When asked to identify two targets embedded within a rapid consecutive stream of visual stimuli, observers are less able to identify the second target when it is presented within half a second of the first. This divided attention deficit has been termed the attentional blink (AB). Object files are temporary episodic representations thought to provide the
basis for object constancy in our perceptual world. Object file research has revealed benefits (faster reaction times) when the same object file can be used across successive visual displays. This contrasts with slower reaction times when there is a mismatch between successive visual displays, possibly because the process of generating new object files is attentionally demanding. In this thesis RSVP methodology was used to investigate the relationship between the AB and object files, and specifically the relationship between the degree of object file continuity inherent in the visual streams and
the AB magnitude. The full stream (24consecutive stimuli) experiments revealed an inverse linear relationship between the degree of object file continuity and AB magnitude. High levels of object file continuity were associated with almost complete
attenuation of the AB, whilst partial disruption of object file continuity led to less attenuation and complete disruption of continuity led to a robust AB effect. Furthermore, these outcomes were not due to differences in perceptual masking across the stream types. Several of the experiments described here were designed to identify which aspect of the full streams provided the locus of object file continuity and the associated AB modulation. No direct evidence was found to support the hypothesis that items preceding Tl serve this function (Raymond and Sorensen, 1994), although such a position cannot be
ruled out. Evidence presented here suggests that an important locus of object file continuity in the full stream experiments was due to items intervening Tl and T2; the underlying process at this locus is consistent with Kahneman et al. 's (1992) notion of 'recency' in object file reviewing. This thesis also provides evidence that, when present, distractor items (other than the masks that accompany Tl and T2) are implicated in determining object file continuity and AB modulation.