Cognitive Representation of Facial Motion: Investigation of Normal and Impaired Cross-Feature Integration

In my D.Phil, I would like to test if our ability to integrate information across facial features influences the stability of our identity representations. Though this has not been explored previously, I hypothesize that the degree to which we are able to integrate dynamic facial information dictates the speed and accuracy of our identity-matching ability. Explicitly, I imagine that representations of identities learnt through successful integration of cross-feature movement are more stable than those learnt without such integration.
Under this assumption, healthy participants should be able to form more stable representations of identities for which they integrate information across features than of identities for which they do not. In turn, I suggest that ASD participants would not show a stability benefit in either case - given that they do not experience an illusory slowing effect, I hypothesize their ability to integrate motion across features to be impaired.
Of course, to ensure that integrating across features is causally involved with representation building and not merely correlated, we ought to conduct a pilot study. I suggest that varying the degree to which participants can derive a benefit of cross-feature integration should predict the stability of a resulting representation. Such initial finding would strengthen the argument that illusory slowing serves as a proxy of our ability to integrate motion information across features.
Once this proxy is established, an initial setup ought to be a behavioural study exploring differences in RG stability for paired-identity visual search performance of healthy participants. Subsequently, we should compare stability of NTs' and ASDs' representations.