Related Experiment Video
Updated: Oct 4, 2025
07:09
Gaze in Action: Head-mounted Eye Tracking of Children's Dynamic Visual Attention During Naturalistic Behavior
Published on: November 14, 2018
10.9K
Remote Data Collection During a Pandemic: A New Approach for Assessing and Coding Multisensory Attention Skills in
Bret Eschman1, James Torrence Todd1,2, Amin Sarafraz3
1Department of Psychology, Florida International University, Miami, FL, United States.
Frontiers in Psychology
|February 7, 2022
Summary
Developmental researchers can now collect infant looking time data remotely using novel audiovisual protocols and machine learning for gaze analysis. This enables continued research during disruptions like the COVID-19 pandemic.
Area of Science:
- Developmental Psychology
- Cognitive Science
- Human-Computer Interaction
Background:
- The COVID-19 pandemic severely disrupted in-person data collection for developmental research, particularly for looking time studies.
- Existing remote platforms are often inadequate for precise infant looking time data collection and processing.
- There was a critical need for adaptable methodologies to maintain research continuity.
Purpose of the Study:
- To introduce novel remote data collection and analysis methods for audiovisual looking time protocols in young children.
- To adapt the Multisensory Attention Assessment Protocol (MAAP) for remote administration.
- To develop and validate automated gaze estimation and classification for remote webcam data.
Main Methods:
- Remote administration of the MAAP using Zoom and Gorilla Experiment Builder on participants' home computers.
- Gaze estimation from webcam recordings using the OpenFace toolkit.
- Development of a machine learning (artificial neural network) algorithm to classify gaze direction (left, right, center) on the participant's screen.
Main Results:
- Successful remote administration of the MAAP, enabling data collection in participants' homes.
- Demonstrated effectiveness and precision of OpenFace for estimating gaze direction and duration.
- High reliability of the machine learning approach for classifying gaze relative to Areas of Interest (AOI), comparable to traditional coding methods.
Conclusions:
- The combination of remote MAAP administration, OpenFace, and a machine learning algorithm provides a robust, automated method for analyzing looking time data.
- This approach facilitates continued developmental research, even during public health crises or when in-lab testing is not feasible.
- Best practices for remote developmental data collection are outlined to guide future research.
