CHAPTER 27 Decision and Attention Andrei Gorea and Dov Sagi 152 ABSTRACT In real life, individuals are faced with more than one perceptual event on which they have to make distinct decisions. It is shown that for a range of such multi- stimulus environments, decision behavior departs from optimality in the sense that subjects do not set their decision criteria in accordance with the require- ments of each individual event. This behavior is explained in terms of a unified internal representation of the multistimulus environment, presumably result- ing from the relaxation of attention to the critical dimension associated with each stimulus. Exceptions are observed for cross-modal (audiovisual) stimula- tions and for stimuli showing sensory interference. It is proposed that decision behavior and the selection process required to segment sensory objects are inti- mately related. Response criterion interaction may account for phenomena such as extinction and may be the substrate of a number of contextual effects. I. INTRODUCTION Despite the progressive and by now practically completed withdrawal of introspective thought from the field of experimental psychology, or perhaps because of it, a little introspection may guide the reader of this chapter. After all, every research report on attention that saves itself a formal definition of this concept—the overwhelming majority of attentional studies—relies on its intuitive understanding. Intu- ition tells us that making a decision is by essence an attentional state, that it requires pondering and hence involves an effortful, intentional, voluntary (terms willy-nilly equivalent to free will) component to be contrasted with a default, “relaxed,” freewheeling behavior. This being said and inasmuch as intention, volition, free will, and perhaps mental effort remain purely intuitive, ill-defined terms, attention itself, at least its endogenous aspect, stands out as an equally indefinite concept. The work to be described here bears on human deci- sional behavior and, more specifically, on a subject’s capacity to deal with a number of decision criteria when faced with an equal number of distinct events. Techni- cally speaking, this is an exploration into the psy- chophysics of decision. The basic results show that when confronted with a range of equally likely but different- strength events, subjects use a unique decision crite- rion although optimal behavior requires the use of criteria proportional to the stimuli strengths (see below). We refer to this behavior as to criterion attrac- tion and we interpret it as the consequence of a unitary internal representation (UIR) of the (physically) distinct events, at least along the dimension under study. Criterion attraction (and its underlying UIR) is not observed under all the experimental conditions con- sidered. A survey of the ensemble of our data led us to conclude that the UIR is the instantiation of a default, nonattentional state whereby a number of stimulus dimensions are merged together, hence sparing the observer the sustained effort of keeping them apart in exchange for a negligible loss in performance. II. HISTORICAL BACKGROUND By the mid-19th century, Fechner (1860) asked (and answered) the question of how to measure sensation; he thereby founded psychophysics. Some 50 years later, the Gestalt school put forward the study of shape per- ception relying on what could be called conceptual observation. Another 50 years later, a group of what one might call “neo-psychophysicists” set forth signal detection theory (SDT) (see Green and Swets, 1966), a set of principles and psychophysical tools by means of which sensation and decision were associated in an Neurobiology of Attention Copyright 2005, Elsevier, Inc. All rights reserved. INO027 11/11/04 12:52 PM Page 152