|
Visual Search for Flicker: High Temporal Frequency Targets Capture Attention
University of California, Santa Cruz, CA
Purpose. Previously, we found that visual search for flicker was dependent on target-distractor discriminability (Dougherty, Verardo & Mayer, ARVO '95). The present study attempted to replicate the previous results with a cueing paradigm to eliminate possible stimulus density confounds.
Methods. Visual search was performed using a two-interval, forced choice paradigm. The stimuli consisted of 8 low-spatial frequency gabors arranged concentricly at 3-5° eccentricity (randomly jittered to break up the circle) and flickering at 2, 4, 8 or 17 Hz. The target, which was always present in one of the two intervals, flickered at a different rate than the distractors. The contrast of each flicker rate was set at twice threshold. Temporal phase of the distractors was randomized and each stimulus presentation ended with a brief, dynamic noise mask. There were 2, 4 or 8 spatial cues presented before each interval and the target always appeared in one of the cued locations. Stimulus duration was adjusted with a staircase procedure and the data for each condition were fit with a Weibull function by a maximum-likelihood minimization procedure to determine the critical duration necessary for 80% correct.
Results. For low frequency targets, search times were consistent with those predicted by flicker discriminability. Generally, less discriminable targets produced longer critical durations. For more discriminable targets, search times were relatively independent of the number of cues, while for less discriminable targets search times increased with the number of cues. High frequency targets (e.g. 17Hz) produced shorter critical durations regardless of discriminability. This was true even for target-distractor pairs of equal discriminability (e.g. 17Hz target with 4Hz distractors had shorter critical durations than 4Hz target with 17Hz distractors).
Conclusions. In general, more difficult discriminations require more processing time. However, the results showed systematic departures from the predicted search-time/discriminability functions for high frequency targets. The data can be accounted for if we assume that the high frequency targets captured attention.
(see the full poster)