Inhibitory control as a moderator of threat-related interference biases in social anxiety

Abstract

Prior findings are mixed regarding the presence and direction of threat-related interference biases in social anxiety. The current study examined general inhibitory control (IC), measured by the classic colour-word Stroop, as a moderator of the relationship between both threat interference biases [indexed by the emotional Stroop (e-Stroop)] and several social anxiety indicators. High socially anxious undergraduate students (N = 159) completed the emotional and colour-word Stroop tasks, followed by an anxiety-inducing speech task. Participants completed measures of trait social anxiety, state anxiety before and during the speech, negative task-interfering cognitions during the speech and overall self-evaluation of speech performance. Speech duration was used to measure behavioural avoidance. In line with hypotheses, IC moderated the relationship between e-Stroop bias and every anxiety indicator (with the exception of behavioural avoidance), such that greater social-threat interference was associated with higher anxiety among those with weak IC, whereas lesser social-threat interference was associated with higher anxiety among those with strong IC. Implications for the theory and treatment of threat interference biases in socially anxious individuals are discussed.

Keywords:

• Social anxiety
• Inhibitory control
• Attentional bias

The authors are thankful for the research assistance and feedback provided by members of the Teachman programme for Anxiety, Cognition and Treatment lab, especially Emily Rakosi, Sean Norton, Eunice Cho, Ashna Contractor and Elizabeth Stevens.

This study was supported by a National Institute on Aging grant to Dr. Bethany Teachman [grant number R01AG033033].

The authors are thankful for the research assistance and feedback provided by members of the Teachman programme for Anxiety, Cognition and Treatment lab, especially Emily Rakosi, Sean Norton, Eunice Cho, Ashna Contractor and Elizabeth Stevens.

This study was supported by a National Institute on Aging grant to Dr. Bethany Teachman [grant number R01AG033033].

Notes

¹ The materials reported here are part of a larger study that examined the effectiveness of a new emotion regulation strategy, termed implementation intentions, in high socially anxious individuals. The strategy manipulation did not show condition effects on the measures included in this paper, so we collapsed across conditions for these analyses. Note, we also reran the main analyses when controlling for condition, and the pattern of results did not change. A full list of measures is available from the first author.

² Note that a masked (subliminal) version of the emotional and colour-word Stroop was also administered in the present study but is not reported here, given this paper's theoretical focus on inhibitory control as a moderator of conscious threat bias. Results from the masked Stroop are available from the first author.

³ Note a small group of participants (18.1% of the full sample) exhibited below-zero colour-word interference scores, suggesting faster responses for incongruent relative to congruent words. Notably, only three of these participants showed a speeding of more than 100 ms (i.e., one-tenth of a second) between their mean incongruent and congruent trial latencies; thus, most of the below-zero scores are quite small and can perhaps most plausibly be interpreted in terms of random noise around the mean. By contrast, 49 participants (36.3% of the sample) showed a slowing of more than 100 ms for incongruent versus congruent trials, consistent with past research that uses the colour-word Stroop task as an IC measure (see MacLeod, 2005).

⁴ For maximum clarity and consistency, the e-Stroop bias index is described in terms of greater and lesser degrees of “threat interference”, rather than “threat facilitation”, throughout this manuscript. However, as seen in Figures 1A–1D, some participants actually exhibited faster response latencies for social-threat relative to neutral words (as indicated by a below-zero e-Stroop bias score), suggesting a threat facilitation rather than a threat interference effect.