Abstract Visual images in the vicinity of visual targets interfere with target discrimination ( d′ ) at high levels of performance but facilitate discrimination at low levels. Here, I review the background literature that supports this conclusion and theorize that an image showing features matching those of the target boost sensitivity at every level of performance, but images of all types (including matching ones) also add multiplicative noise. This noise increases in proportion to the signal level, eventually overcoming the boosting effect and causing interference (i.e., the Perky effect). Experiment 1 demonstrates that image type (vivid or weak, static or dynamic) has no effect on the changeover from facilitation to interference; the critical factor is signal level. Equations for both the mean and standard deviation of d′ are derived and shown to fit data from Experiment 1, two previous large-scale studies from my lab, and a recent study from another lab.
Adam Reeves (Fri,) studied this question.