. Righi G. et al. J. Autism Dev. Disord. 48, 3647-3657 (2018) PubMed
  6. Williams D.L. et al. J. Autism Dev. Disord. 48, 3668-3677 (2018) PubMed
  7. Northrup J.B. et al. Autism Res. Epub ahead of print (2022) PubMed
  8. Tager-Flusberg H. and C. Kasari Autism Res. 6, 468-478 (2013) PubMed