Visual Attention and Early Cognitive Development in Rural Malawian Infants

Visual attention is one of the earliest developing neurocognitive and behavioral systems in infants. Thus, its assessment has been viewed as a possible method to examine infants’ perceptual and cognitive abilities before fine motor and language skills have started to develop and before conventional tests of cognitive function become feasible. However, infant visual attention and its predictive value with respect to long-term cognitive development has been less studied in low- and middle-income countries where children’s developmental status and predictions have been based on gestational age and nutritional status rather than direct measures of cognitive function.

This thesis presents one of the first studies applying state-of-the-art eye tracking measures to assess infant attention as a marker of early cognitive development in low-resource settings. The thesis was designed to assess whether early-life risk factors (i.e., gestational age, growth, maternal cognition and pyschosocial well-being, socioeconomic status, and care practices) and visual attention are related at 9 months and predict conventional developmental assessments at 18 months. In addition, the thesis assessed whether a dissociation of attention to faces and patterns, previously shown in studies in Western countries, is a characteristic of infant attention that is consistently found in data from different cultures.

For a prospective cohort study in rural Malawi, a total of 444 newborn children were enrolled based on their gestational aged at birth (i.e., preterm, early-term, and full-term) and were followed up until the age of 18 months. Besides measures of early-life risk factors, data on infant visual attention was collected at 7 and 9 months with remote, infrared eye tracking. Traditional developmental assessments were administered at 18 months. Eye tracking measures of visual attention consisted of measures of visual search, anticipatory attention shifts, and attention to faces. Development at 18 months was assessed using tests of motor skills, language acquisition, socioemotional behavior, and executive functioning. Dissociation of attention to faces and patterns was assessed by analyzing attention disengagement times in a pooled data set including a total of 637 infants aged 6–9 months from three studies conducted in Finland, Malawi, and South Africa.

Gestation age groups did not differ on their visual attention scores. Correlation coefficients between early-life risk factors and visual attention scores were low (-0.17– 0.14). Also, correlation coefficients between risk factors or visual attention and later development were low (-0.10–0.19). On cross-cultural analysis, delayed disengagement occurred on 52% of face trials compared to 4% of pattern trials, demonstrating a clear attentional bias for faces. The analyses further showed that the likelihood of delayed disengagement from faces was statistically dissociable from the likelihood of delayed disengagement from patterns.

In conclusion, preterm birth, nutritional status, or other early-life risk factors were not associated with the 9-month visual attention in rural Malawi. Also, early visual attention or early-life risk factors were not associated with 18-month developmental attainment. These results do not support the use of current measure of visual attention as a predictor of development in rural Malawian infants till 18 months of age, but leaves open the possibility that associations are detected at older ages. Analyses on disengagement times provided evidence for cross-cultural similarities in infants’ attentional bias for faces, and suggested variations in this bias are independent of variations in attention to non-face patterns. The results also showed that the distinction between faces vs. patterns was robust compared to other factors that affect infant attention (i.e., facial expression, face identity, and habituation).