Autism Research on Detecting Emotions
More often than not, when people see others yawning, they find themselves yawning as well. This phenomenon is known as social yawning and it involves a deeper set of emotions. Yawning in this scenario reflects a person’s empathy for another. Such instinctual display of empathy usually strengthens the social group and the relationship among individuals. However, recent research shows that contagious yawning is not always the case for people on the autism spectrum (ASD, Autism Spectrum Disorder).
Research offers many explanations for the deficiency to perceive emotions typical for the ASD population. The most dominant one is that autistic children tend to confuse the expressions being displayed and therefore find it difficult to interpret them successfully.
Facial Perception in Autism
In 2011, I was visiting the MIT Media Lab and met Dr. Rosalind Picard, an MIT Professor, who leads a number of research projects on assistive technologies for people with autism. Dr. Pickard tells us that many autistic children are brilliant in reading facial expressions if they analyze them on a computer or observe another person from a distance. The distinction, however, arises when we try to measure face-to-face interaction. An autistic child focuses hard on comprehending what we are saying when we talk to them and therefore ignores our facial expressions.
To help autistic children counter these challenges, Picard and her team at the MIT Media Lab are trying to develop special assistive technology for expression analysis. The software uses six affective-cognitive mental states defined by Professor Baron-Cohen from the University of Cambridge:
The technology tracks the facial points, monitors face transitions, records the head poses and extracts the facial features. As the facial expressions change, the software keeps recording the degree of each emotion as seen in the different expressions. Professor Picard emphasizes the importance of dynamic analysis for face transitions. The problem is that static face expressions are not always representative of the expressed emotion and it is the history of face transitions that gives us cues to deciphering another person. For example, if someone looks confused as they didn't understand or missed something in our speech, we might mistakenly perceive their facial expression as disagreement with our statements.
It turns out that, based on the dynamic analysis of facial transitions, the computer can easily detect what the person is feeling. When tested on different categories of contexts and behaviors, the computer software developed at MIT Media Lab appeared to be more successful in recognizing facial transitions than people in general. This technology is a scientific breakthrough and marks a significant step towards availability of mainstream assistive tools for individuals with Autism.
Dr. Mari Davies and Dr. Susan Bookheimer, neuropsychology researchers from the University of California, Los Angeles, conducted a study to compare the brain activity of 16 typically developing children and 16 high-functioning autistic children. These children were subjected to a series of faces showing emotions of anger, fear, happiness and neutral expressions while undergoing Functional Magnetic Resonance Imaging. Half of the faces had their eyes averted, the other half stared directly back at the children.
It was found that, Ventrolateral Prefrontal Cortex (VLPFC), the part of the brain which evaluates emotions, became active when the direct-gaze faces came up and quieted down when the averted-gaze faces were displayed to the typically developing children. However, the autistic children showed no reaction to either set of faces. This shows that autistic children do not perceive any difference in emotion whether the face stares back at them or looks away from them.
Emotions are of second nature to the typically developing children; however, for autistic children recognizing emotions is a very difficult process. Yet, autistic children are often able to recognize simple emotions. In a study conducted by Professor Baron-Cohen, it was found that autistic children could make out faces that showed happy or sad emotions but had difficulty identifying faces carrying expressions of surprise or fear.
According to Dr. Angelique Hendriks from Radboud University, the reason for this deficiency could be a weak central coherence. This term defines the inability of autistic children to combine the parts of information or signals they receive into one whole coherent picture. This is why they treat different parts of information separately and are unable to connect and relate them to the situation at hand.
Dr. Ellie Wilson, in her PhD research at Macquarie University, tested the hypothesis of whether autistic children can match images onto real life people. The study demonstrated that the key difference with neurotypical children is in the way autistic children move their eyes around the face. It may be possible that training might improve their recognition skills, though the results from a few training studies in the past haven’t been particularly convincing.
Among many problems faced by autistic children, having no perceptual ability to read facial expressions is the most serious and pressing of them all. Researchers and technologists are working together to develop mechanisms which will aid the learning of autistic children and help them navigate in the social world.
This article was written by:
Dr. Tali Shenfield is a Child Psychologist and a Clinical Director of Richmond Hill Psychology Center. She is a member of the Canadian Psychological Association and Ontario College of Psychologists. When not busy with psychological assessments and psychotherapy, she is writing on topics related to parenting and psychology. You can read her blog here.
To be a guest author on the Your Mental Health Blog, go here.
Author, G. (2014, January 28). Autism Research on Detecting Emotions, HealthyPlace. Retrieved on 2021, October 18 from https://www.healthyplace.com/blogs/yourmentalhealth/2014/01/autism-research-on-detecting-emotions