Although efficient visual search and effortless texture segmentation are assumed to rely on preattentive processing, they are considered to be different tasks. We believe that the debate receives new insights from studies on the detection of pop-out targets varying the number of context elements. Previous studies reported a non-monotonic relationship between the number of display elements and detection performance, indicating that processing small display sizes (as usually observed in visual search tasks) may be different to processing in large set sizes (as usually observed in texture segmentation).
Our paradigm allowed a transition from stimulus displays typically used in search tasks to stimulus displays typically used in segmentation paradigms. Stimulus arrays differed in set size, ranging from 2 to 49 elements. Detection performance and ERPs were recorded. We assumed that different processes may underlie the processing of the pop-out stimuli dependent on the number of elements in the stimulus display.
Our experiments showed the expected non-monotonic relationship between set size and detection performance: performance first decreased slightly and then increased again. ERP-effects were as follows: The N2p was increased for blank compared to target trials, however, only for large set sizes. An N2pc was observed for a broader set size range, but not for very large set sizes. P3 was enlarged for target compared to blank trials for all set sizes larger than 4. The data can be regarded as evidence in favor of different processing modes depending on the number of elements within a stimulus display.