Using dynamic neural fields to understand the development of metric representations in typically developing and at-risk infant populations

During the past half-century, the experimental use of looking measures have led to many new discoveries about the origins of cognition. Across the first year, infants' looking changes in predictable ways, they form memories more quickly, and they begin to discriminate between subtly different stimuli. However, a rich understanding of the link between looking and cognitive dynamics has yet to be achieved. This was the overarching goal of this thesis.

I developed a new Dynamic Field Theory of infant looking and memory and formally implemented this theory in a Dynamic Neural Field model. In Experiment 1, I tested and confirmed a prediction of the model with 10-month-old infants. The prediction was that robust memory can induce both a familiarity and novelty bias depending on the metric similarity of the familiar and novel items at which infants look. This prediction is a radical one because all existing theories posit that familiarity biases arise from weak memory.

One central innovation of the DNF model is that it captures developmental change in the rate at which memories are formed and discrimination within the same system and from the same developmental mechanism. With a validated theory in hand, in Experiment 2 I tested this theoretical assumption. In particular, I measured the looking dynamics and discrimination performance of 5-, 7-, and 10-month-old infants. The results showed that infants' exhibited an increased ability to discriminate between dissimilar familiar and novel items between 5 and 7 months of age. The results also showed that three well-known looking indices of memory formation also generally change between 5 and 7 months of age. Additionally, individual differences in these looking indices were predictive of infants' discrimination performance. These findings indicate that, indeed, looking and discrimination change together, and are linked within individuals, over development.

In Experiment 3 I tested developmental change in the discrimination abilities of at-risk infants. Previous studies have shown that the looking dynamics and recognition performance of at-risk infants is delayed but, critically, follows the same developmental trajectory as typically developing infants. Consistent with these previous studies, the looking dynamics of at-risk infants did change in predictable ways over development. However, their discrimination performance did not - young at-risk infants, unlike young typically developing infants or older at-risk infants, discriminated between dissimilar familiar and novel items.