Investigative Ophthalmology & Visual Science, 41:425. 2000. [ARVO 2000; Ft. Lauderdale, FL.]
Purpose. To reconcile estimates of visual memory capacity derived from flicker experiments (5-6 items) with estimates derived from eye-movement, tracking, and partial-report studies (3-4 items).
Methods. Visual search was carried out using arrays of horizontal and vertical rectangles under flicker conditions, where displays continually alternated between "on" (each exposure 800 ms) and "off" (blank fields of various durations). Targets had a changing orientation and the distractors a constant orientation. Changing items were present in a random half of trials, and observers were asked to report for each trial whether a changing item was present or absent. Reaction times were measured for set sizes 2,6, and 10 items. An estimate of memory capacity was then calculated from the resultant search slopes. This was done for a set of interstimulus intervals (off-times) ranging from 120 ms to 960 ms.
Results. For off-times of 120 ms, the capacity estimate was 5.7 items, a value close to those of earlier flicker studies. For off-times of 360 ms, this fell to 3.5 items. The capacity estimate remained approximately constant at 3.5 items for off-times between 360 ms and 960 ms.
Conclusions. When off-times are 360 ms or longer, the flicker technique leads to estimates of memory capacity that are similar to estimates of attentional span derived from other techniques--a value of about 4 items. The capacity "boost" found at shorter off-times is unlikely to be due to guidance of search by simple first-order motion signals, since these generally have a time window of less than 100 ms. Rather, some other low-level mechanism appears to be involved.