In the Blink of an Eye: Visual Processing, Hippocampal Subfields, and Memory Function in Children

• In Event: 1-163 - Memory and the Developing Brain: Sampling Behavior and Defining Constructs

Integrative Statement

Studies of long-term memory performance in children suggest that recollection undergoes a protracted period of improvement throughout childhood and early adolescence. This developmental pattern has been considered in the context of maturing hippocampal subfields and the computational processes (i.e., pattern separation and pattern completion) that they support (e.g., Keresztes et al, 2018). However, performance differences between children and adults are only illuminating insofar as they reflect differences in the construct of interest. We note that the encoding phase of standard memory tasks like the traditional picture recognition paradigm may disadvantage younger participants at the earliest stage of visual exposure and processing, introducing variability in the measures of long-term memory performance that are not related to long-term memory mechanisms per se. Children’s diminished capacity to exert sustained attention, reduced visual exploration of scenes, inexperience with encoding strategies, and inchoate distributed mental schemas may introduce age-related confounds that muddy interpretations of children’s performance on such tasks. To minimize or assess the impact of these confounds, we created a novel task designed to differentiate between the contributions of early perceptual processing and classically defined mnemonic functions toward picture recognition performance. Item viewing time at encoding was limited to 200 ms, or about the duration of a single fixation, to better equate visual input and exploration across ages. The task included an immediate and delayed test, allowing assessment, on a subject-by-subject basis, of the relative retention of information in short- and longer-term memory. We included decoy items at both testing times to assess the perceptual precision of stimulus representations at each timepoint. One condition included an 800 ms inter-stimulus interval, whereas a second, more challenging condition omitted the inter-stimulus-interval to prevent the use of encoding strategies like rehearsal. We tested 103 individuals age 7-28 on the behavioral task and obtained hippocampal volume measures using structural MRI from 57 of the child and adolescent participants. Recognition performance (assessed by target hit rates corrected for guessing) at both immediate and delayed test correlated positively with age (r=0.45 and 0.39; ps<0.001). Differences in hit rate between immediate and delayed test (“Loss”) were equivalent for children, adolescents, and adults with no correlation between the Loss measure and age (r=-0.02). A mediation analysis (95% CI) found that 36% of the effect of age on hit rate at delayed test was explained by hit rate at immediate test. We found no significant relationship between hippocampal subfield volumes and Loss or perceptual precision (PP: target hits vs decoy false alarms) at delayed test, however we did find that CA1 volume was significantly correlated with the Perceptual Precision measure at immediate test (r=0.33, p=0.01). Our findings indicate that children’s poor performance on certain visual long-term memory tasks may not reflect memory function differences per se, but rather differences in viewing behavior and early visual processing – functions that have themselves recently been linked to the hippocampus (Voss et al, 2017).

Authors

• Rebecca F Schwarzlose, Washington University in Saint Louis
  Presenting Author

• Qijing Yu, Wayne State University
  Non-Presenting Author

• Sruthi Ramesh, Wayne State University
  Non-Presenting Author

• Roya Homayouni, Wayne State University
  Non-Presenting Author

• Zhijian Chen, Wayne State University
  Non-Presenting Author

• Noa Ofen, Wayne State University
  Non-Presenting Author