Alertness and the control of attention

Abstract

The thesis aims to examine the ways in which alertness influences the relationship between attention and performance in a reaction situation. Alertness is operationally defined as the state induced following the presentation of a warning signal. The relationship of alertness to attention is considered, and research on alertness, both physiological and behavioural, is reviewed. Attempts to incorporate the observed effects of alertness into models of attention, generally based on traditional arousal theory, are examined, and a number of unresolved issues are identified. Empirical work carried out to investigate these issues shows that (a) expectancy does not interact with alertness at foreperiods of less than one second in duration; (b) alertness focusses attention towards more probable stimulus locations, in line with the predictions of arousal theory; (c) simultaneous improvements in both speed and accuracy of responding can be produced by alertness with stimuli of up to at least 800 milliseconds in duration, and this cannot be explained as a shift in speed-accuracy tradeoff; (d) explicit task instructions are more effective when subjects are alert than when they are not, and tentative evidence suggests that this may also be the case with the effects upon behaviour of implicit task context. These results support and extend the proposal by Kahneman (1973) that alertness produces a general facilitation of attention, the behavioural manifestations of which are task dependent. This conclusion contrasts with the more mechanistic role ascribed to alertness by Posner (1974), i.e., that of simply producing earlier sensory input sampling. A simple theory of alertness is presented which embodies these conclusions, and some further research topics of interest and relevance are identified and discussed