Poster #4 - Neural mechanisms underlying preferential looking in a visual recognition memory task in pre-verbal infants

• In Event: Poster Session 18

Abstract

Since the 1960s researchers have used behavioral looking patterns as a window to the mind of infants (e.g., Fantz, 1961). These behavioral looking measurements have not only revealed infants' perceptual and conceptual abilities such as spatial orientation (Atkinson, 1988), face discrimination (Young-Browne et al., 1977), and physical rules (Baillargeon et al., 1987), but they are also predictive of later cognitive abilities. (e.g., Rose et al., 2004). However, looking patterns are the culmination of many sensory, cognitive, and motor processes and it is unclear which of these processes are reflected in looking patterns. Here we applied event-related potentials (ERPs), a brain measure with high temporal resolution that allows perceptual, attentional, and memory processes to be isolated and tracked, to better understand the nature and origin of visual preferences.

Specifically, we assessed visual working memory for faces behaviorally using an established paired comparison visual recognition memory (VRM) paradigm with faces (Rose et al., 1999; Dzwilewsky et al., 2020). In a separate session, we also measured the brain response to an analogous passive viewing face memory task by recording event-related potentials (ERP) from the scalp. A total of 118 7.5-month-old infants contributed usable ERP data and 85 contributed useable data on both tasks. The VRM task consisted of 5 blocks with a familiarization trial and two test trials in each block. From it, we extracted three previously established metrics: average run duration (average time spent looking at the stimuli before looking away in seconds) and time to familiarization (time to reach familiarization criterion in seconds) from the familiarization trials and novelty preference (the proportion of time spent looking at the novel stimulus) from test trials. These are measures are thought to reflect information processing speed, visual attention, and visual recognition memory, respectively with shorter run duration, shorter time to familiarization, and more novelty preference reflecting better performance (Dzwilewsky et al., 2021), For the ERP task, we extracted amplitudes characterizing established components thought to reflect earlier perception (N290/P400),mid-latency attention (Nc), and later memory updating (PSW) (Guy, Zieber & Richards, 2016; Reynolds, 2015).

At the group level, the behavioral task showed memory for the familiarized face, as indicated by a novelty preference in the paired comparison (t(84)=10.76, p<.001). We also found ERP group differences between novel and familiar face processing on later components related to attentional orienting (Nc, t(117)=2.77, p=.006) and memory updating (PSW, t(117)=3.10, p=.002), but not on earlier components related to structural encoding and perceptual processing of faces (N290/P400, all ps>.05). Preliminary correlational analyses were then run between the three behavioral metrics and each of the 4 ERP components of interest and revealed only that Nc component was significantly correlated with time to familiarization (TTF) metric (r(84)=-.24, p=.028). Time to familiarization has been characterized as a marker of attention during the learning or encoding phase of the behavioral task. While analysis is ongoing, these results suggest that differences in behavioral novelty preferences may reflect, at the least, differences in attentional orienting during the learning phase.

Authors

• Yiyu Liu, University of Illinois at Urbana-Champaign
  Presenting Author

• Elizabeth A Enright, University of Illinois at Urbana-Champaign

• Daniel C. Hyde, University of Illinois at Urbana-Champaign