The effect of local-global processing on contextual learning

Abstract

Repeated contexts allow us to find relevant information more easily. The ability to use such information has been called contextual learning and has been investigated using the contextual cuing paradigm. In this task, participants search for and identify the orientation of a target letter T amongst rotated Ls. Over blocks, RTs become faster for repeated spatial configurations (old displays) compared with novel configurations (new displays). Learning such repeated contexts has been proposed to depend on either global processing of the repeated contexts, or alternatively processing of the local region surrounding the target. The present study takes a novel approach by measuring differences in participant’s attentional bias (i.e. local or global) in relation to the amount of contextual learning. Twenty participants completed a shape detection task followed by a contextual cueing task. In the shape detection task participants were presented with large shapes consisting of smaller shapes and they had to decide whether the target shape was present or absent. The presentation of the target shape could either be presented at the local level (i.e. the smaller shape) or the global level (i.e. the larger shape). In the contextual cuing task participants completed 16 blocks with 24 trials in each (12 old and 12 new displays). Participants were then split into two groups, depending on whether their local RT - global RT was below (local-bias group) or above (global-bias group) the median. Participants in the local-bias group showed significantly stronger contextual cueing effects than participants in the global-bias group (152 vs. 39 ms, respectively). Furthermore, there was a strong negative correlation between global bias and contextual cueing (r = -0.61, p<.005). In conclusion, this study suggests that contextual learning depends more on local information than on global information. Furthermore, it highlights the importance of observer variables in relation to contextual learning.