Holistic and Part-based Face Recognition in Children With Autism

Joseph RM, Tanaka J
Department of Anatomy and Neurobiology, Boston University School of Medicine, Massachusetts rmjoseph@bu.edu

Abstract: J Child Psychol Psychiatry. 2003 May;44(4):529-42

Findings

Research shows that compared with typically developing children, those with autism cannot recognize faces readily. In normally developing children, when faces are turned upside down, recognition is known to be severely disrupted, much more than after inversion of non-face objects. This has been taken as evidence that faces are recognized holistically, i.e., all at once as an entire unit. Some researchers suggest that the child with autism cannot easily recognize a person’s whole face at once and instead they rely on face parts to try and put them into the context of the whole face. However, this theory has never been tested in children with autism.

Dr Joseph and his colleagues studied face recognition in two experiments. Study 1 studied typically developing children: 27 normal 9-year-olds and 30 normal 11-year-olds. Children were asked to look at whole faces and match them to the same face or a similar face differing in one feature (eyes, nose, or mouth). Study 2 studied children with autism: tests compared 22 children diagnosed with high-functioning autism to 20 non-autistic children with language impairment or delay. Children saw only the isolated face part (eg, mouth) that was different in the face comparison test and they were asked to identify that face part that differed in the whole face. Both upright and inverted faces and face parts were shown to all children.

As expected, typically developing children in Study 1 were better at recognizing face parts represented in the whole face than in parts in all the upright faces but not inverted faces. These children were most accurate in identifying faces based on differences in eyes. In Study 2, autistic children showed an advantage over typical children in recognizing differences in the mouth, but were far less efficient than normal children at telling differences in faces from the eyes. The non-autistic language-delayed children performed more closely to the typically developing children.

Conclusions

Autistic children process faces holistically and this was mainly evident when recognition depended on the mouth. These findings suggest that face recognition abnormalities in autism are not fully explained by an impairment of whole face processing, and that there is an unusual significance of the mouth region when children with autism process information from people's faces. The authors suggest this may be a result of the autistic child’s specific impairment in processing information from the eyes, or because of an aversion to looking at eyes, so that the mouth takes on greater significance as a primary medium of communication for the autistic child. This is consistent with studies that show children with autism are delayed (by several years) in spontaneously following shifts of gaze from others, and they depend on vocal cues to establish attention. Another possibility for these findings is that autistic impairments in language functioning foster an early and enduring tendency to attend to mouths so they can make sense of speech via lip reading, especially when other communicative cues from the eyes are inaccessible.