Sensitivity of eye-tracking measures to child nutritional status and home environment in rural Malawi

• In Event: 2-043 - Objective measures of early development in low- and middle-income countries: EEG, Eye-tracking, fNIRS

Integrative Statement

Undernutrition and low nurturing and stimulation in the home contribute to an estimated 250 million children in low- and middle-income countries (LMICs) not fulfilling their developmental potential. Evaluating the efficacy of interventions targeting pregnancy and infancy requires objective and sensitive measures of infant cognition, during a period when infants have a limited behavioral repertoire. Measures of infant looking behavior may be more sensitive than global behavioral measures to factors that constrain early child development in LMIC contexts. We hypothesized that measures derived from two automated eye-tracking tasks administered to children age 5.5 to 10 months in rural Malawi would be associated with measures of nutritional status and the home environment.
Participants were enrolled in a randomized child nutrition trial conducted in Mangochi district. Children were eligible for enrollment if they resided in the catchment area of Lungwena Health Clinic or Malindi Hospital, were between 5.5 and 9.9 months of age, singleton birth, did not have acute illness or malnutrition, or history of egg allergy. At enrollment, we assessed 260 children (mean (SD) age 7.1 (1.2) mo) on a visual paired comparison (VPC) task, based on Rose (1994), and an Infant Orienting with Attention (IOWA) task based on Ross-Sheehy, Schneegans, & Spencer (2015). In the VPC task, children were presented with 4 faces for 20 s of familiarization each, then tested for novelty preference by presenting the familiar face paired with a novel face in a 20-s test trial. In the IOWA task, a central fixation was followed by a 100 ms cue (a small black dot) then a target image on the left or right side of the central fixation. On valid trials, the target and cue appeared in the same location, on invalid trials the target and cue appeared in opposite locations, on double cue trials the cue appeared in both locations, and on no cue trials no cue appeared.
Preliminary analyses revealed that, as expected, infants’ novelty preference was related to looking duration during familiarization, such that novelty preference was stronger with longer duration of familiarization time (Figure 1). Accuracy on the IOWA task differed between stunted (length-for-age z-score <-2) and non-stunted children, with stunted children showing lower accuracy on invalid trials (Figure 2). Ongoing analyses will evaluate infants’ responses to these tasks as a function of other child and environmental characteristics typically associated with global developmental assessments in LMICs. Specifically, we will evaluate infants’ visual cognitive performance and maternal education, Home Observation for the Measurement of the Environment Inventory score, child length-for-age z-score (LAZ), weight-for-length z-score, head circumference z-score, and hemoglobin concentration, all measured at enrollment. The results of these analyses will constitute some of the first evidence for the determinants of individual differences on these infant looking tasks measured by eye-tracking in an LMIC context.

Authors

• Elizabeth Prado, University of California, Davis
  Presenting Author

• Michaela Carol DeBolt, University of California, Davis
  Non-Presenting Author

• Bess L Caswell, University of California, Davis
  Non-Presenting Author

• Matthews George, University of Malawi College of Medicine
  Non-Presenting Author

• Ken Maleta, University of Malawi College of Medicine
  Non-Presenting Author

• Christine P Stewart, University of California Davis
  Non-Presenting Author

• Shannon Ross-Sheehy, University of Tennesee, Knoxville
  Non-Presenting Author

• Lisa Oakes, University of California, Davis
  Non-Presenting Author