Ananya Mishra¹, Tony Thomas¹; ¹Indian Institute of Technology Roorkee

INTRODUCTION Contextual cueing involves implicit learning of the inter-spatial relationships among search items, enhancing search efficiency. The effect is considered pre-attentive and argued to be rapid, automatic and effortless. The present study investigates the robustness of the contextual cueing effect under conditions of featural level disruptions. METHOD 30 participants completed a letter search task that had 10 search configurations repeated (unbeknownst to participants), and each configuration was presented 30 times in random order (Training Phase). Each letter was in white font and size (0.57° × 0.47°). The Testing phase had the configurations used in the Training Phase presented again. However, items appeared in a different colour (red and green) or size (20pt and 50pt; 0.38° × 0.28° and 0.95° × 0.66°, respectively). The featural level changes were not expected to affect retrieval of the learnt invariant spatial relationship of the search items, provided the implicit learning was robust. RESULTS Training phase demonstrated robust learning F(5,145)=21.90, P<.001, ηp2=.43. In testing phase, a significant RT cost (vis-à-vis repeated (Old) configuration) was obtained: when items appeared randomly in green and red (Rand_Color: t(29)= -2.65, p=.013), in two different sizes (Rand_Size: t(29)= -6.98, p<.001), perceptually grouped in size (Percept_Size: t(29)= -6.38, p<.001); suggesting disruptive effects of featural changes on Contextual Cueing effect. CONCLUSION Extraction of regularities in contextual cueing effect seems to be dependent on bottom-up factors, and not a generic phenomenon as suggested previously. Featural-level violations (for instance, colour and size in the present case) stand out perceptually and seem to override the advantage offered by the implicitly learnt patterned spatial information. Features get perceptual prioritization earlier in time than the advantages offered by the contextual cueing effect.