Tracking eye movements proves informative for the study of gaze direction detection in autism

TL;DR: This article found that participants with autism were significantly less accurate at interpreting eye direction and detecting a target item (array sizes 4 and 6 items) than typically developing participants of comparable nonverbal ability.

About: This article is published in Research in Autism Spectrum Disorders.The article was published on 2009-07-01 and is currently open access. It has received 27 citations till now. The article focuses on the topics: Eye tracking & Task analysis.

Summary

Introduction

• Considerable research effort has been dedicated to exploring how well children with autistic spectrum disorders infer eye gaze direction from the face of an actor.
• The current research emphasises the value of tracking eye movements for revealing components of task performance in domains known to be of difficulty.
• Furthermore, even when this ability does develop, it remains severely delayed relative to participants of equivalent mental age (Leekam, Hunnisett, & Moore, 1998).
• The development of gaze direction detection and gaze perception are likely to follow atypical (and independent) pathways (Webster & Potter, 2008).

Participants

• All participants had been diagnosed by clinicians and satisfied the diagnostic criteria for autism according to the DSM-IV (APA, 1994).
• All participants completed a calibration phase at the beginning of the study and all were reported to have normal or corrected-to-normal vision.
• For typically developing participants teachers completed the Strengths and Difficulties Questionnaire (Goodman, 2001), scoring each individual on emotional symptoms, conduct, hyperactivity, peer relationships and pro-social behaviour.
• This measure of nonverbal ability is one of the most frequently used matching measures (Mottron, 2004) and provides a quick assessment across a wide age range.

Materials

• Colour digital photographs were taken of a woman directing her gaze towards one of several targets.
• In the six-item array, targets were placed on either side at 10˚, 20˚ and 30˚ visual angle from the model’s midline.
• The system was transported to the testing location of each individual, who sat approximately 50cm from the screen during testing.
• The eye tracker was calibrated for each participant using a 9 point calibration of each eye.
• The Tobii system recorded gaze behaviour to pre-specified areas of interest (AOI).

Autism and Developmental Disorders.

• Scan Path Differences and Similarities During Emotion Perception in those With and Without Autism Spectrum Disorders.
• Eye Direction Detection Improves with Development in Autism, also known as Brief Report.

Figures

Figure 1: Task example, the 6-item array task with gaze cues provided by the eyes and head (target item ‘sheep’).

Table 1: Accuracy (percentage correct) and Gaze Behaviour (fixation length in milliseconds) as a Function of Cue and Group Membership.

Figure 3: Examples of the ‘gaze plots’ for the 6-item array task trial depicted in Figure 1, with gaze cue provided by the head+eyes. Target item is the sheep. Images display the gaze plots of i) a participant with autism and ii) their matched comparison participant who is developing typically

Figure 2: Target fixation length (ms) as a factor of task accuracy, combining cue type for each array size

Frequently asked questions

Q1. What are the contributions in this paper?

Here the authors combine task performance ( accuracy to correctly label a target item ) and eye movement information ( ‘ where ’ the participant fixates when completing the task ) to understand more about the components involved in completing eye direction detection tasks. The data have implications for the design of tasks for individuals with autism and provide insights into the usefulness of including measures of visual attention in understanding task performance. RUNNING HEAD: eye-tracking, gaze direction & autism Tracking eye movements proves informative for the study of gaze direction detection in autism Autism is a neuro-developmental disorder characterized by qualitative impairments of social communication, accompanied by unusual repetitive or stereotyped behaviours ( DSM IV, American Psychiatric Association, 1994 ). Clifford and Dissanayake ( 2008 ) studied home videos of infants later diagnosed with autism, identifying poor quality and timing of eye contact even during the first year of life. The development of gaze direction detection and gaze perception are likely to follow atypical ( and independent ) pathways ( Webster & Potter, 2008 ). This suggestion is supported by research showing more pronounced problems for younger children with the disorder, indicating an unusually protracted development ( Webster & Potter, 2008 ). With an abundance of research revealing deficits in the domain of social cognition and attention in autism there have been calls for a shift away from characterising overall task performance and towards the study of the processes and strategies used to perform these tasks ( Klin, Jones, Schultz, Volkmar, & Cohen 2002b ; Volkmar, Lord, Bailey, Schutlz, & Klin, 2004 ; Boraston & Blakemore, 2007 ). Information about which aspects of the stimuli the participant fixates upon can provide insights into the strategies they are using. The current study extends the combined use of eye-tracking and behavioural performance to the study of gaze direction detection, a domain that is regularly implicated in explorations of autism. In typical development the evidence therefore suggests that explicit judgement of eye-direction is a skill that arises by approximately 3-years of age and continues to gradually improve in precision during early childhood. Furthermore, even when this ability does develop, it remains severely delayed relative to participants of equivalent mental age ( Leekam, Hunnisett, & Moore, 1998 ). Evidence supporting this suggestion comes from apparent dissociations between the ability to identify where someone is looking ( Leekam et al., 1997 ) and a lack of spontaneous gaze monitoring ( e. g. Mundy, Sigman, Ungerer, & Sherman, 1986 ) which prevents triadic interactions. Boraston and Blakemore ( 2007 ) suggest that eye tracking is particularly useful when it is combined with tests of cognitive performance because it provides information in addition to a participant ’ s overall test score. 

Q2. How many pupils with autism attended schools?

Five attended the special education unit of a mainstream secondary school and 10 attended schools for pupils with additional educational needs. 

Q3. What was the use of the Tobii 1750?

The Tobii 1750 eye-tracking screen was interfaced and controlled by a Dell Latitude D820 computer for presenting stimuli and recording responses (using Tobii Studio software). 

Q4. How many participants with autism were matched to a typically developing child?

Each individual with autism was matched to a typically developing child (TD) on the basis of nonverbal ability (11 males and 4 females, mean chronological age 6 years 10 months, ranging 5 years 1 month to 8 years 3 months). 

Q5. How many participants with autism scored within the range of the CARS?

To comply with their inclusion criteria, all participants scored within the autistic range, with CARS scores ranging between 33-50. 

Q6. What is the effect of the task on the accuracy of the target?

However when the task gets harder (6 items for typically developing participants) participants look at the target for less time when they respond incorrectly.