Resumen de Estudio de los mecanismos atencionales en el procesamiento afectivo de participantes con ansiedad elevada vs. baja
Many theorists argue that emotion and attention are related to one another because they both deal with information processing priorities (Oatley, & Johnson-Laird, 1987). For example, fear is associated with activation of defensive systems involved in detecting and responding to danger and this includes attentional mechanisms to facilitate identification of potentially dangerous stimuli in the environment (Öhman, Flykt, & Esteves, 2001). From the evolutionary perspective, people direct more cognitive processing toward those important aspects of the world like threatening stimuli because of its obvious survival value. However, this adaptative function might lead to the development of emotional disorder like anxiety when the presence of threat stimulus interferes with the requirements of other goals (Matwes, & MacLeod, 1985, 1986).
In this way, many models of anxiety have put forward that there is increased tendency in people with high anxiety to orient attention towards threat, and this bias might be involved in the origin and/or maintenance of anxiety disorders (Eysenck, 1992; Mathews, & MacLeod, 1994; Williams, Watts, MacLeod, & Mathews, 1997). Frequently used cognitive-experimental paradigms to examine these attentional biases are the dot probe task (MacLeod, Mathews, & Tata, 1986) visual search task (Hansen, & Hansen, 1988) or stroop task (Williams, Mathews, & MacLeod, 1996), all of them supporting the attentional bias towards treat in high anxiety. However, the pattern of results observed with these tasks is ambiguous as to whether the attentional bias can be better described as an increased attentional capture by threating stimuli, or rather as a reduced capacity to disengage attention from them. So, results can be interpreted either as threatening stimuli excessively capturing attention, thus leading to a threateningly view of the world by people with high anxiety (capture hypothesis) (Mathews, & Mackintosh, 1998; Mogg, & Bradley, 1998; Williams et al., 1997), or as an increased difficulty that they suffer to shift or disengage attention from threat-related stimuli (disengaging hypothesis) (e.g. Fox, Russo, Bowles, & Dutton, 2001).
The exogenous cueing paradigm (Posner, 1980) has been used to investigate this problem, as this task was originally designed to investigate spatial covert orienting of visual attention, which can be broken down into three elementary operations, namely attentional disengaging, moving and engaging (Posner, Inhoff, Friedrich, & Cohen, 1987). In this paradigm, participants have to respond only to targets, which ignoring visual cues. Typically targets can appear in one of two peripheral markers, one to the right and the other to the left of fixation. The target is preceded by the abrupt onset of a peripheral cue, which can be presented in one of the same two target locations. Thus, in most of the trials, cue and target appear at the same spatial location (valid trials), whereas in the remaining trials, cue and target appear in opposite locations (invalid trials). Especially when targets appear shortly after the cue, participants are faster and more precise responding to targets on valid trials as compared to invalid trials, which is called 'orienting or validity effect'.
In the adaptation of this paradigm to the study of emotional biases in attentional orienting, participants with high vs. low trait or state anxiety carry out an exogenous cueing task, in which the target stimulus is preceded by an emotional cue, usually a threatening vs. emotionally positive or neutral stimulus. Any evidence regarding an increased orienting effect for threatening cues observed in participants with high trait or state anxiety is taken as evidence of an attentional bias in these individuals. Moreover, if the increased orienting effect is due to responses on valid trials involving negative cues being faster or more precise than those on valid trials with neutral or positive cues, this is taken as evidencing a bias in the engagement process of attention, which would support the engaging or attentional capture hypothesis. However, if responses on invalid trials with negative cues are slower or less precise than those on invalid trials with neutral or positive cues, this taken as evidence for a deficit in disengaging attention from threatening material, thus supporting the disengaging hypothesis. As most studies have found results supporting this last hypothesis, negative cues typically only influencing response times on invalid trials, researchers have concluded that the studied attentional bias involves a specific difficulty in disengaging attention from the location of threatening stimuli in people with high anxiety (Fox et al., 2001, Yiend, & Mathews, 2001, Broomfield, & Turpin, 2005).
Fox, Russo, y Dutton (2002) tested the disengaging hypothesis directly. In this case, emotional cues were presented long time before the target, in order to investigate involuntary orienting of attention. Thus, in contrast to previous studies, the cue-target stimulus onset asynchrony (SOA) was increased up to one second, and the same proportion of valid and invalid trials was presented. Under these conditions we expect participants to be slower and/or less precise responding to targets in valid trials as compared to invalid trials, an effect that is called 'Inhibition of Return effect' (IOR; Posner, Rafal, Choate & Vaughan, 1985). In fact, this was the effect observed in low anxiety individuals for all types of cues. However, in anxious people the IOR effect disappeared with threat cues but not with non-threat cue (Fox et al., 2002). The reduction or elimination of the IOR effect with threatening cues was interpreted as reflecting the reluctance of anxious people to disengage attention from negative stimuli; as long as attention is not disengaged from the cued location, the IOR effect is not present. Taken together the evidence seems to support the disengaging hypothesis. On the other hand, many models of anxiety propose that the bias towards threat is presumed to operate at an automatic, preattentive stage (Beck, 1976; Bower, 1981; Williams et al., 1997). The exogenous cueing paradigm has been used to study whether the exaggerated capture or the hindered disengaging observed in people with high anxiety operates at a preattentive stage by presenting the emotional cue below conscious thresholds (Correa, Fox, Carmona, Noguera, Lupiáñez, & Tudela, 2002). Results supported the attentional capture hypothesis in high trait anxiety participants, as responses on valid trials were faster for threatening than for neutral cues, especially for individuals who did not consciously perceive the emotionality of the cue.
However, it is important to note that the standard exogenous cueing procedure might not be sensitive enough to measure differences in attentional capture, as a function of the emotionality of the cue, due to ceiling effects. Note that a single abrupt onset stimulus is presented before the target (i.e., the cue), and that any peripheral salient enough cue might produce the standard maximum orienting effect (Ruz, & Lupiáñez, 2002). Therefore attention might be first directed towards the exogenous cue, independently of its emotional value, thus leading to a ceiling effect (e.g. Fox et al., 2002). This is even more likely, if we take into account that in many of the experiments cues were predictive of the target location, and long cue-target SOAs were used. As a result, research aiming at distinguishing between the attentional capture and disengaging hypotheses might have only investigated the disengaging hypothesis.
The aim of the current research was to adapt the exogenous cueing paradigm to study the attentional capture hypothesis in anxious people. A different strategy was used in order to avoid ceiling effects. In particular, we manipulated the emotional valence of the target instead of manipulating the emotional valence of the cue, and used long enough cue-target SOAs suitable to measure Inhibition of Return (slower responses on valid trials). Considering that IOR reflects a reduced attentional capture for targets appearing at previously cued locations (i.e., a bias against attention being captured again at or returning to, this location, Lupiáñez, Decaix, Sieroff, Chokron, Milliken, & Bartolomeo, 2004), if threatening stimuli do have an advantage in capturing attention, IOR might be overcome by the privileged capacity of threatening stimuli to capture attention when they are the target, so that no IOR will be observed for these stimuli. Besides, we wanted to explore whether the typical masked Stroop interference effect founded in anxious people (e.g. Bradley, Mogg, Millar, &White, 1995) is modulated by spatial attention. By doing so, we wanted to study whether attentional biases towards threatening stimuli take place independently of spatial attention at a preattentive stage. More specifically, the main goals of the thesis are, by Chapters, the following: In Chapter 2 are presented two experiments aiming at investigating whether participants with high vs. low trait anxiety have attentional biases regarding attentional capture by ontogenetically learnt stimuli (i.e., words). The IOR procedure described above was used with neutral vs. emotionally charged words, with the prediction that high trait anxiety individual would show reduced IOR for negative words, as these words would be special for anxious people in capturing attention, thus overcoming the cognitive inhibitory effect. In order to appropriately select the words we had to elaborate and validate a data base of emotional words, which is presented on Chapter 5.
In Chapter 3 two experiments are presented, which aimed at investigating whether participants with high and low trait (Experiment 1) and state (Experiment 2) anxiety do show attentional biases regarding attentional capture by phylogenetically learnt stimuli like threatening faces. To induce anxiety we designed and validated with physiological and self report measures an original procedure, which is presented on Chapter 6.
In Chapter 4 an experiment is reported, aiming at investigating whether attentional capture by threatening stimuli is mediated by spatial orienting. Participants with high vs. low trait anxiety were to categorize the gender of target faces, in a masked emotional Stroop paradigm, embedded within an exogenous cueing procedure. Faces were first shortly presented either with a neutral, positive or negative expression, and rapidly substituted by the same face with neutral expression, so that participants did not notice the emotional manipulation. Shortly before the target, a peripheral cue was presented either in the same (valid trials) or the opposite location (invalid trials). By doing so, we wanted to know whether the mechanism underlying emotional Stroop is related to, or is independent of spatial orienting of attention.
Finally, in Chapter 7 an extended summary and an integrative discussion of the results are presented, in the context of the discussion about attentional capture vs. disengagement as the mechanisms underlying attentional biases in anxiety.
