Poster #3 - Surface indicators of deeper processes: Pupil dilation reflects covert attentional shifts

• In Event: Poster Session 01

Abstract

Infant visual attention develops rapidly, influencing visual behavior from the earliest moments of life. Cued attention tasks have long been used to assess visual attention (Hood & Atkinson, 1993; Posner & Petersen, 1990), and recent work using this approach demonstrates substantial increases in covert attention between 5 and 11 months (Ross-Sheehy, et al., 2015). In this task, infants were first oriented to a brief central fixation stimulus. Once fixated, the stimulus was removed, and infants were presented with a 100ms spatial cue (small black dot to the left or right of fixation) followed by a colorful target. Measures included latency to orient to the target (reaction time, or RT) and accuracy. Critically, the location of the spatial pre-cue influenced orienting, but only if attention was sufficiently developed. Specifically, a cue presented in the same location as the target decreased the RT, and a cue presented opposite the target increased RT. Thus, the appearance of these spatial cueing effects suggests covert attentional orienting.

Changes in pupil dilation (PD) may also relate to covert attention shifts. Recent adult research suggests that covert attention shifts are frequently preceded by brief PD, particularly if the task is difficult, or stimuli appear in the periphery (Brocher et al., 2018; Vilotijević & Mathôt, 2022). Thus, we hypothesized that if spatial cueing effects reflected covert attention orienting, infants might demonstrate PD just prior to their eye movement to the target. Specifically, if PD reflects covert attentional orienting, then we should see a negative correlation between PD and subsequent RT. To test these hypotheses, infants were assessed using the IOWA Task (Ross-Sheehy et al., 2015) and eye movements and PD were assessed via eye tracking. Participants included 66 5-month-olds, 71 7-month-olds, and 113 11-month-olds who are part of a larger ongoing study.

RT by condition was assessed using a 3x5 repeated measures ANOVA, and results revealed a main effect of both condition, F(1,246)=146.95, p<.001, and age, F(1,246)=25.878, p<.001 (Figure 1A). Follow-up simple effects tests revealed significantly faster RT to the valid cue (all ps<.001), and significantly slower RT to the invalid cue (all ps <.001, Figure 1A). We next conducted Pearson’s r correlations comparing baseline corrected pupil during the target interval (prior to eye movement) and RT (Figure 1B). Results revealed a significant negative correlation at 5-months for the invalid cue, r(432)=-.178, p<.001, and positive correlation for the tone cue baseline, r(566)=.173, p<.001. In addition, 11-month-olds demonstrated significant negative correlations for nearly every cue condition, including the double cue, r(961)=-.204, p<.001, the no cue, r(1121)=-.071, p=.018, the tone cue, r(1120)=-.102, p=.001, and the valid cue, r(891)=-.072, p=.032. Though these coefficients appear low, they are highly significantly, particularly for the double cue. In addition, these effects are highly consistent, with larger pupils predicting faster RTs, a finding that is consistent with adults. These correlations are preliminary only, full analyses will include linear mixed effect models to better capture the relation between moment-to-moment changes in pupil, and subsequent attention orienting

Authors

• Victoria Jones, University of Tennessee - Knoxville
  Presenting Author

• Esther Reynolds, University of Tennessee - Knoxville

• Micalee Anne Segers, University of Tennessee - Knoxville

• Shannon Ross-Sheehy, University of Tennessee - Knoxville