Social anxiety: The elephant in the room of autism

Studies on "theory of mind" and autism seem to assume, at least implicitly, that the better a person's ability to observe and understand others, the greater their real-world social success--and vice versa. The reality is a lot more complex, because of a third variable: social anxiety.

Consider a young couple I know.  The boy, outgoing and expressive, has many acquaintances, a dozen or so friends, and a group of close friends with whom he has stayed in touch since middle school.  His parties successfully bring together friends from his middle school, high school, college, and post-college days.  The girl, perfectly NT, has a few online friends and a couple friends from high school and college.  She doesn't throw parties.  She socializes most often with her boyfriend's friends and family.

So, does the boy observe and understand other people better?  No.

The boy has a poor memory for names and faces.  Since graduating college, several acquaintances have greeted him briefly on the street and he has had brief, pleasant conversations with them, though he could not recall their names or the context in which he met them. He almost never remembers what he said to whom, and sometimes forgets what others have said to him.  He relies on the girl to remember these things for him.  His timing is bad, and he can come off as awkward, particularly on the phone.  By contrast, the girlfriend reads people well, and she can tell you in detail who said and did what at any of her boyfriend's parties.

So, who has better social skills: the one with better social perception and cognition, or the one with more friends?

The boy enjoys social success despite his poor social memory and awkwardness because he is outgoing, with an expressive face and engaging smile.  The girl has few friends despite her social perceptiveness because she experiences social anxiety.  When in groups, she barely says a word.  She responds to others' overtures, but rarely initiates interactions herself.

The larger point: social anxiety plays an important role in social interactions.  People often think of it as a response to social failure, but as with this girl, it can also create social failure.  When the girl approaches a group, she stands too far away from them, literally placing herself outside the group.  They behave accordingly, making little eye contact and no verbal overtures.  She, sensing the literal and figurative distance, feels isolated.  Socially anxious people, like this girl, have fewer friends and more difficulty dating than other equally socially-adept people because they initiate fewer conversations, and may make it harder for others to approach them.

Children with autism often bring social anxiety to their interactions, for all sorts of reasons.  If they have often been criticized or rejected, they may fear similar treatment in the future.  If they know about their diagnosis, they may fear misunderstanding someone or saying the wrong thing.  

In fact, a recent study found that simply being invited to play increased autistic children's stress levels.  Furthermore, the older the children were, the more it increased their stress.

In this study, there were no large, noisy groups, and no one teased or rejected the child.  If the autistic child did not want to play, the other children simply accepted it and played in close proximity.  Nonetheless, the prospect of playing with another child was stressful.  As the authors put it:
We reasoned that if a participant exhibited social stress under such benign conditions, then responses may be even more notable on a typical school playground with many children, enhanced stimulation and challenging social exchanges.

How the study worked
This study measured stress using a chemical called cortisol, which the body releases during stress to help it return to its normal, pre-stress state.  Cortisol is released and builds up slowly in the body, and so is best measured about 20 minutes after a stressful event.  Researchers collected saliva to measure cortisol at key points in the study.  They defined the 40% of participants with the highest cortisol response to stress as the "cortisol responder group" and the 40% with the lowest response as the cortisol nonresponder group, to look at the effects of stress on social behavior.

Participants were 21 children with autism and 24 typically developing children (equal numbers of autistic and typically developing children were tested, but three with autism had unusable data).  They ranged from 8-12 years old, with an average age of 10.  Autistic participants had an IQ of at least 75 and a diagnosis of autism (not Asperger's or PPD-NOS), and had no co-occurring disorders such as seizures or Fragile X.  To avoid including neurotypicals with below-average social skills, neurotypical participants with scores greater than 10 on a test of social and communication ability were not included (compared to a threshold of 15 for possible ASD). 

What happened during the experiment:
  1. Participants, at home, measured their cortisol levels several times a day for several days to get a reliable baseline.
  2. Participants came to the lab and sat undisturbed in a quiet waiting room for about 15 minutes.  During this time, researchers collected the first cortisol sample.
  3. Children with autism went outside to a playground just behind the lab, as did two other children: a typically developing participant and another child, a typically developing confederate trained by the experimenters.  All three children were allowed to play freely for five minutes.
  4. After five minutes, the confederate approached each participant and invited him to play on the equipment.  If the participants said no, the confederate repeated the invitation up to two more times, each a minute apart.  
  5. After ten minutes, a box of toys designed for cooperative play (balls) was introduced to the playground, and the children continued to play freely.
  6. After fifteen minutes, the confederate again invited each participant to play, this time with the balls.
  7. After twenty minutes, participants went back inside and another sample of cortisol was collected.
  8. Researchers collected two more samples of cortisol 20 minutes and 40 minutes after the playground experiment.

What the study found
First, not surprisingly, typically developing children interacted with others more than children with autism did.  Overall, they spent more time interacting whether playing with playground equipment or cooperative toys.  Children with autism initiated fewer interactions, and rejected others' invitations to play more often.  However, they did not simply avoid other children: instead, they often came near other children, then left without interacting, much as socially anxious people hang around the fringes of a group without joining in.  

The table below compares how frequently autistic and typically developing children interacted using the playground equipment ("motor play") or the balls ("cooperative play").  Notice that children with autism show a much wider range of willingness to interact.  Particularly in the cooperative play condition, a number of them interacted as often as neurotypical children.  However, there were more autistic children who interacted less often, bringing down the group average.  Meanwhile, a few neurotypical outliers interacted with a frequency more typical of the autistic group.  Thus, one cannot necessarily tell how much a child will interact with others based on diagnosis as autistic or NT. 
The table below breaks down the interaction patterns for autistic and typically developing children during both motor and cooperative play.  Some of these patterns are the straightforward ones described earlier: for example, during group motor play, children with autism spend 46% of the time interacting compared to 79% for NTs; during cooperative play, children with autism spent 59% of the time interacting compared to 91% for NTs.

Oddly, while children with autism reject invitations to play about equally often whether invited to motor or cooperative play, NT children rejected cooperative play invitations more often (.41 vs. .18).  This difference was not tested for statistical significance, and the reasons are unexplored.  Perhaps NTs perceive ball games as potentially higher-stake and higher-risk than simply playing with playground equipment?  However, NTs initiated cooperative play more often than motor play, while autistic children showed the opposite pattern, frequently initiating motor play but almost never initiating cooperative play.  All of this suggests the complexity of decisions to interact even for NTs, and the fact that the sort of play involved matters.

The most interesting finding was that autistic children's cortisol levels during the study increased with age, while typically developing children's did not change over development.  See the figure below, where the crucial time points are S2 (immediately after play), S3 (20 minutes after play), and S4 (40 minutes after play). Younger autistic children showed lower cortisol levels even than the NT children, while older ones had much higher cortisol levels than the NTs, particularly after interacting on the playground.

What does it all mean?
The authors argue:
At first glance, it may be interpreted that socialization [for autistic children], at least with peers, is stressful and may be deemed threatening in some way. Interestingly, however, many of the children with autism did not completely remove themselves from the interaction as one might expect from a threatening situation. Older children with autism tended to play more in a group on the equipment during both motor and cooperative play and avoided less, indicating that they were voluntarily engaged in the interaction. The younger children tended to avoid interacting at a higher rate than the older children with autism.  Thus, taken together, it may be that children with autism responded to the playground situation with an approach/avoidance conflict; they wanted to play but found it possibility threatening. We conjecture that younger children with autism responded to the conflict with greater avoidance, while older children with autism yielded to the attraction...
Why does this matter?
The fact that the children with autism exhibited increased stress under such benign social conditions is concerning. Social stress may be even more significant under conventional school playground environments with more children, increased stimulation and more rejecting social exchanges.
The authors interpret older autistic children's increasing social stress as being due to "preparation for social interaction amid greater awareness of their own social imitations."  In this particular paper, they do not mention another, more obvious cause: the teasing, bullying and rejection many autistic children face, which may lead them to fear interacting with others.

However, the authors do point out that, as the elephant in the room, social anxiety must be accounted for and incorporated into educational interventions.  First, developing social skills is not enough; one must also learn to cope with stress:
 "Teaching coping strategies along with social skills may go far in improving social competence while ameliorating the increased reactivity to novel social situations."
Second, the nature of the social skills taught and the educational methods used should change based on a child's age and degree of social anxiety.  
It may be reasoned that the older child characterized by more social interactions, increased self-awareness and higher biological and psychological stress would likely benefit from treatment comprised of direct social skills training, stress reduction techniques and safe opportunities to interact and practice skills with others. Conversely, younger children characterized by withdrawn behavior, reduced motivation and lower stress may benefit from treatment approaches that utilize video or computer media and less reliance on face-to-face interaction to acquire skills.
I completely agree. I would add that because of the harmful effects of social anxiety, a child with autism may not even be able to use the social skills he learns unless he copes with his social anxiety.

Corbett, Blythe A., Schunn, Clayton W., Simon, David, Ryan, Niles & Mendoza, Sally (2010).  Elevated cortisol during play is associated with age and social engagement in children with autism.  Molecular Autism, Vol. 1.