An exploration of the filtering mechanisms involved in infants' selection of events for deeper encoding and memory retention

Graepel-Csink, Viktoria Zsuzsanna (2021) An exploration of the filtering mechanisms involved in infants' selection of events for deeper encoding and memory retention. PhD thesis, Birkbeck, University of London.

Abstract

The work presented in this thesis was motivated by the question as to how relevant information is selected for focused attention, visual exploration and memory encoding in infancy. This question involves the initial paradox that the cognitive system can only discard information once it has already attended to it, thus wasting cognitive resources. Based on both the adult and the infant literature, my main hypothesis was that ‘surprise’, i.e. a violation-of-expectation might act as a mechanism that highlights a piece of information that is ‘worth learning’. In three experiments, I explored whether events that –presumably –contradict the infant’s expectations would result in better learning outcomes. In Experiments 1 and 2, 12-month-olds saw an object magically disappear (Experiment 1) or appear at a new location following an occlusion event (Experiment 2). These scenarios –and their possible counterparts –were followed by an opportunity to learn a new property about the object, with the prediction that violations-of-expectation would facilitate learning. In Experiment 3, 17-month-olds were presented with familiar objects that were either labelled correctly (dog = ‘dog’) or mislabelled (dog = ‘banana’/dog = ‘cumbles’), followed by a subsequent recognition memory test. In brief, infants in these experiments did not show enhanced learning and memory for the items that had previously surprised them, despite unambiguous evidence for a surprise response in Experiment 3. I used eye-tracking to explore infants’ gaze duration, fixation patterns and pupil dilation in response to the surprising outcomes, as well as their looking times to the entire screen and their social referencing behaviour. The interpretation of infants’ looking behaviour and the potential inferences to be drawn regarding the underlying cognitive mechanisms are discussed throughout the thesis. Furthermore, motivated by the experimental evidence, I explored infants’ underlying representations that are violated in these experimental paradigms. I reasoned that the source of surprise –i.e. the strength of a representation and the degree of certainty with which it is held –is likely to determine the cognitive impact of surprise. This question echoes previous discussions in infant cognition on the correct inference to be drawn from infants’ responses to violations-of-expectation, which are questions that I discuss throughout the thesis. In order to explore whether infants rely on purely sensory memories to solve violation-of-expectation tasks, I tested 12-month-olds in a paradigm that involved the impossible appearance of an item at an empty location (Experiment 4). A surprise response in this scenario –measured via looking times –would indicate that infants represented the absence of an item at the location, which is difficult to accomplish using sensory memories alone. Similarly to previous studies with younger infants, 12-month-olds did not respond to this violation, which is compatible with the view that these early representations are derived from sensory memories. Furthermore, motivated by the findings of Experiment 4, I tested adults in a novel paradigm that involved the concurrent encoding of multiple items and empty locations and responding to changes that occur at one of the locations (Experiment 5). I analysed fixation patterns during encoding, pupil dilation during the delay phase, as well as accuracy and reaction times at the response phase. In summary, all of these measures revealed the same finding, namely that –unlike items –empty locations were exclusively encoded as a property of the global pattern as not as unique pieces of information. In summary, the results from Experiments 1, 2 and 3 argue against the hypothesis that violations of firm prior knowledge facilitate infants’ learning. Additionally, the results from Experiments 4 and 5 suggest that forming a representation in the absence of clear visual input is remarkably challenging for the encoding and retrieval processes, even in adulthood. These findings are interpreted in light of earlier work in the adult and infant literature on the link between perceptual novelty and surprise, with a special emphasis on their likely cognitive impact.

Metadata

Statistics

DownloadsShow export options

Export as XMLJSONCSV

Activity Overview

6 month trend

149Downloads

6 month trend

163Hits

Additional statistics are available via IRStats2.

Archive Staff Only (login required)

Edit/View Item