1. Introduction
Children and adolescents all have uncertainties, anxiety and worries and some develop into problems (disorders). Anxiety disorders are highly prevalent among children and adolescents. They represent a substantial portion of mental disorders in children and adolescents, reaching approximately 10% by age 16 (Costello, Egger & Angold, 2003). In a Dutch survey (312 participants) Graaf, Have and Dorsselaer (2010) found that more than 10% of the Dutch adolescents (13-17 years) had an anxiety disorder, a half year before the survey. Most of them had social and specific phobia. 4-8% of the children (0-12 years) suffered from internalizing problems, such as anxiety and mood disorders (Graaf et al., 2010). Anxiety disorders in childhood are associated with high levels of emotional distress and result in academic and social impairment (Alkozei, Cooper & Creswell, 2014; Mychailyszyn, Mendez & Kendall, 2010). Moreover, children with anxiety disorders are at increased risk for anxiety disorders later in life (Bress, Meyer & Hajcak, 2015). Because of these consequences, it is important to identify factors that contribute to the development and maintenance of anxiety.
It is generally assumed that cognitive factors, such as abnormalities in information processing (e.g. catastrophic misinterpretation of bodily sensations and symptoms) and certain beliefs (e.g. anxiety symptoms have harmful consequences and uncertainty is unacceptable), play an important role in maintaining (childhood) anxiety disorders. Therefore, cognitive behavior therapy has been established as an effective treatment for childhood anxiety disorders (Creswell, Murray & Cooper, 2013). However, 40-50% of children retain their primary anxiety disorder diagnosis after treatment (James et al., 2005). That’s why a more developed understanding of the cognitive factors of maintaining anxiety in children is important. The cognitive theory of Beck (1985) considers anxiety as a result of abnormalities in information processing and feedback processing. When we look at information processing, three cognitive mechanisms play an important role.
Interpretation bias First of all, a number of studies have demonstrated that anxious children interpret ambiguous situations more negative. For example, clinically anxious children (7-14 years) are more likely than non-anxious controls to interpret ambiguous stories in a threatening manner (Creswell, Schniering, & Rapee, 2005). Additionally, adolescents (11-16 years) with high levels of social anxiety symptoms had significantly higher levels of threat interpretation bias than those with low social anxiety symptoms (Miers, Blöte, Bögels & Westenberg, 2008). But findings are inconsistent, particularly among studies with children at a younger age (7–12 years). While Alkozei, Cooper and Creswell (2014) have found
significant group differences regarding interpretation of threat on tasks involving ambiguous scenarios, other studies have failed to find differences between children with anxiety disorders and non-anxious children of the same age on similar tasks (Creswell, Murray & Cooper, 2013; Waters, Wharton, Zimmer-Gembeck & Craske, 2008).
Attentional bias The second mechanism that is assumed to play an important role in information processing in anxious children is the attentional bias, which refers to the tendency to selectively attend to signals of threat. Bar-Haim and colleagues (2017) analyzed the results of 172 studies that investigated attentional biases in anxiety. Their analyses provided evidence that anxious individuals (adults) are more sensitive and in favour of threat-related stimuli (Bar-Haim et al., 2007). Since the publication of this study there has been an increase in research investigating this attentional bias in children with anxiety, but these studies vary in the age of the sample. Therefore, Dudeney, Sharpe & Hunt (2015) did a systematic research to determine whether an attentional bias for threat-related information is evident for children with anxiety and, if so, under what conditions this bias is present. The meta-analysis concluded that anxious children do show a similar bias towards threatening stimuli as has been documented in adults, but to a lesser degree and this bias is moderated by age, such that the difference between anxious and control children increases with age. The ability of a child to exert effortful control by inhibiting automatic responding develops with age as executive functioning ability improves throughout childhood and adolescence. The results of this meta-analysis suggest that children with anxiety may fail to develop an adequate level of effortful control in attentional and inhibitory processes as they mature (Dudeney et al., 2015).
So far, there seems to be evidence that childhood anxiety is a result from an interpretation bias as well as an attentional bias. Anxious children interpret ambiguous situations as more threatening and focus more on signals of threat (Muris, Merckelbach & Damsma, 2000), but this is moderated with age.
Intolerance of uncertainty Thirdly, a more recently investigated mechanism, that seems to play an important role in developing and maintaining anxiety is called Intolerance of Uncertainty (IU). IU has been defined as a cognitive bias that affects how a person perceives, interprets and responds to uncertain situations. More specific, people high in IU believe that negative events—how small their probability of occurrence is—are unacceptable, should be avoided, reflect badly on themselves and block constructive action (Boelen et al., 2010). They feel very uncomfortable in unstructured and uncertain situations (Li, Li & Luo, 2005).
Intolerance to uncertainty is often linked with an attentional bias for threatening information (Bar-Haim et al., 2007). When anxious individuals are exposed to uncertain situations, anxious individuals tend to focus on threatening aspects of the situation (Muris et al., 2000). They get nervous and interpret the ambiguous information as if it is in fact threatening information. IU has contributed to our understanding of adult anxiety disorders, but there is a paucity of research in child and/or adolescent samples (Boelen et al., 2010). Boelen and colleagues (2010) used the Intolerance Uncertainty Scale short-form (IUS-12) to measure IU in relation to anxiety. He found, using data from 191 Dutch adolescents (14-18 years), that IU was specifically related to worry and social anxiety, but not depression.
Laugesen et al. (2003) examined the relationship between worry and 4 cognitive variables (i.e., IU, positive beliefs about worry, negative problem orientation and thought suppression) in 528 adolescents (14 -18 years). They found that IU was the most important variable in distinguishing between moderate and high worry groups. Barahmand (2008) examined association between IU and worry among 197 adolescents from Iran and found IU to be associated with worry among girls but not boys, suggesting that the role of IU in adolescent worry differs as a function of gender.
Summarizing, the interpretation bias, the attentional bias and IU seem to be important aspects of developing and maintaining anxiety. These cognitive errors play an important role in the manner information from the outside world is been processed.
The information processing errors may also influence the way feedback and performance outcomes are processed (Gu, Huang & Luo, 2010). To explore the relationship between feedback processing and anxiety more deeply, we use psychophysiological correlates as an objective measure. Psychophysiological correlates of feedback processing which can be used, is the Feedback Related Negativity (FRN) (Takács, Kóbor, Janacsek, Honbolygó, Csépe & Németh, 2015). The FRN is elicited by external feedback. It occurs when outcomes are worse than anticipated (Bress et al., 2015). The FRN is related to processing the outcome value and motivational significance of ongoing events. It peaks between 200 and 300 ms following the feedback stimulus. The FRN is larger (more negative) after losses than after gains (Bellebaum & Daum, 2008). In addition, the FRN has been demonstrated to be larger for unexpected negative outcomes than expected ones (Hajcak, Moser, Holroyd & Simons, 2007; Holroyd, Hajcak & Larsen, 2006). Holroyd and colleagues (2006) showed that the magnitude of a reward/punishment did not affect the FRN amplitude. Negative and neutral
feedback elicited equal FRN amplitudes.
Altered performance monitoring is often reported in anxiety related syndromes (Takács et al., 2015) and could lead to difficulties in everyday risky decision making. That’s why
Takács et al. (2015) investigated feedback processing measuring the FRN in 26 undergraduate students (mean age: 21.30) while performing the Balloon Analogue Risk Task (BART). To understand ambiguous decision making, the balloon analogue risk task (BART) is a widely used experimental tool. In this gambling paradigm, participants are asked to pump a balloon on a screen, and each pump is associated either with a reward (positive feedback) or with a balloon burst (negative feedback). After each pump, participants could decide to collect the accumulated reward or to take the risk and pump the balloon further. If the balloon burst, the reward is lost. As a rule, each successful pump increases the probability of a burst in the next trial; however, this regularity is not transparent for the participants. Participants make a series of choices with increasing risk before they face an external punishment or change their course of action and choose the safe option. Therefore, without explicitly knowing the probabilities of rewards and losses, BART performance requires ambiguous decision making. Participants were divided into a High Trait Anxiety (HTA) group and a Low Trait Anxiety (LTA) group by median split of the state-trait anxiety inventory (STAI). The STAI a widely-used tool to measure trait anxiety. Takács et al. (2015) found that the FRN was smaller for the HTA group than for the LTA group after receiving negative feedback (after the balloon popped). They proposed that students of the HTA-group had more pessimistic expectations about whether or not the balloon would burst. These pessimistic expectation bias is triggered by the ambiguity in the BART-task. Because the FRN reflects the prediction error of the outcome, it seems that negative events met the expectations of HTA individuals, indicated by a smaller FRN (Takács et al., 2015).
Because anxious people find it difficult to deal with uncertain situations it would be interesting to find out how ambiguous feedback is processed. Gu, Huang & Luo (2010) added this ambiguous feedback condition. They aimed to explore the relationship between anxiety and ambiguous outcome evaluation by using the FRN. Seventy-nine college students participated in a mass screening with the Spielberger’s State-Trait Anxiety Inventory (STAI). Participants who scored high in trait anxiety (upper 25% of the distribution) were assigned to the high-trait anxiety group (HTA), whereas the participants who scored low were assigned to the low-trait anxiety group (LTA). As a result, 33 participants (mean age = 23.61) were
included in the study, 16 in the HTA group and 17 in the LTA group. They played a monetary gambling task by choosing between two alternatives (for example 5 or 25 coins) presented on a screen. After they had made a choice there were three possible outcomes. A positive outcome meant that the participant had won as many coins as he had chosen, a negative outcome indicated that he/she had loss the coins, neutral feedback meant the participant neither won nor lost and ambiguous feedback meant that the valence could be positive, negative or neutral. Gu et al. (2010) found the amplitude of the FRN after negative feedback versus positive feedback was significant larger for LTA individuals compared to HTA individuals. However, there was no significant difference in the FRN elicited by ambiguous feedback between the two groups. Maybe the processing of ambiguous feedback consumed more cognitive resources than definite outcomes. Therefore, this kind of the FN, in response to ambiguous information, was maybe more complicated than the classic (well-studied) FN response (Gu et al., 2010).
In another study 253 Chinese students (mean age = 20.53) participated in a mass screening with the STAI and were assigned to the HTA-group (upper 25 % of the distribution) or the LTA group (lower 25% of the distribution). It resulted in a group of 34 participants, 17 in the HTA group and 17 in the LTA-group. The same monetary gambling paradigm was used. The ‘+’ symbol indicated that the participant won as many points as he/she chose in this trial, while the ‘−’ symbol indicated the reverse. The ‘0’ indicated a neutral feedback, which means the participant neither won nor lost. The ‘*’ symbol indicated an ambiguous feedback. Prior to the experiment, the participant was told that the ambiguous feedback was an uninformative outcome, valence could be positive, negative or neutral, but it was impossible for him/her to guess the real valence in any given trial. Compared to the LTA group, the HTA group exhibited a smaller FRN following negative feedback. Because HTA adults are more likely to expect negative outcome, their expectation error is less, which could lead to a smaller FRN. However, Gu et al. (2010) found the FRN after ambiguous feedback to be larger for the HTA group than for LTA adults. As said before, anxiety is linked with Intolerance to Uncertainty. Consequently, anxious people might think the ambiguous outcome is very intolerable. Consistent with this idea, the FRN was larger following ambiguous outcomes than following negative outcomes in the HTA group, suggesting that the ambiguous outcomes were even worse than negative outcomes in the HTA participants’ minds (Gu et al., 2010).
The former studies show different findings. Couple of studies showed that the FRN is smaller
after negative feedback for a HTA group versus a LTA group (Takács et al., 2015; Gu et al., 2010). Although, Gu, Huang and Luo (2010) did not find a significant difference. Gu et al. (2010) found FRN larger for HTA than LTA after ambiguous feedback, however Gu, Huang and Luo (2010) did not find any differences.
Not many studies investigated the FRN in early adolescents. Hämmerer, Müller and Lindenberger (2011) found the FRN was larger for both positive and negative feedback for children (9-11 years) compared to early-adolescents (13-14 years) and adults. Participants performed a probabilistic reinforcement learning task. They were presented with different pairs of Japanese characters that were each associated with probabilistic gains and losses. Choosing one of the two symbols resulted in gaining or losing 10 points. However, within each pair, one symbol had a higher probability of resulting in a gain than did the other symbol (Hämmerer et al (2011). This result supported an earlier study of Eppinger, Mock and Kray (2009) with 18 adults (19-24 years) and 17 children (10-12 years). Participants performed a probabilistic learning task in which the participants were showed 36 colored images of objects, that belonged to one of six categories (clothes, vehicles, fruit, vegetables, furniture, domestic appliances). Participants were asked to make a two-choice decision upon presentation of the stimulus and to press one of two response keys. FRN-amplitudes after positive feedback did not differ significantly between the two groups. However, children showed a larger FRN after negative feedback than adults. Findings suggest that children might be more sensitive to negative feedback during learning than adults (Eppinger et al., 2009). Findings of a study by Crowley et al. (2013) showed that FRN amplitudes were larger for non-rewarded than rewarded feedback. They also found that the FRN amplitudes of the children (10-12 years) and the early-adolescents (13-14 years) were larger compared to the adolescents (15-17 years). Crowley et al. (2013) showed that developmental effects reported in learning tasks (Eppinger et al., 2009; Hämmerer et al., 2011) are not only a function of a learning process, but can also be seen in a non-learning task. Concluding, it looks like in childhood FRN amplitudes are larger than in adolescence, where smaller (adult-like) FRN responses appear.
Studies that have investigated the association between FRN amplitude and anxiety in (early-) adolescents are scarce. Bress et al. (2015) investigated the association between FRN and anxiety in a study with 25 adolescents (11-13 years). They did not find an association between
the FRN and anxiety. Participants in this task could either win or lose; ambiguous feedback wasn’t investigated.
In summary, given the high prevalence rates and lasting consequences of childhood anxiety a more deeply understanding of anxiety and the role of feedback processing in early adolescents is important. Intolerance of uncertainty is a cognitive mechanism that seems to play an important role in developing and maintaining anxiety. It might also influence ambiguous feedback processing, measured by psychophysiological correlates.
In the current study, we investigated the FRN in (early-) adolescents, aged 9 to 13 years old, after positive, negative and ambiguous feedback. By adding an ambiguous feedback condition, we investigated whether high anxious adolescents show different FRN amplitudes in an ambiguous and uncertain feedback situation compared to low anxious adolescents. To our knowledge no studies have examined feedback processing after ambiguous feedback in (early) adolescents before.
Participants filled out two self-report questionnaires on anxiety: ZBVK and the SCARED-NL. The ZBVK questionnaire is widely used to measure anxiety symptoms in children and adolescents. But from a clinical point of view, it might also be useful to use a questionnaire which is more related to the DSM classification system. The items of the SCARED-NL parallel the DSM-IV classification of anxiety disorders. Therefore, we decided to use both questionnaires. Based on their scores on the questionnaires we assigned the participants to either the HTA or LTA group, by means of a median split. We used the developed research paradigm based on the 2-option forced-choice gambling task used by Gu and colleagues (2010).
Firstly, we expected the FRN to be larger after losses, thus after receiving negative feedback, than after gains (positive feedback). We also investigates whether differences occur in the FRN after ambiguous and positive feedback, and the FRN after negative and ambiguous feedback. Secondly, we investigated whether anxiety influences the FRN after positive, negative or ambiguous feedback. In line with previous findings (Gu et al., 2010), we expect that the FRN after negative feedback will be smaller in the HTA group than the LTA group. HTA-children will have more negative expectations of the outcome. Therefore, negative feedback will not be very different from their expectations and their FRN will be smaller (Gu et al., 2010). However, we expected the FRN in the HTA group after ambiguous feedback to be larger than in the LTA group. The HTA children interpret the ambiguous feedback as more
threatening and they might be intolerant to this kind of uncertainty, which results in an enlarged FRN compared to the LTA-group. We did not expect differences in FRN between HTA and LTA after positive feedback.