Abstract
PURPOSE. To explore the interactions (interference or facilitation) between successive unfamiliar real-life scenes in RSVP tasks.
METHODS. Stimuli were 416 color images of real-life scenes unfamiliar to the subjects. During each of 32 trials per subject we used the RSVP procedure to present 7 pictures, 1 unrepeated and 3 repeated with 1, 2 or 3 intervening images (counterbalanced across subjects), with durations per picture ranging from 13 to 1710 ms. Following each 7-picture stream subjects were shown 8 pictures singly (4 seen and 4 new), and asked to judge whether the given picture was present in the stream they just saw. All 64 naïve subjects saw all the pictures, each picture never more than twice (only in the repeated condition). Accuracy data were corrected for guessing and fitted with psychometric functions.
RESULTS. All data sets are well fitted by exponential functions. For the unrepeated condition the time constant (corresponding to 63% correct) is ∼400 ms. In the repeated condition performance is virtually identical across all lags, with a time constant of ∼200 ms.
CONCLUSION. We find no evidence of negative interference between repeated unfamiliar pictures at any lag. Given the exponential form of the psychometric function, the fact that the time constant differs by a factor of 2 between the repeated/unrepeated conditions suggests complete independence between successive glances at a picture (probability summation). This implies that for sequences of pictures that are complex, not easily namable and unfamiliar to the subject, sequential effects, such as repetition blindness and attentional blink, do not apply.
NIH EY13155 to V. Maljkovic