Abstract
When subjects search for two targets presented among a rapid sequence of visual distractors, they are severely impaired at detecting the second target (T2) if it occurs within 500 ms of the first one (T1). Models of this ‘attentional blink’ (AB) suggest that it reflects limitations in consolidating target representations into visual short-term memory (VSTM). However, the relation between the AB and capacity limits of VSTM has not been directly established. Here, we tested the specific prediction that if the AB is subject to the capacity limits of VSTM, then T2 performance should 1) decrease as the number of encoded T1 items increases, but 2) level off after the number of T1 items reaches VTSM capacity. We tested this prediction in 11 subjects where the T1 task consisted in encoding a variable number (1–6) of letters briefly displayed around an imaginary circle, and the T2 task consisted in detecting a large X shown inside the imaginary circle at variable SOAs after T1. To provide converging estimates of VSTM capacity, we employed both partial and whole report procedures for the T1 task. Both types of report demonstrated that subjects encoded more objects as T1 setsize increased, but only up to setsize 4, when their capacity leveled off even with further increase in the number of objects displayed. Reciprocally, T2 performance decreased as T1 setsize increased, but again only up to setsize 4, and leveled off thereafter. To rule out the possibility that these findings were due to verbal working memory interference between T1 and T2, an additional 16 subjects performed a similar experiment but with a purely perceptual T2 task that required detecting a gap in a square (Landolt figure). The results were identical to the first experiment.
These findings suggest that the AB is tightly coupled to the limited capacity of visual short-term memory.
Supported by NSF Grant BCS-0094992.