Subliminal gamma flicker draws attention even in the absence of transition-flash cues.

We recently reported evidence indicating that selective attention is deployed to a target location in a multi-object display, when the target event (a change of one of the objects) is preceded by subliminal flicker in the gamma range. However, concerns have been raised regarding the stimuli used in this study and the possible contribution of an artifactual cue: a "transition flash" between pretarget flicker offset and target onset. Here, we report a series of experiments investigating the existence and potential contribution to selective attention of this transition-flash cue under different presentation conditions. We find that, although the transition flash is a real phenomenon (detection rates ≃ 15% > chance), it cannot, on its own, explain the original effects of gamma flicker on the response time to target detection. Even after eliminating this flash, detection was significantly faster, or more accurate, for targets preceded (vs. not preceded) by flicker. This congruency effect (≈ 15 ms) demonstrates that gamma flicker on its own is sufficient to engage selective attention. This interpretation is further strengthened by a reevaluation of 1) experiment 7 reported by van Diepen and colleagues and 2) the validity effect experiment reported by Bauer and colleagues. Possible reasons for the discrepant results are also discussed.