Pre-attentive visual selection

In this note, we consider the critical function of the selection from an entire visual scene of visual locations or objects for detailed or attentive processing. Such selection of some places at the expense of others is necessary because attention has only a meagre capacity (estimated at just 40 bits/second; Sziklai 1956). The same bottleneck implies that the selection process itself cannot generally be attentive (bar an explicit cue or some other form of effective guidance). Fortunately, pre-attentive mechanisms, operating in parallel across the entire visual input, seen most prominently in pop-out (with, for instance, a red dot popping out among the green ones, or a vertical bar among horizontal bars), operate to offer a vastly simpler substrate of salience for attentional selection (Neisser 1967; Triesman & Gelade, 1980, Julesz 1981). Concomitantly, tasks that require the selection of non-cued locations that are not pre-attentively salient, are much more daunting.