Comparing Automatic Eye Tracking and Manual Gaze Coding Methods in Young Children with Autism Spectrum Disorder.
Courtney E Venker, Ron Pomper, Tristan Mahr and 3 others
PMID 31622050WHAT IT FOUND
Manual gaze coding kept more usable data than automatic eye tracking in young children with autism, but the two systems did not always lead to the same conclusions about word processing.
Key findings
01Manual gaze coding yielded more usable trials and more looking time than automatic eye tracking: 12.41 and 12.53 trials versus 10.49 and 10.75, and 90.05% and 91.22% versus 87.88% and 88.73%.
02When analysed separately, manual coding retained 65 children and showed a significant Target Present versus Target Absent difference (p = .012), while eye tracking retained 53 children and did not (p = .164).
03After separate cleaning, 17 children were excluded from the automatic eye-tracking dataset but included in manual coding, and they had higher autism severity (mean 9.18) than the 53 retained in eye tracking (mean 8.17; p = .006).
STILL TO COME
How it was doneWhat they foundWhat it means for SLPs
Read the rest of this summary
You get three full summaries a month, free, and we do not ask for a card. Search, the TL;DRs and your library stay unlimited either way.
What it does not show
The study compared only one automatic eye tracker and one manual gaze-coding procedure, so other systems may behave differently. It used one set of data-cleaning rules; changing child-level or trial-level criteria could change the results. The findings are most relevant to 2-large-AOI designs where both systems can measure the same outcome; they do not show how to choose systems for tasks requiring fine-grained fixations or pupil measures. The separate single-system analyses used different samples, 65 children for manual coding and 53 for eye tracking, so differences could reflect statistical power or child characteristics. The study did not directly measure gaze accuracy, so it cannot tell whether a missed or shifted look was technical loss or true looking away. Children excluded from eye tracking had higher autism severity, so eye-tracking results may not represent more severely affected children.
Declared interests
The article is marked as Research Support, N.I.H., Extramural. The supplied text does not include an author conflict-of-interest declaration or statements about sponsor role.
The easy way to misread this
Do not treat automatic eye tracking and manual gaze coding as interchangeable methods. In this study, manual coding retained more data and showed a significant Target Present versus Target Absent difference (p = .012), while eye tracking did not (p = .164), and eye tracking excluded more children with higher autism severity.