Introduction
The temperamental factors behavioral inhibition (BI)
1 and attentional control (AC) have often been associated with the development of internalizing behaviors in youth as well as in adults (e.g. Johnson et al.
2003; Jorm et al.
1999; Muris et al.
2005). BI can be seen as reactive temperament and represents a proneness towards anxiety and a sensitivity towards signals of punishment and non-reward (Carver and White
1994). From a developmental point of view one could argue that BI causes the child to avoid stressful situations, which in turn could enhance the anxious symptoms. AC can be seen as regulative temperament, and represents the ability to focus and switch attention, and is a specific component of the overarching temperament trait of effortful control (Rothbart et al.
2004). Reactive temperament involves individual differences in emotional arousability, whereas regulative temperament modulates this reactivity. Eisenberg et al. (
2000) have proposed that attentional control may especially be important in reducing internalizing symptoms, such as anxiety and sadness in children. As such, AC may be a moderator or buffering factor in the association between BI and internalizing disorders. High levels of AC would thus prevent those high on BI to develop anxiety disorders and depression disorders (Muris and Ollendick
2005; Nigg
2006).
In support of the view that high BI sets people at risk for developing internalizing psychopathology, previous work found positive relationships between BI and symptoms of DSM-IV based anxiety disorders and depression. Cross-sectional studies showed a firm relation between BI and anxiety in children (Muris et al.
2005) and between BI and depression in children (Nigg
2006). In prospective research Johnson et al. (
2003) showed that BI was associated with anxiety disorders and depression in young adults. Some research has been done into the relevance of AC for the various internalizing disorders. There is empirical evidence that low AC is linked to self-reported anxiety symptoms in children between age 5–8 (Eisenberg et al.
2001), children aged 9–13 (Muris et al.
2004; Muris et al.
2006), adolescents aged 12–15 (Muris
2006), and adolescents aged 12–18 year (Vervoort et al. in press). In addition, it has been shown that effortful control is associated with symptoms of depression in children aged 9–13 (Muris et al.
2006) and adolescents aged 11–17 (Verstraeten et al.
2009). Prospectively, averaged deficits in attentional control predicted which individuals still had high scores on internalizing problems or even deteriorated over time at time 2 compared to time 1 (Eisenberg et al.
2009). With regard to the proposed moderating role of AC in the association between BI and internalizing disorders, three previous studies focusing on negative emotionality (similar to behavioral inhibition) have shown that the combination of high negative emotionality and low effortful control was associated with the highest levels of internalizing symptoms (Muris
2006; Muris et al.
2007; Oldehinkel et al.
2007).
Thus far research on BI, AC, and their interaction has focused on specific disorders (e.g. social phobia), or on internalizing behavior in general. It remains therefore to be examined whether low AC, and/or high BI can be best considered as more general characteristics of people suffering from symptoms of anxiety and/or mood disorders, or that the importance of low AC and/or high BI varies across the various internalizing disorders. The same holds for the proposed synergistic influence of AC and BI on the development of anxiety and mood symptomatology.
From a theoretical point of view we have no a priori reason to believe that BI and AC are more relevant to some anxiety disorders than others (see also: Nigg
2006). The various etiological models of specific DSM-IV anxiety disorders include both the presence of BI as a proneness towards perceived threat as well as the absence of AC as an impairment in the ability to regulate this proneness towards perceived threat. In these models the tendency to perceive threat involves negative social evaluations as in social phobia (Clark and Wells
1995), loss of control as in panic disorder (Clark
1986), future disaster as in generalized anxiety disorder (Wells
2005), and so forth, while low AC constitutes the inability to disengage attentional focus from such threat stimuli (Clark and Wells
1995; Wells
2005). While we thus regard BI as relevant for all anxiety disorder, depressed individuals are not primarily pre-occupied with threat, rather, they have a negative focus on self, the future, and the world (Beck et al.
1979). Such a negative focus is dissimilar to BI as defined as a proneness towards anxiety and a sensitivity towards signals of punishment and non-reward (Carver and White
1994). However, AC does have a role in the model by Beck et al. (
1979), in which depressed people are unable to switch their attention away from their negative thoughts. Note that non of the existing theoretical models of anxiety and depression are explicit as to a mutually enhancing effect of high BI and low AC or, alternatively, on a buffering role of AC on the relation between BI and anxiety and depression. The above leads to the hypothesis that both BI and AC will play an important role in all anxiety disorders. In depression, however, we hypothesize a smaller role for BI than in anxiety disorders and a similar role for AC compared to anxiety disorders. Intuitively, the proposed synergistic effect is plausible for all anxiety disorders as well as for depression, and this will be explored.
In sum, the present study seeks to expand our insights into the role of low AC combined with high BI in symptoms of the various DSM-IV anxiety and mood disorders. In addition to considering main effects we will test the hypothesis that AC may exert its protective effect mostly among those who are highly behaviorally inhibited (i.e., as a moderator). We extend previous work by studying the full range of anxiety and mood symptomatology, and expect a substantial role for BI and AC in all symptomatology related to the main anxiety disorders, and a more substantial role for AC than for BI in depressive symptomatology.
Discussion
The present study examined the independent and joint associations of BI, AC, and their interaction with DSM-IV based internalizing problems. Results of the current study can be summarized as follows: (i) BI was positively and AC was negatively related to the severity of symptoms of all DSM-IV based internalizing disorders; (ii) The BI × AC interaction effect consistently showed cumulative predictive validity where high AC reduced the effect of BI on all DSM-IV based internalizing problem dimensions; (iii) In symptoms of depression, AC played a larger role than in the DSM-IV based anxiety dimensions. Symptoms of depression were most strongly related to AC, while symptoms of social phobia, separation anxiety disorder, generalized anxiety disorder, panic disorder and obsessive compulsive disorder were most strongly related to BI.
Replicating previous research (e.g., Muris et al.
2005; Johnson et al.
2003), the present results showed a robust relationship between high BI and DSM-IV based internalizing disorders. Next to that our results broaden current knowledge by showing this relation for different DSM-IV based anxiety disorders. Importantly, a negative and robust association was also found for low AC, as was shown before by Eisenberg et al. (
2001). Although BI systematically showed the strongest relationship with symptoms of anxiety, AC was also found to have substantial additional explanatory power to symptoms of DSM-IV anxiety disorders. Thus the present results indicate that low AC is an important additional characteristic of all DSM-IV based anxiety domains. Comparing the various DSM-IV based disorders, it appears that for symptoms of social phobia BI is significantly more influential than in the other measured dimensions. This difference however may be due to the fact that the items in the BIS subscale partly focus on a sensitivity for social evaluation, hence, to the specific operationalization of BI in the present study (Carver and White
1994) rather than reflecting a stronger association with social phobia per se. These findings would hold only if the results were replicated using different models such as the Kagan model of BI (Kagan et al.
1988) or the broader construct of negative affectivity from the temperament model of Rothbart et al. (
2000). For all other DSM-IV based anxiety disorders, the strength of the association with BI was highly similar. For symptoms of depression, we found that AC plays a larger role than for the DSM-IV based anxiety dimensions which was according to expectations. Also as expected, no differences emerged across the various anxiety dimensions with respect to AC.
Perhaps most important, BI × AC independently contributed to the model. This again was a robust finding in all measured anxiety and depression dimensions. These results are consistent with previous research that focused on the role of effortful control in internalizing problems in general (Oldehinkel et al.
2007), in depression (Verstraeten et al.
2009; Muris
2006), and in anxiety (Muris et al.
2004). These findings extend previous findings by showing that this buffering effect holds for symptoms of generalized anxiety disorder, social phobia, separation anxiety disorder, panic disorder, obsessive–compulsive disorder, and major depressive disorder. The results are consistent with the view that high AC may function as a buffer between BI and the development of DSM-IV based internalizing complaints (Muris and Ollendick
2005), while individuals with high BI combined with low AC report the highest levels of DSM-IV based internalizing symptoms. If prospective studies would show that high BI and low AC precede DSM-IV based internalizing symptoms, this might be of importance for identifying adolescents at high risk of developing an anxiety or mood disorder. It should be noted however that the interaction had small additive value on top of BI and AC, with BI and AC each having the largest associations, thus the identification of potentially at risk adolescents should focus primarily on high BI or low AC.
Our results are not only of theoretical interest but also provide clues for clinical interventions. More specifically, our findings support the view that training AC might be a helpful strategy in the prevention and/or treatment of internalizing disorders. In support of this, there is already some preliminary evidence for the useful application of cognitive control training in depressed adults (Siegle et al.
2007) and of task concentration training in social phobia (Mulkens et al.
2001). It would be interesting to see whether also in children and/or adolescents AC training would help to reduce symptoms of internalizing disorders. In a similar vein it would be interesting to examine whether AC training may also be helpful in the prevention of the onset or recurrence of internalizing disorders in adolescence.
One possible limitation is that AC was assessed by a self-report measure. It can not be ruled out that subjective AC does not (completely) match with actual capacity of self-regulation by AC (Reinholdt-Dunne et al.
2009). It would therefore be important for future research to test whether similar results would emerge when using a behavioral measure of executive functioning (EF) such as the Attentional Network Task (Reinholdt-Dunne et al.
2009; ANT; Fan et al.
2002). Some promising research in this field has been done by Derakshan et al. (
2009), who showed that in undergraduate students anxiety is related to attentional control as measured by task-switching tasks. Adding a parent or teacher report of AC may be another helpful strategy to differentiate between individuals’ perception of their AC and their actual capacity to regulate their attention. This differentiation is also of importance for training applications, with training either focused on direct training AC through a specific attentional control training or focused on changing the perceptions of AC, for instance through cognitive therapy, or both.
Second, the present sample may have been prone to selection bias, since invited adolescents were all informed beforehand on the nature of our research. That is, the current data were collected as part of a large screening with the ultimate aim to select groups of adolescents to be included in training for the prevention of anxiety. When comparing the means on the self-report questionnaire (RCADS) with same aged adolescents measured in the context of a large longitudinal cohort study in the Netherlands (Van Oort et al.
2009), the adolescents in our sample have higher scores on all DSM-IV based internalizing dimensions. However, there seems to be no obvious reason to assume that the present correlational findings would be different in a less anxious sample.
Third, one could argue that in particular the cross-sectional study of the association between BI and anxiety is somewhat tautological. While the BIS scale measures the predisposition to anxiety, the DSM-IV scales measure the actual experience, i.e., the state, of anxiety (see Jorm et al.
1999). We acknowledge that these measurements are likely to be confounded by one another. Nonetheless, the association between BI and anxiety and mood disorders has already been established prospectively, as well as by means of different methods and operationalizations of the BI construct (Kagan et al.
1988; Biederman et al.
1993). The additional value of the current paper lies in the fact that we showed that AC had additional value above and beyond BI for all DSM-IV based anxiety and mood problems, by itself and in interaction with BI.
Finally, it should be acknowledged that the cross-sectional design of our study does not allow any firm conclusion regarding the direction of found associations. To arrive at more solid grounds in this respect it would be important to test presently found associations in a longitudinal design and to determine prospectively whether the combination of AC and BI and their interaction have prognostic value for the onset of mood and/or anxiety disorders.
To conclude, both high BI and low AC were found to have independent and mutually enhancing associations with all anxiety and depression dimensions that were derived from the DSM-IV. Since AC may buffer the relationship between BI and internalizing problems as well as protect for internalizing problems by itself, the training of attentional control strategies may be a promising route for modifying or preventing symptoms of anxiety and depression in adolescents.