The last two decades have seen a large body of research devoted to the study of attentional biases in addiction and other disorders. A variety of experimental tasks, derived from those used in mainstream experimental psychology, have been used to assess attentional bias. For example, MacLeod et al. (1986) adapted the visual probe task, which is described below, to demonstrate that clinically anxious patients but not nonanxious controls tend to direct their attention toward threat-related words. Since this seminal study, a large volume of research has been devoted to the characterization of such attentional biases in a variety of emotional disorders (for a recent review, see Bishop 2007). More recently, researchers have used these tasks to study attentional biases for drug-related cues in addiction. With few exceptions, these tasks do not provide a direct readout of attentional processes; instead, the allocation of ► attention must be inferred based on a secondary measure, such as response time.
Perhaps the most commonly used task is a modified version of the classic "Stroop" task. In the classic Stroop task, participants are required to name the color in which different words are printed. A highly robust observation is that when color-related words are presented in an incongruent color (e.g., the word GREEN printed in red ink), participants are relatively slow to specify the ink color, compared to a control condition (e.g., the word TABLE printed in red ink). The interpretation of this "Stroop interference" is that people automatically process the semantic content (meaning) of words that they encounter; when the semantic content of a word (e.g., the word GREEN) conflicts with the required response (to say "red"), this leads to a slowing down of color-naming, or errors in color-naming. In the modified version of this task (the addiction Stroop), participants are required to name the color in which drug-related words are printed, and their color-naming times for these drug-related words are compared to those for a control category (e.g., words related to musical instruments). See Fig. 1 for an example of the procedure and some illustrative findings. As reviewed by Cox et al. (2006), a highly robust finding in the literature is that drug users are slower to name the color in which drug-related words are presented, compared to words from a control category, but control participants do not exhibit this pattern of Stroop interference. Such Stroop interference has been demonstrated in users of a variety of different drugs, including alcohol, cannabis, cocaine, heroin, and tobacco (reviewed by Cox et al. 2006). This suggests that, compared to nonusers, drug users engage in excessive semantic processing of drug-related words, i.e., those words can "grab their attention.'' However, it is important to note that other explanations for this Stroop interference have been put forward, including delayed color-naming as a consequence of attempts to suppress the processing of drug-related words, or a generic slowdown in cognitive processing induced by the drug-related words, perhaps independently of any bias in selective attention (see Field and Cox 2008). Indeed, any Stroop interference that is observed may reflect the combined influence of all of these factors, and the specific influence of selective attention to drug-related words may be minimal; these issues mean that results from this task must be interpreted with some caution.
The visual probe task is an alternative measure of attentional bias that has been extensively used in psycho-pharmacology research, particularly over the past 5 years.
In this task (see Fig. 2), a pair of pictures are presented on the left and right of a computer screen. One of the pictures is drug-related (e.g., an image of a person smoking a cigarette), and the other picture contains no drug-related content. The pictures are shown for a short period (typically between 50 and 2,000 milliseconds (ms)), and after pictures have been removed from the computer screen, a visual probe stimulus (for example, a small dot) is presented on either the left or right of the screen, in the spatial position that had been occupied by either the drug-related or the control picture. Participants are instructed to respond to the probe as quickly as possible. As reviewed by Field and Cox (2008), a robust observation is that drug users are faster to respond to probes that replace drug-related pictures, than to probes that replace control pictures. Control participants are equally fast to respond to probes that replace drug-related and control pictures. This has been demonstrated not only with tobacco smokers in several studies, but also with cannabis users, heroin users, and heavy drinkers.
Given that people are faster to detect and respond to a stimulus if it appears in a location of the visual field that they were attending to, this finding (faster reaction times to probes that replace drug-related pictures, rather than control pictures) is usually interpreted as indicating that drug users were directing their attention toward drug-related pictures. Indeed, when participants' eye movements are monitored while they complete the task (as detailed below), there is usually a large positive correlation between the index of "attentional bias'' (derived from reaction times to probes) and the amount of time that
Attentional Bias to Drug Cues. Fig. 1. The addiction Stroop test. Participants are required to quickly identify the color in which words are printed. One list of words is drug-related; the other is neutral. In this study, adolescent heavy drinkers were slower to name the color of alcohol-related words, than to name the color of neutral words, but light drinkers took a similar amount of time to name the color of words in the two word lists. (Reprinted with permission from Field M, Christiansen P, Cole J, Goudie A (2007) Delay discounting and the alcohol Stroop in heavy drinking adolescents. Addiction 102:579-586.)
Attentional Bias to Drug Cues. Fig. 2. The visual probe task. Following a centrally presented fixation cross, a drug-related picture and a neutral picture are presented side by side on a computer screen for a brief period. After the pictures disappear, a visual probe is presented, and participants are required to rapidly respond to the probe. In this study, tobacco smokers were faster to respond to probes that replaced smoking-related pictures, than to probes that replaced neutral pictures, but nonsmokers did not show this difference. (Reprinted with permission from Mogg K, Bradley BP, Field M, De Houwer J (2003) Eye movements to smoking-related pictures in smokers: relationship between attentional biases and implicit and explicit measures of stimulus valence. Addiction 98:825-836.)
people maintain their gaze on the drug-related cues. So, it appears valid to use reaction times to visual probes to infer the allocation of attention to drug-related cues that precede those visual probes. A further issue with the visual probe task is that researchers have experimentally manipulated the amount of time that picture pairs are presented on the screen before they are removed and replaced by the visual probe. By varying the stimulus onset asynchrony (SOA), in this way, for example, using very short exposure durations (e.g., 50-200 ms) or longer durations (e.g.,
2,000 ms), it is assumed that reaction time can capture the extent to which cues "grab" the attention (with very short SOAs) or "hold" the attention (with longer SOAs). Some studies suggest within-subject and between-group differences in attentional bias when short and long SOAs are used, which is consistent with the notion that these different SOAs can be used to measure different aspects of cognitive processing of drug cues. However, there is currently some controversy over whether reaction times obtained from the visual probe task can ever be used to index orienting of attention, regardless of how brief the SOA is (Field and Cox 2008).
Given these issues with interpreting results from the visual probe and addiction Stroop tasks, there is a need to use measures of attentional bias that provide more direct readouts of selective attention. Eye movement monitoring is one such measure, because people are usually attending to whatever is currently the focus of gaze. Some researchers have monitored participants' eye movements while they complete a visual probe task in which drug-related and control pictures are presented. It is possible to then look at the orienting of attention to drug-related cues, by measuring which picture (drug-related or control) participants initially direct their gaze toward on each trial of the task. It is also possible to measure the maintenance of attention on, or latency to disengage attention from, drug-related cues, by comparing the amount of time that participants direct their gaze toward drug-related pictures with the amount of time that they direct their gaze toward control pictures. Tobacco smokers, ► cocaine users, and cannabis users (but not control participants) tend to preferentially maintain their gaze on drug-related pictures; however, the evidence for a bias in the initial orienting of attention toward drug-related pictures is currently equivocal (Field and Cox 2008).
Additional laboratory measures have been used in recent years to investigate attentional bias for drug-related cues. These include tasks such as the flicker induced change blindness task and the attentional blink task (see Field and Cox 2008). However, these are not discussed in detail in this section as they are not currently widely used, although both methods may have advantages over existing measures. The final measure that we consider here is ► electroencephalography (EEG), particularly the study of event-related potentials (ERPs) elicited by drug-related cues. EEG recording involves attaching electrodes to various sites on the scalp of participants (see Fig. 3) that measure the electrical activity produced by the brain. Participants are then shown drug-related (or control) pictures, and activity in the cortex can be measured in the form of electrical activity on the scalp in response to presentation of the pictures (ERPs). Although there are many different forms of ERPs, which differ in their latency, magnitude, and brain region of origin (see separate entry), researchers have focused on ERP components such as the P300 (a slow positive wave that typically occurs about 300 ms after stimulus presentation). This particular ERP component has been implicated in selective attention, such that its magnitude (in response to a visually presented stimulus) seems to correlate highly with the degree to which participants attend to that stimulus. This has particularly been observed for motivationally relevant stimuli. In several studies, it has been demonstrated that P300 magnitude in response to drug-related
Attentional Bias to Drug Cues. Fig. 3. Measuring event-related potentials (ERPs) in response to smoking-related cues. Smoking-related and neutral pictures were presented for 2 s, and participants were asked to watch the pictures attentively. The ERP results indicate that both the P300 and the subsequent slow wave amplitudes in response to smoking-related pictures are significantly enhanced in smokers compared to nonsmokers at frontal and central sites, whereas the magnitude of the P300 and SPW amplitudes in response to neutral pictures does not differ between the groups. Accordingly, it can be concluded that smokers show more bias for smoking-related pictures than smokers. (Data published in Littel M, Franken IHA (2007) The effects of prolonged abstinence on the processing of smoking cues: an ERP study among smokers, ex-smokers and never-smokers. J Psychopharmacol 21:873-882.)
cues is significantly larger in drug-users than in control participants (see Fig. 3 for an example). This has been demonstrated in heroin, cocaine, alcohol, and tobacco users (e.g., Franken et al. 2008).
Was this article helpful?