Revisiting the effects of configuration, predictability, and relevance on visual detection during interocular suppression

Statistical regularities and predictions can influence the earliest stages of visual processing. Studies examining their effects on detection, however, have yielded inconsistent results. In continuous flash suppression (CFS), where a static image projected to one eye is suppressed by a dynamic image presented to the other, the predictability of the signal may facilitate or delay detection. To identify the factors that differentiate these outcomes and dissociate the effects of expectation from those of behavioral relevance, we conducted three CFS experiments that addressed confounds related to the use of reaction time measures and complex images. In experiment 1, orientation recognition performance and visibility rates increased when a suppressed line segment completed a partial shape surrounding the CFS patch, demonstrating that valid configuration cues facilitate detection. In Experiment 2, however, predictive cues marginally affected visibility rates and did not modulate localization performance, challenging existing findings. In experiment 3, a relevance manipulation was introduced; participants pressed a key when they detected lines of a particular orientation and ignored the other possible orientation. Visibility and localization were enhanced for relevant orientations. Predictive cues modulated visibility rates, orientation recognition sensitivity, and response latencies, but not localization—an objective measure sensitive to partial breakthrough. Thus, while a consistent surround can strongly enhance detection during passive observation, predictive cueing primarily affects post-detection factors—such as response readiness and recognition confidence—by introducing invalid cue interference. Relevance and predictability did not interact, suggesting that the contributions of these two processes to detection are mostly orthogonal.