How visual and semantic information influence learning in familiar contexts.

Abstract

Previous research using the contextual cuing paradigm has revealed both quantitative and qualitative differences in learning depending on whether repeated contexts are defined by letter arrays or real-world scenes. To clarify the relative contributions of visual features and semantic information likely to account for such differences, the typical contextual cuing procedure was adapted to use meaningless but nevertheless visually complex images. The data in reaction time and in eye movements show that, like scenes, such repeated contexts can trigger large, stable, and explicit cuing effects, and that those effects result from facilitated attentional guidance. Like simpler stimulus arrays, however, those effects were impaired by a sudden change of a repeating image's color scheme at the end of the learning phase (Experiment 1), or when the repeated images were presented in a different and unique color scheme across each presentation (Experiment 2). In both cases, search was driven by explicit memory. Collectively, these results suggest that semantic information is not required for conscious awareness of context-target covariation, but it plays a primary role in overcoming variability in specific features within familiar displays.

A Configural Dominant Account of Contextual Cueing: Configural Cues are Stronger than Colour Cues

“…For example, Henderson (2006a, 2006b) found that a repeated background using photographs of natural scenes facilitated both RTs (2006a) and eye movements (2006b) to a single target letter. Furthermore, a contextual cueing benefit was found if the repeated context was defined by complex 'fractals' and geometrical images that made up the background of the search display (Goujon et al, 2012, Kunar et al, 2006, the colour of the search stimuli (Kunar et al, 2006;Jiang & Song, 2005;Jiang & Chun, 2001) or, important to this paper, the colour of the search background (Kunar et al, 2006). Understanding how these different contextual cues interact is important to obtain a unifying theory of how predictive contexts work together to benefit search overall.…”

“…In contrast, Goujon et al (2012) used geometrical figures as the repeated context. In their experiments, instead of using the configuration of distracter items, Goujon et al (2012) used the pattern of different complex fractals (presented as display backgrounds) as the predictive cues.…”

“…Furthermore, when the colours of the image were varied across the experiment (similar to the methodology of Ehinger & Brockmole, 2008) the contextual cueing effect was reduced compared to a condition in which the colours of the repeated geometrical images were invariant. Goujon et al (2012) suggested that with non-semantic displays, composed of visually complex images, colour played a more important role. From these previous studies it seems that the importance of colour as a cue depends on what other contexts were present.…”

“…Note that in the work of Goujon et al (2012) the target was the only stimulus present in the display. In their experiments, as the predictive context was established by the repeated geometrical images, there was no need for a predictive configural context outlined by the distracter layout.…”

“…In their experiments, instead of using the configuration of distracter items, Goujon et al (2012) used the pattern of different complex fractals (presented as display backgrounds) as the predictive cues. Here, participants learned to associate a specific fractal image with a specific target location.…”