Selective attention to unmasked and masked threatening words: relationships to trait anger and anxiety

https://doi.org/10.1016/S0191-8869(00)00160-4Get rights and content

Abstract

In two experiments, unmasked and masked versions of emotional Stroop tasks were used to investigate the relationships between trait anger, trait anxiety and selective attention to threatening words. For the masked tasks individual neutral-emotional recognition thresholds were determined. In Experiment 1, participants were selected on low and high trait anger scores. Results showed a significant interaction between GROUP (low vs. high anger) and WORD-TYPE (neutral vs. threat) in the unmasked task exclusively, due to interference in the high trait anger group and facilitation in the low trait anger group. In Experiment 2, participants were selected on low and high trait anxiety scores. Here, a significant GROUP×WORD-TYPE interaction was restricted to the masked task, with the high trait anxiety group showing an attentional bias for threatening words. Theoretical and methodological issues are discussed.

Introduction

Cognitive theories of emotion (e.g. Beck, 1976, Bower, 1981, Williams et al., 1997) claim that anxiety leads to selective attention for cues related to threat. This so-called attentional bias for threat is seen as a crucial factor in the etiology and maintenance of anxiety disorders (Williams, MacLeod & Mathews, 1996). The most successful paradigm for demonstrating attentional bias is the emotional Stroop task. In this task the color of different words has to be named while the content of the words must be ignored. Attentional bias is observed as increased latencies (interference) when color-naming threat as compared to neutral words. A large number of studies have shown that anxious individuals display this bias (for a review, see Williams et al., 1996).

Emotional Stroop effects have also been shown in masked conditions. In the masked task, conscious recognition of the word is prevented by short presentation and fast replacement of the target stimulus by a masking display in the same color. Attentional bias for masked threat words has been shown in clinically anxious individuals and high anxious normals (e.g. Mogg et al., 1993, Van den Hout et al., 1997, Van den Hout et al., 1995). A further intriguing observation is the occurrence of attentional bias in the masked but not in the unmasked task in high anxious normals (MacLeod & Hagan, 1992, MacLeod & Rotherford, 1992, Mogg et al., 1993). A consciously mediated mood-controlling strategy in the unmasked task is suggested to be responsible for this dissociation (Mathews & McLeod, 1994). Given that the dissociation has not been shown in clinically anxious individuals, the inability to control interference for unmasked threatening words is suggested to be diagnostic of pathology (see Williams et al., 1996; but see Amir, McNally, Riemann, Burns, Lorenz & Mullen, 1996).

Section snippets

Experiment 1

Research on selective processing of negatively valenced material has been primarily concerned with anxiety and depression (Williams et al., 1997). However, anger too might be associated with a heightened sensitivity to threat. Beck (1976) states that the threat of being harmed results in anxiety while the threat of being provoked or ill-treated results in anger. Whether a specific situation is perceived as being harmful or provoking requires a personal interpretation. Some people are more

Discussion

The results of Experiment 1 demonstrate differential responding to unmasked threatening words in the low and high trait anger groups. The significant GROUP×WORD-TYPE interaction depended on individuals high in trait anger showing interference and individuals low in trait anger showing facilitation (i.e. the low trait anger participants were faster at color naming threatening words compared to neutral words). Measured from zero (no attentional bias) the effects of these groups were, however, not

Experiment 2

In Experiment 2, using a research design comparable to the foregoing, we again aimed to investigate selective attention to threatening words, but now used participants scoring low and high on trait anxiety measures. As in Experiment 1, selective attention to threatening words was expected in the masked task for the high anxious group whereas no firm predictions could be made for the unmasked task.

References (28)

  • G. Bower

    Mood and memory

    American Psychologist

    (1981)
  • B.P. Bradley et al.

    Attentional bias for threatening facial expressions in anxietymanipulation of stimulus duration

    Cognition and Emotion

    (1998)
  • B.P. Bradley et al.

    Selective processing of negative informationeffects of clinical anxiety, concurrent depression and awareness

    Journal of Abnormal Psychology

    (1995)
  • E. Diener et al.

    The independence of positive and negative affect

    Journal of Personality and Social Psychology

    (1985)
  • Cited by (56)

    • Neutral and threatening distracter word stimuli are unnecessarily stored in working memory but do not differ in their degree of working memory storage

      2021, Biological Psychology
      Citation Excerpt :

      This suggests that the distracter words were semantically processed in our task design. Previous reports demonstrate that individuals show enhanced attention towards threat-related stimuli (Anderson & Britton, 2019; Dowd et al., 2016; Hopkins et al., 2016; Mulckhuyse, 2018; Schmidt et al., 2015; Schupp et al., 2004), including when emotional word stimuli are used (Algom et al., 2004; Boehme et al., 2015; Kousta et al., 2009; Van den Heuvel et al., 2005; van Honk et al., 2001; Wabnitz et al., 2016; Yeung & Fernandes, 2019a). In addition, although anxious individuals have been shown to demonstrate attentional bias towards threat-related stimuli (Aftanas et al., 2003; Fox et al., 2001; Kim et al., 2018; Mogg & Bradley, 2016; Mogg et al., 1997; Nelson et al., 2015; Pacheco-Unguetti et al., 2010; Quigley et al., 2012; Raeder et al., 2019; Shah et al., 2018; Wieser et al., 2018; Yao et al., 2019), the findings regarding this effect using emotional words is mixed, with some reporting a lack of this enhanced attentional bias effect (Moritz et al., 2008; Pishyar et al., 2004).

    • Trait anger in relation to neural and behavioral correlates of response inhibition and error-processing

      2016, International Journal of Psychophysiology
      Citation Excerpt :

      HTA participants scored 21 or higher and LTA scored 15 or lower on the Dutch version of the Trait Anger Subscale (TAS) from the State Trait Anger Expression Inventory-2 (Hovens et al., 2014) respectively at the time of testing. The TAS has adequate psychometric qualities (Hovens et al., 2014; Spielberger, 1999) and similar cut-off criteria have been used in previous studies to select subclinical angry samples (e.g., Eckhardt and Cohen, 1997; van Honk et al., 2001). The resulting LTA group consisted of 45 participants (71.7% women; M age = 20.76, SD = 2.44) with a mean score of 13.13 (SD = 1.42) on the TAS.

    • Nonconscious attention bias to threat is correlated with anterior cingulate cortex gray matter volume: A voxel-based morphometry result and replication

      2012, NeuroImage
      Citation Excerpt :

      Thus, early preconscious attention biases to threat appear to influence states of worry and stress reactivity, which may be associated with elevated ACC volume. In addition to anxiety, other negative emotion-related individual differences and/or disorders, may have hyper-vigilance to threat, exaggerated amygdala activity, and abnormal ACC volumes, such as post-traumatic stress disorder, depression, and heightened trait anger (Bryant and Harvey, 1997; Carlson et al., 2010a; Kasai et al., 2008; Kitayama et al., 2006; Mogg and Bradley, 2005; Mogg et al., 1995; Rauch et al., 2000, 2003; Sheline et al., 2001; van Honk et al., 2001a, 2001b). Taken together, these results highlight the close relationship between the disposition for negative emotion-related traits, the amygdala–ACC network, and biased attention to threat.

    • Transdiagnostic cognitive processes in high trait anger

      2011, Clinical Psychology Review
      Citation Excerpt :

      Smith and Waterman (2005) found that HTA, in comparison with LTA, was associated with significantly longer response latencies for unmasked aggression-related words in males on the emotional stroop test. Van Honk, Tuiten, de Haan, van den Hout, and Stam (2001b) found that HTA was associated with selective attention for unmasked threat-related words on the emotional stroop test. This effect was not found for words that were masked.

    View all citing articles on Scopus
    View full text