The effect of attention training on emotional vulnerability and food consumption following a stressor.

摘要

Individuals with anxiety typically display an attentional bias toward threat that may contribute causally to the development and maintenance of anxiety. C. MacLeod, E. Rutherford, L. Campbell, G. Ebsworthy, and L. Holker (2002) showed that manipulating attentional bias toward and away from threat can modify emotional vulnerability. This experiment attempted to replicate and extend this finding to undergraduates (N = 67) reporting average anxiety, but above-average emotional overeating tendencies. An objective outcome was added (calories consumed during a "taste test").;Participants were double-blindly assigned to an "attend-neutral" attention training condition of the dot probe task (in which the probes replaced neutral words to train a bias toward neutral words) or an "attend-negative" condition (in which the probes replaced negative words). It was hypothesized that the attend-neutral group would report less negative affect following a stressor and consume fewer calories than the attend-negative group.;Reaction times to each of the two types of trials (where probes replaced neutral or negative words) showed high internal consistency. However, Cronbach's alpha for attentional bias scores (the difference between reaction times to detect probes replacing neutral words and probes replacing negative words) was low pre- and posttraining (.50 and .33).;Perhaps related to the dot probe task's low reliability, the attend-neutral group's bias score did not change. The attend-negative group, however, developed the predicted bias toward negative words. Contrary to predictions, both groups reported equivalent negative affect increases following the stressor and consumed equivalent calories during the "taste test." In exploratory analyses of the top one-third of the sample on trait anxiety, the attend-negative group showed a trend toward the predicted greater increase in negative affect following the stressor compared with the attend-neutral group, r = .39 (a medium effect size). The two groups, however, consumed equivalent calories. A clinically or subclinically anxious sample that displays a bias toward threat seems to increase the likelihood of training a bias away from threat. At 1-month follow-up, unexpectedly, the attend-negative group reported decreased general distress compared to the attend-neutral group, who reported an increase, possibly suggesting that training toward threat could function as exposure and decrease anxiety.