Avoiding the dichotomy between endogenous and exogenous attention: Selective attention as an attentional priority map fed by multiple interacting attentional signals.

For more than 40 years, selective attention has been summarized as operating through two distinct mechanisms: a bottom-up mechanism (exogenous attention, driving attentional selection based on stimulus features such as salience) and a top-down mechanism (endogenous attention, driving attentional selection based on goal-related information). This dichotomy is still an integral part of discourse on selective attention, despite contemporary research clearly showing that this is an oversimplification. In the present theoretical synthesis, we summarize nine key lines of arguments showing that exogenous and endogenous attention are not cleanly distinguishable in terms of either neurocognitive or computational mechanisms, and that they do not exhaustively represent the types of attentional signals that contribute to attentional selection. We then outline how the framework of priority maps can help better describe the combined influence of salience-based, goal-based, and other stimulus- and observer-related sources of attentional bias. In this view, selective attention is more usefully summarized in the form of various concurrent attentional signals being integrated in a priority map, which represents attentional weights at different space locations. This framework also helps conceptualize possible types of interactions between attentional signals, including early and late interactions during the course of attentional processing.