Introduction

``It all looks so negative. I just can't see anything good in this situation.'' Statements such as these are common refrains of people caught in the cognitive whirlwind of negative thoughts and feelings traditionally labeled depression. Yet, the role of characteristic attention to negative aspects of one's environment in the evolution and maintenance of depression is less clear than is its ubiquitous presence. It would certainly be comforting for an aspiring therapist to be able to recite the ills of paying attention to negative things and to thus propose that to heal themselves, his or her depressed clients need only attend more to positive stimuli. Two factors stand in the way of this instant and painless approach to alleviating one of society's most common and debilitating disorders.

First, the notion of attention biases in depression has been the subject of a great deal of debate over the last decade. Some researchers provide evidence suggesting that depressed people (e.g., Gotlib & McCann, 1984) and formerly depressed people (e.g., Ingram, Bernet, & McLaughlin, 1994) do selectively attend to negative stimuli when presented with both positive and negative information. Such findings can be interpreted to suggest that attention biases play a causal role in the maintenance and propagation of the disorder (Ingram, 1984). Other researchers suggest that attention biases in depression are epiphenomena or byproducts of a disorder which are not cognitively maintained (e.g., Teasdale & Barnard, 1993). Were this the case, a therapy directed at changing a negative attentional bias might not serve to alleviate the true cause of depression. Still other researchers suggest that depressed people do not differentially attend to negative, positive, and neutral information, and that the results observed by other researchers are a product of other cognitive processes such as negative memory or decision making biases which can be illustrated using methodologies designed to reveal these processes (Hill & Dutton, 1989; MacLeod & Mathews, 1991; MacLeod, Mathews, & Tata, 1986). The resolution of this debate has broad theoretical implications for a conceptualization of depression since the presence of negative attentional biases in attention are critical to the validation of many popular cognitive theories of depression (e.g., Beck, 1967a, 1974; Bower, 1981). Similarly, therapies advocating cognitive treatments for depression, involving, for example, the alleviation of negative attention biases (e.g., Beck, 1976; Brown & Lewinsohn, 1986; Ingram & Hollon, 1986), are based on such cognitive models. Before the aspiring therapist is to be confident in advocating such a therapy, evidence directly supporting its rational would be useful.

The second obstacle to our idealistic therapist's use of attention mediation techniques in therapy is more practical. Even if depressed people are biased to attend more to negative than positive information it is not clear how this statement would translate to the creation of a therapy, given the many forms such a bias could take. If indeed depressed people attend only to negative information, a treatment directed only to helping them pay attention to positive information, or to learning positive information might be useful. Alternately, if depressed people pay attention to information of all valences, but quickly filter out positive information, pay more attention to some negative stimuli than other negative stimuli, or have, in any way, a more subtle bias than a general predisposition towards attending to negative information, a clearer picture of the role of attention biases in depression than has been provided in the past will be useful in forming a relevant therapy. It will be suggested in the current thesis that attention is not necessarily biased towards all negative information in depression, but towards the negativity or affective content of any information. Thus, it will be suggested that changing depression must focus on pervasive devotion to negativity itself, rather than to specific effectively valenced aspects of the environment. Attaching a positive affective valence to otherwise negative information will be considered at least as important in therapy as redirecting one's attention. To better understand the rational for this hypothesis, an empirical and theoretical exploration of possible relationships between attention to depression will be presented in the thesis.

The current investigation begins by quantitatively aggregating results from existing literature devoted to the examination of attention biases in depression. Specifically, a meta-analytic review of experiments employing the affective lexical decision task, an information processing task which has been used to provide evidence both for, and against the role of biased attention to negative stimuli in depression is discussed. Various approaches to understanding performance on the task are shown to be compatible in light of the idea that attention may be biased towards the negative affective valence of stimuli, but not the semantic content of such stimuli in people who are depressed, and thus, depressed people may be distracted from responding to negative words on the lexical decision task by the affective valence of negative stimuli. An experiment designed to measure attention to the affective valence of stimuli is then discussed, involving the explicit identification of the affective valence of stimuli, without reference to their semantic content. Finally, a computational neural network model which implements an analog of biased attention towards the affective valence of stimuli is shown to produce behaviors similar to those illustrated by depressed and nondepressed people on the affective lexical decision task and on the valence-identification task. Conclusions are drawn from the meta-analysis, empirical data, and neural network model regarding the process by which information processing biases might be induced in a depressed individual and what might be done to help make such an individual become less depressed.