In the mood for adaptation

Cognitive conflict plays an important role in tuning cognitive control to the situation at hand. Based on earlier findings that have demonstrated emotional modulations of conflict processing, we predicted that affective states may adaptively regulate goal-directed behavior that is driven by conflict. This hypothesis was tested by measuring conflict-driven control adaptations following experimental induction of four different mood states that could be differentiated along the dimensions arousal and pleasure. After mood states were induced, 91 subjects performed a flanker task, which provides a measure of conflict adaptation. As predicted, pleasure level impacted conflict adaptation: less pleasure was associated with more conflict-driven control. Arousal level did not influence conflict adaptation. The present study suggests that affect adaptively regulates cognitive control. Implications for future research and psychopathology are discussed. Mood and adaptation 3 In the mood for adaptation: how affect regulates conflict-driven control Introduction Emotions seem to have evolved to guide organisms and their conspecifics in their struggle for survival and affective states are assumed to facilitate behavior that is adaptive to the current situational context (Morris, 1992). In particular, it has been suggested that negative mood stimulates the processing of stimuli that have a negative valence and, therefore, deserve priority. Indeed, low pleasure levels seem to induce negative information biases in attention and memory. Although it has been suggested that these biases systematically change the way people cope with negative events (cf. Gendolla, 2000), it has yet to be demonstrated how affect may play this regulating role in cognitive control adaptations. The main function of cognitive control is to adapt the cognitive system to situational demands. It has been proposed that this adaptation is driven by the detection of cognitive conflict (Botvinick et al., 2001). Evidence supporting this view comes from conflict tasks. On a given trial, subjects respond slower to target information if distracting flanker information suggests a different response. On trials following this conflict, however, flanker interference is reduced (Egner, 2007; Gratton et al., 1992), indicating that facing conflict enhances control (Botvinick et al., 2001). Numerous studies have shown that low-pleasure affect facilitates neural conflict monitoring (e.g., Luu et al., 2000). They illustrate that moods that are congruent with the negative valence inherent to conflict (Botvinick, 2007) facilitate conflict registration (cf. Rusting, 1998). Given that conflict registration is important for tuning goal-directed behavior (cf. Kerns et al., 2004), affective states that prioritize conflict processing should also induce stronger behavioral adaptations to cognitive conflict. We therefore predict that people in a low-pleasure Mood and adaptation 4 mood adapt more strongly to cognitive conflict, and are thus more likely to recruit control, than people in a high-pleasure mood. Some authors have postulated that, independent of pleasure, changes in arousal level may also influence conflict-adaptation by altering the signal-to-noise ratio of conflict information (Verguts & Notebaert, 2009). If so, conflict-driven cognitive control may be influenced by the arousal level of the current affective state. (FOOTNOTE 1). Given that pleasure level and arousal level are the two fundamental dimensions mood is assumed to vary on (Yik et al., 1999), we investigated four groups of participants who underwent a standard mood-induction manipulation before performing a conflict-evoking flanker task. Each mood group occupies one of the four quadrants derived by crossing the dimensions of pleasure and arousal (see Figure 1; cf. Jefferies et al., 2008). The four derived moods that were induced were: anxiety (low pleasure, high arousal), sadness (low pleasure, low arousal), calmness (high pleasure, low arousal), and happiness (high pleasure, high arousal). As pointed out, we predicted stronger conflict-driven adaptation effects (i.e., reductions of flanker-induced interference after conflict trials) for participants with low pleasure levels (anxious and sad participants) than for participants with high pleasure levels (calm and happy participants). Methods Participants and design Ninety-eight students participated either for payment or course credits (18-30 years old; 24 males; 11 left-handed). They were randomly assigned to one of the four mood induction groups where anxious, sad, calm, and happy moods were induced. Data from seven subjects were excluded from analyses because of response omissions on more than 20% of the trials (2), chance level task performance (3), or incompliance with instructions (2). Mood induction and assessment Mood and adaptation 5 A standard mood-induction procedure was used that combined music with imagination. This is known to induce reliable mood changes (Eich, 2007). Subjects used a headphone to listen to specific classical music samples which were validated by previous research (Jefferies et al., 2008). They were instructed to develop a particular mood by imagining and writing down a mood-appropriate event in detail, either by using the written vignette given or by recalling a similar event from their past. Music continued to play during task performance. In order to check the induction manipulation, subject were to rate their mood on a 9 x 9 pleasure x arousal grid (Russell et al., 1989). Ratings (reported as values ranging from -4 to 4) were given on a grid occasionally presented on the computer monitor during the experiment. Task We used a version of the classic flanker task (Eriksen & Eriksen, 1974) where centrally presented target stimuli are vertically flanked on either side by two response-compatible or response-incompatible stimuli. Four Dutch color words were used as targets and flankers (using a word set including “brown”, “gray”, “yellow”, and “red”, or “purple”, “green”, “orange”, and “blue”). Subject were instructed to use their index fingers, making a left response to two specific central target words and a right response to the other two target words (stimulus-response mapping counterbalanced within mood groups). A reminder of the stimulus-response mapping was shown for 15 seconds before the start of each of the two blocks of 72 trials. All trials started with a fixation cross (randomly varying intervals of 800, 1000, or 1100 ms), followed by the stimulus, which was presented until response registration or, in the case of omission, for 1500 ms. In half of the trials target and flanker stimuli would call for different responses (responseincompatible condition; I) whereas in the other half physically identical target and flanker stimuli would call for the same response (response-compatible condition; C). All trials were presented in Mood and adaptation 6 an unconstrained random sequence. Stimuli appeared in black lower-case in Arial bold font (3.5 cm wide and 5.4 cm high for the entire array) and were presented on a grey background. Participants viewed the stimuli on a 17” monitor from ≈ 60 cm.