- Email: [email protected]
Words may not be enough! No increased emotional Stroop effect in obsessive–compulsive disorder
Behaviour Research and Therapy 46 (2008) 1101–1104
Contents lists available at ScienceDirect
Behaviour Research and Therapy journal homepage: www.elsevier.com/locate/brat
Shorter communication
Words may not be enough! No increased emotional Stroop effect in obsessive– compulsive disorder Steffen Moritz*, Benny-Kristin Fischer, Birgit Hottenrott, Michael Kellner, Susanne Fricke, Sarah Randjbar, Lena Jelinek University Medical Center Hamburg-Eppendorf, Department of Psychiatry and Psychotherapy, Martinistrasse 52, 20246 Hamburg, Germany
a r t i c l e i n f o
a b s t r a c t
Article history: Received 6 February 2008 Received in revised form 27 April 2008 Accepted 1 May 2008
Conflicting evidence has been obtained whether or not patients diagnosed with obsessive–compulsive disorder (OCD) share an attentional bias towards disorder-related stimuli. Some of these inconsistencies can be accounted for by suboptimal stimuli selection. In consideration of the heterogeneity of OCD, we investigated Stroop interference effects for two classes of OCD items (i.e., washing and checking) in 23 OCD patients and 23 healthy controls. In order to cover prevalent OCD concerns, item compilation was based on experts’ appraisals. Patients neither displayed greater immediate as well as delayed Stroop interference nor any bias for OCD and subtype-congruent stimuli. On the contrary, for washing-related items, OCD patients, and here especially washers, displayed facilitation relative to healthy controls. Although the present study at first sight refutes the notion of an attentional bias in OCD in contrast to other anxiety disorders, several potential moderators need to be considered before this account is ultimately dismissed. In particular, an attentional bias may only be elicited using visual material that is more attention-grabbing than verbal stimuli. Finally, blockwise instead of random item administration and greater consideration of individual relevance may be crucial prerequisites for the effect to emerge. Ó 2008 Elsevier Ltd. All rights reserved.
Keywords: Obsessive–compulsive disorder Stroop Washer Checker Attentional bias Interference
Introduction There is equivocal evidence whether or not patients with obsessive–compulsive disorder (OCD) share an attentional bias for concern-related material, a response pattern consistently reported for patients with other anxiety disorders and depression (Mathews & MacLeod, 2005; Williams, Meathews, & MacLeod, 1996). While OCD is classified among the anxiety disorders, the absence of such a bias in many studies (Kampman, Keijsers, Verbraak, Naring, & Hoogduin, 2002; Kyrios & Iob, 1998; McNally, Riemann, Louro, Lukach, & Kim, 1992; McNeil, Tucker, Miranda, Lewin, & Nordgren, 1999; Moritz, Jacobsen et al., 2004; Moritz & von Muhlenen, 2005; Moritz & von Mu¨hlenen, 2008; Unoki, Kasuga, Matsushima, & Ohta, 1999: supraliminal presentation) but not all studies (Foa, Ilai, McCarthy, Shoyer, & Murdock, 1993; Lavy, van Oppen, & van den Hout, 1994; Unoki et al., 1999: subliminal presentation) has been interpreted (Summerfeldt & Endler, 1998) as a further evidence to segregate OCD from the anxiety disorder spectrum besides differences regarding phenomenology and treatment. Alternatively, problems to detect an attentional bias in patients with OCD may be rooted in the idiosyncratic nature of OCD: Stimuli
* Corresponding author. Tel.: þ49 40 42803 6565; fax: þ49 40 42803 7566. E-mail address: [email protected] (S. Moritz). 0005-7967/$ – see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.brat.2008.05.005
eliciting concern or arousal largely differ even among patients who share the same subtype (i.e., washing, checking, ordering, hoarding). While one checker may only be concerned when leaving his house (i.e., worry that the door has not been properly locked), another patient may only worry while driving (i.e., fear to have hit someone over). In contrast, in phobias, for example spider phobia, the set of fear-eliciting stimuli is more homogeneous. Moreover, not all prior relevant studies have looked at OCD-relevant stimuli but some administered general anxiety, panic, or depression stimuli (McNally et al., 1992, 1994; McNeil et al., 1999; Moritz, Jacobsen et al., 2004). As many patients do not experience fear/ panic but disgust, tension, ‘‘not just right’’ experiences (Coles, Heimberg, Frost, & Steketee, 2005) or an unspecified urge preceding compulsions, such material may not be best suited to capture an attentional bias. A prominent test to elicit an attentional bias is the emotional Stroop paradigm (MacLeod,1991). Whereas in a conventional Stroop task (Stroop, 1935), the print-colour of an incongruently typed colour word has to be named resulting in strong interference (i.e., reading is more automatized than the target response colour naming), the emotional Stroop uses disorder-related words. As personally or disorder-relevant items are attention-grabbing, the subject’s attention is diverted from the primary task towards the distractor, thus delaying the response. As an alternative account, rumination or a lowered threshold for disorder-relevant material has been
1102
S. Moritz et al. / Behaviour Research and Therapy 46 (2008) 1101–1104
postulated to explain this response pattern in mental disorders. Response latencies in the experimental condition are compared to a control condition (neutral stimuli, colour bars, or meaningless strings of characters). Usually, the emotional Stroop effect is weaker than interference effects elicited by colour words. For the present study, we compared OCD and healthy subjects on an emotional variant of the Stroop task involving OCD-relevant stimuli. We looked separately at items relevant for the two most prevalent OCD subtypes (checking and washing). This was done, as an attentional bias may not operate for all subtypes alike and to take into account the heterogeneous phenomenology of OCD (e.g., the word ‘‘dirt’’ is likely more arousing for washers than checkers). Stimuli were chosen via expert ratings to cover the most prevalent concerns for each subtype. Apart from subtype other prominent moderators for group differences in OCD apart from subtypes, such as severity of OCD as well as comorbid affective symptoms, were carefully considered. Finally, in view of recent studies (McKenna & Sharma, 2004; for a discussion see Phaf & Kan, 2007) that have detected slow but not fast emotional Stroop interference, we have looked at reaction times both for immediate (trial N) and subsequent trials (N þ 1). To the best of our knowledge, it has not been investigated whether interference effects need more time to evolve in OCD patients. Methods Participants Twenty-three patients meeting DSM-IV criteria for OCD were recruited from the Department of Psychiatry and Psychotherapy of the University Medical Center Hamburg/Germany (gender: 11 male, 12 female; age: M ¼ 35.57, SD ¼ 10.54; years of formal school education: M ¼ 11.04, SD ¼ 1.69). Presence of OCD was confirmed with the Mini International Neuropsychiatric Interview (MINI, Sheehan et al., 1998). Any history of psychotic and manic symptoms (grandiose or paranoid ideas, hallucinations of any modality) led to exclusion. Most patients were medicated with psychotropic drugs (15 patients were prescribed antidepressant drugs, six of these also with a neuroleptic agent). Eleven patients had a depressive episode secondary to OCD and eight patients were co-diagnosed with another anxiety disorder. Twenty-three subjects served as healthy controls who were recruited via an established subject pool or word-of-mouth (gender: 12 male, 11 female; age: M ¼ 31.39, SD ¼ 10.66; years of formal school education: M ¼ 11.74, SD ¼ 1.48). The MINI interview verified absence of any axis I psychiatric disorder in controls. None of the participants suffered from neurological disorders (e.g., stroke, epilepsy) including OCD spectrum disorders (e.g., Tourette’s syndrome). Written informed consent was obtained from all participants prior to baseline assessment. The German version of the YaleBrown Obsessive–Compulsive Scale (Y-BOCS) (Goodman et al., 1989; German translation by Hand & Bu¨ttner-Westphal, 1991) served as an index of OCD severity (total: M ¼ 22.73; SD ¼ 6.61). For assessing depression, the Hamilton depression rating scale (HDRS; 17 item version; Hamilton, 1960) was administered blind to neurocognitive status (total score: M ¼ 12.18, SD ¼ 7.08). Subscores for these scales were computed using algorithms derived from factor analytic studies (Moritz et al., 2002; Moritz, Meier, Hand, Schick, & Jahn, 2004). We administered the Hamburg Obsessional Compulsive Inventory (HOCI; Klepsch, Zaworka, Hand, Lu¨nenschloss, & Jauernig, 1991) to specify OCD subtypes. The HOCI is a self-rating instrument that assesses core obsessions (e.g., thoughts of doing harm to self/others) and compulsions (e.g., checking) along six scales. Seven of the patients displayed both checking and washing symptoms, while each three patients only showed checking or washing symptoms.
Experimental task The emotional Stroop task was constructed using SuperLabÒ software and was individually presented via a Macintosh computer. Following a short practice task with six items to familiarize subjects with the task requirements, 15 stimuli from each of the nine different conditions were presented in random order. Words for the non-OCD conditions were similar to stimuli used in a prior experiment (Moritz, Jacobsen et al., 2004) and compiled following consensus ratings by experienced clinicians. The OCD items were newly collected after an iterative reduction process guided by ratings from six clinical psychologists or psychiatrists with extensive experience with OCD patients. The final set of words was rated OCD-relevant by the majority of assessors. The conditions were similar according to word length, initial letters (plosive vs. non-plosive) and word frequency (all contrasts p > 0.1). The emotional word conditions were as follows: OCD-checking relevant (e.g., lock, accident), OCD-washing relevant (e.g., blood, dirt), anxiety (e.g., anxiety, panic), depression (e.g., weakness, loneliness), positive (e.g., beauty, success), and paranoia-relevant (e.g., spy, surveillance). There were two non-affective control conditions: neutral (e.g., bag, table) and Stroop interference (e.g., incongruently written colour words). Naming of horizontal colour bars served as baseline condition. As noted, the non-OCD conditions were similar to a previous experiment with one exception: in the neutral condition, two formerly presented items that could be considered OCDrelevant (soap, sink) were substituted for the present experiment. Colors were evenly distributed across all conditions: red (Ger¨ n), yellow (gelb), and blue (blau). Words were man: rot), green (gru presented in German language; all participants were German native speakers. Each trial consisted of the following sequence: a small fixation point (300 ms) was first followed by a 200 ms blank trial. Subsequently, the target stimulus (font 30, type Geneva) was shown until the individual triggered the voice-key via microphone. Participants were instructed to respond as fast as possible and to avoid making mistakes. Following each response, the experimenter pressed a button to denote correct (i.e., ‘‘C’’) and incorrect responses (i.e., ‘‘X’’) which initiated the next trial. Strategy of data analysis Reaction time latencies were recorded for correct trials only. Reaction times faster than 300 ms and slower than 4000 ms were excluded from analyses. As dependent variables, we computed reaction times from the median response latencies for target words (N) as well as subsequent words (trial N þ 1), as we were interested if an attentional bias persists over time and/or manifests after a delay due to, for example, rumination. A 2 9 mixed analysis of variance (ANOVA) was conducted with the experimental conditions as within-subject and Group (healthy, OCD) OCD as betweensubject factor. Since our hypothesis was undirected, for exploratory purposes we also computed single comparisons. Results Sociodemographic background variables and immediate interference Samples did not differ with respect to sociodemographic background variables (age, gender, school education; all p > .1). In the mixed ANOVA neither the effect of Group, F(1,44) ¼ 0.84, p > .3, h2partial ¼ .02, nor the interaction of Group Condition, F(8,352) ¼ 0.78, p > .3, h2partial ¼ .02, surpassed the level of significance. As can be seen in Fig. 1, the significant effect of Condition is mainly owing to discrepant reaction times in the control condition
S. Moritz et al. / Behaviour Research and Therapy 46 (2008) 1101–1104
reaction times in ms
850 healthy OCD
800 750 700 650 600
ty
na colo m r in g St ro op
xi e
n pr e de
an
ss io
ve si ti po
no id
l ut ra
ra pa
ne
as hi w
ch e
ck in
g
ng
550
Fig. 1. The two-way ANOVA did not yield any group differences regarding condition. However, post hoc analyses revealed a slight speeding of OCD patients for washingrelevant words in the colour bar naming condition but not in the neutral condition relative to controls.
colour naming versus the conventional Stroop interference condition, F(8,352) ¼ 28.50, p < .001, h2partial ¼ 0.39. When we subtracted reaction times in the two OCD and the Stroop interference condition from the neutral word and non-word control condition (colour naming), no evidence for greater interference emerged for OCD patients even before Bonferroni-correction for multiple comparisons. On the contrary, for washing-relevant material, the difference achieved significance (washing vs. colour bars, p ¼ .03) or trend level (washing vs. neutral word, p ¼ .09) indicating facilitation in patients. For checking-relevant items (p > .3) and conventional Stroop interference (p > .1) no such effects emerged. Delayed interference For trials following the critical trials (N þ 1) groups displayed a similar response pattern: the main effect of Group and the interaction did not produce significant effects (both p > 0.3, both h2partial < 0.028). Moreover, none of the post hoc comparisons (e.g., washing condition – neutral) yielded significance (p > 0.3). Subgroup analysis According to HOCI scores, the OCD sample was split into washers and checkers. None of the ANOVA results yielded significance except for the washing condition with colour naming as baseline, F(2,41) ¼ 3.34, p < .05. Post hoc comparisons showed that washers (M ¼ 48 ms) showed significantly less interference than healthy participants (M ¼ 124 ms). OCD-non-washers performed intermediate (M ¼ 92 ms). Finally, we investigated whether a subgroup of patients would display strong interference. To our surprise, for washing-relevant material six healthy subjects were among the 10 participants with the highest interference latencies, whereas for checking-relevant items both five healthy and OCD patients displayed highest interference. Discussion The present study revealed no evidence for greater interference on any of the experimental conditions of an emotional Stroop task in OCD. Specifically, no retardation was detected for OCD-relevant material – irrespective of subtype. The OCD stimuli were collected after a thorough iterative reduction process involving experts’ opinion to assure that central OCD concerns were covered. Moreover, special care was devoted to equate conditions for several important moderators such as words length and language frequency. Our findings unlikely reflect a lack of statistical power, as
1103
group differences were rather small and for the washing-related words even a counter-intuitive facilitation occurred: here, especially washers, displayed less interference! While it is thus tempting to infer that OCD is not associated with an attentional bias, particularly in view of many studies unable to detect such a cognitive preference in OCD (see introduction) and in line with claims made by Summerfeldt and Endler (1998), we should not jump to conclusions. First, although we gathered stimuli that deal with the prominent concerns of OCD washers and checkers, we cannot rule out that for a subgroup of patients sharing idiosyncratic and very isolated worries these items were not attention-grabbing enough thus attenuating group differences. The counter-intuitive facilitation for washing-related material could be owed to greater familiarity in patients with such stimuli acting against a pop-out effect of the distractor, which may only be found for the most fear-evoking stimuli. We would like to point out that Kyrios and Iob (1998) also found facilitation for OCD and other anxiety words in both OCD and healthy subjects. Interestingly, Lavy et al. (1994) reported greater interference for negatively valenced OCD words while for positively valenced OCD words the response pattern reversed in OCD patients. Nonetheless, future studies may benefit from greater consideration of individual concerns and may also employ paradigms that may tease apart the relative contribution of vigilance and disengagement problems in the emergence of Stroop interference (Amir, Elias, Klumpp, & Przeworski, 2003). If the attentional bias in OCD is existent, it is apparently more subtle than in PTSD patients or patients with simple phobias. Therefore, a verbal paradigm may not be sufficiently sensitive to elicit a heightened attentional bias, especially when involving random item administration precluding prolonged confrontation with disorder-related items. The worries of OCD patients are mostly triggered by visual cues or images (e.g., dirt on the table, key lock) and a visual paradigm may thus be fairer to put this hypothesis to test. Indeed, a recent study using neutral, anxiety and OCD pictures in the context of a modified inhibition of return (IOR) paradigm was able to show greater distractibility of OCD patients for OCD-relevant but not neutral and anxiety material relative to controls (Moritz et al., submitted for publication). Therefore, while we are safe to assume that any attentional bias in OCD is less pronounced than in other anxiety disorders, especially PTSD, researchers should continue to pursue this account with other paradigms before ultimately dismissing it. Moreover, subjective appraisal ratings may help to determine whether stimuli sufficiently cover patients’ concerns. In line with this, we found that washers but not checkers were slowed on a primed Stroop task for OCD-relevant targets when the sample was split in subjects endorsing more than one-third of the items as personally relevant versus those that did not meet this criterion (Fischer, 2007). So far, stimuli of past research were either rated by other subjects for relevance (Foa et al., 1993; Lavy et al., 1994; Unoki et al., 1999), experts (Moritz, Jacobsen et al., 2004, and present study) or not at all. None of these studies performed a subanalysis of items with special personal relevance. Finally, the blockwise administration of conditions may be more powerful to elicit effects (e.g., Lavy et al., 1994) if an attentional bias in OCD needs time to evolve and/or needs multiple triggers and prolonged exhibition (see Phaf & Kan, 2007), although the present analysis of delayed effects in trial N þ 1 did not provide evidence for this assumption.
References Amir, N., Elias, J., Klumpp, H., & Przeworski, A. (2003). Attentional bias to threat in social phobia: facilitated processing of threat or difficulty disengaging attention from threat? Behaviour Research and Therapy, 41, 1325–1335. Coles, M. E., Heimberg, R. G., Frost, R. O., & Steketee, G. (2005). Not just right experiences and obsessive–compulsive features: experimental and selfmonitoring perspectives. Behaviour Research and Therapy, 43, 153–167.
1104
S. Moritz et al. / Behaviour Research and Therapy 46 (2008) 1101–1104
Fischer, B.-K. (2007). Untersuchung kognitiver Einbahnstraßen bei Zwang unter Verwendung einer semantischen Priming-Aufgabe mit sto¨rungsrelevantem Material [Investigation of cognitive ‘‘one-way streets’’ in obsessive–compulsive disorder using a semantic priming-task with disorder-relevant material]. Master thesis, University of Hamburg, Hamburg. Foa, E. B., Ilai, D., McCarthy, P. R., Shoyer, B., & Murdock, T. (1993). Information processing in obsessive–compulsive disorder. Cognitive Therapy and Research, 17, 173–189. Goodman, W. K., Price, L. H., Rasmussen, S. A., Mazure, C., Fleischmann, R. L., Hill, C. L., et al. (1989). The Yale-Brown obsessive compulsive scale. I. Development, use, and reliability. Archives of General Psychiatry, 46, 1006–1011. Hamilton, M. (1960). A rating scale for depression. Journal of Neurology, Neurosurgery and Psychiatry, 23, 56–62. Hand, I., & Bu¨ttner-Westphal, H. (1991). Die Yale-Brown obsessive compulsive scale (Y-BOCS): Ein halbstrukturiertes Interview zur Beurteilung des Schweregrades von Denk- und Handlungszwa¨ngen. Verhaltenstherapie, 1, 223–225. Kampman, M., Keijsers, G. P., Verbraak, M. J., Naring, G., & Hoogduin, C. A. (2002). The emotional Stroop: a comparison of panic disorder patients, obsessive– compulsive patients, and normal controls, in two experiments. Journal of Anxiety Disorders, 16, 425–441. Klepsch, R., Zaworka, W., Hand, I., Lu¨nenschloss, K., & Jauernig, G. (1991). Derivation and validation of the Hamburg obsession/compulsion inventory-short form (HOCI-S): first results. Psychological Assessment, 3, 196–201. Kyrios, M., & Iob, M. A. (1998). Automatic and strategic processing in obsessive– compulsive disorder: attentional bias, cognitive avoidance or more complex phenomena? Journal of Anxiety Disorders, 12, 271–292. Lavy, E., van Oppen, P., & van den Hout, M. (1994). Selective processing of emotional information in obsessive compulsive disorder. Behaviour Research and Therapy, 32, 243–246. MacLeod, C. M. (1991). Half a century of research on the Stroop effect: an integrative review. Psychological Bulletin, 109, 163–203. Mathews, A., & MacLeod, C. (2005). Cognitive vulnerability to emotional disorders. Annual Review of Clinical Psychology, 1, 167–195. McKenna, F. P., & Sharma, D. (2004). Reversing the emotional Stroop effect reveals that it is not what it seems: the role of fast and slow components. Journal of Experimental Psychology: Learning, Memory, and Cognition, 30, 382–392. McNally, R. J., Amir, N., Louro, C. E., Lukach, B. M., Riemann, B. C., & Calamari, J. E. (1994). Cognitive processing of idiographic emotional information in panic disorder. Behaviour Research and Therapy, 32, 119–122.
McNally, R. J., Riemann, B. C., Louro, C. E., Lukach, B. M., & Kim, E. (1992). Cognitive processing of emotional information in panic disorder. Behaviour Research Therapy, 30, 143–149. McNeil, D. W., Tucker, P., Miranda, R., Jr., Lewin, M. R., & Nordgren, J. C. (1999). Response to depression and anxiety Stroop stimuli in posttraumatic stress disorder, obsessive–compulsive disorder, and major depressive disorder. Journal of Nervous and Mental Disease, 187, 512–516. Moritz, S., Jacobsen, D., Kloss, M., Fricke, S., Rufer, M., & Hand, I. (2004). Examination of emotional Stroop interference in obsessive–compulsive disorder. Behaviour Research and Therapy, 42, 671–682. Moritz, S., Kloss, M., Jacobsen, D., Wein, C., Fricke, S., & Hand, I. (2002). Dimensional structure of the Yale-Brown obsessive compulsive scale (Y-BOCS). Psychiatry Research, 109, 193–199. Moritz, S., Meier, B., Hand, I., Schick, M., & Jahn, H. (2004). Dimensional structure of the Hamilton depression rating scale in patients with obsessive–compulsive disorder. Psychiatry Research, 125, 171–180. Moritz, S., Mu¨hlenen, A. v., Jelinek, L., Randjbar, S., Ruhe, C., Fischer, B.-K., et al. Evidence for an attentional bias for washing and checking relevant stimuli in obsessive–compulsive disorder, submitted for publication. Moritz, S., & von Muhlenen, A. (2005). Inhibition of return in patients with obsessive–compulsive disorder. Journal of Anxiety Disorders, 19, 117–126. Moritz, S., & von Mu¨hlenen, A. (2008). Investigation of an attentional bias for fear-related material in obsessive–compulsive checkers. Depression and Anxiety, 25, 225–229. Phaf, R. H., & Kan, K. J. (2007). The automaticity of emotional Stroop: a metaanalysis. Journal of Behavior Therapy and Experimental Psychiatry, 38, 184–199. Sheehan, D. V., Lecrubier, Y., Sheehan, K. H., Amorim, P., Janavs, J., Weiller, E., et al. (1998). The MINI International Neuropsychiatric Interview (M.I.N.I.): The development and validation of a structured diagnostic psychiatric interview. Journal of Clinical Psychiatry, 59(Suppl. 20), 22–33. Stroop, J. R. (1935). Studies of interference in serial verbal reactions. Journal of Experimental Psychology, 18, 643–662. Summerfeldt, L. J., & Endler, N. S. (1998). Examining the evidence for anxietyrelated cognitive biases in obsessive–compulsive disorder. Journal of Anxiety Disorders, 12, 579–598. Unoki, K., Kasuga, T., Matsushima, E., & Ohta, K. (1999). Attentional processing of emotional information in obsessive–compulsive disorder. Psychiatry and Clinical Neuroscience, 53, 635–642. Williams, J. M. G., Meathews, A., & MacLeod, C. (1996). The emotional Stroop task and psychopathology. Psychological Bulletin, 120, 3–24.
Contents lists available at ScienceDirect
Behaviour Research and Therapy journal homepage: www.elsevier.com/locate/brat
Shorter communication
Words may not be enough! No increased emotional Stroop effect in obsessive– compulsive disorder Steffen Moritz*, Benny-Kristin Fischer, Birgit Hottenrott, Michael Kellner, Susanne Fricke, Sarah Randjbar, Lena Jelinek University Medical Center Hamburg-Eppendorf, Department of Psychiatry and Psychotherapy, Martinistrasse 52, 20246 Hamburg, Germany
a r t i c l e i n f o
a b s t r a c t
Article history: Received 6 February 2008 Received in revised form 27 April 2008 Accepted 1 May 2008
Conflicting evidence has been obtained whether or not patients diagnosed with obsessive–compulsive disorder (OCD) share an attentional bias towards disorder-related stimuli. Some of these inconsistencies can be accounted for by suboptimal stimuli selection. In consideration of the heterogeneity of OCD, we investigated Stroop interference effects for two classes of OCD items (i.e., washing and checking) in 23 OCD patients and 23 healthy controls. In order to cover prevalent OCD concerns, item compilation was based on experts’ appraisals. Patients neither displayed greater immediate as well as delayed Stroop interference nor any bias for OCD and subtype-congruent stimuli. On the contrary, for washing-related items, OCD patients, and here especially washers, displayed facilitation relative to healthy controls. Although the present study at first sight refutes the notion of an attentional bias in OCD in contrast to other anxiety disorders, several potential moderators need to be considered before this account is ultimately dismissed. In particular, an attentional bias may only be elicited using visual material that is more attention-grabbing than verbal stimuli. Finally, blockwise instead of random item administration and greater consideration of individual relevance may be crucial prerequisites for the effect to emerge. Ó 2008 Elsevier Ltd. All rights reserved.
Keywords: Obsessive–compulsive disorder Stroop Washer Checker Attentional bias Interference
Introduction There is equivocal evidence whether or not patients with obsessive–compulsive disorder (OCD) share an attentional bias for concern-related material, a response pattern consistently reported for patients with other anxiety disorders and depression (Mathews & MacLeod, 2005; Williams, Meathews, & MacLeod, 1996). While OCD is classified among the anxiety disorders, the absence of such a bias in many studies (Kampman, Keijsers, Verbraak, Naring, & Hoogduin, 2002; Kyrios & Iob, 1998; McNally, Riemann, Louro, Lukach, & Kim, 1992; McNeil, Tucker, Miranda, Lewin, & Nordgren, 1999; Moritz, Jacobsen et al., 2004; Moritz & von Muhlenen, 2005; Moritz & von Mu¨hlenen, 2008; Unoki, Kasuga, Matsushima, & Ohta, 1999: supraliminal presentation) but not all studies (Foa, Ilai, McCarthy, Shoyer, & Murdock, 1993; Lavy, van Oppen, & van den Hout, 1994; Unoki et al., 1999: subliminal presentation) has been interpreted (Summerfeldt & Endler, 1998) as a further evidence to segregate OCD from the anxiety disorder spectrum besides differences regarding phenomenology and treatment. Alternatively, problems to detect an attentional bias in patients with OCD may be rooted in the idiosyncratic nature of OCD: Stimuli
* Corresponding author. Tel.: þ49 40 42803 6565; fax: þ49 40 42803 7566. E-mail address: [email protected] (S. Moritz). 0005-7967/$ – see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.brat.2008.05.005
eliciting concern or arousal largely differ even among patients who share the same subtype (i.e., washing, checking, ordering, hoarding). While one checker may only be concerned when leaving his house (i.e., worry that the door has not been properly locked), another patient may only worry while driving (i.e., fear to have hit someone over). In contrast, in phobias, for example spider phobia, the set of fear-eliciting stimuli is more homogeneous. Moreover, not all prior relevant studies have looked at OCD-relevant stimuli but some administered general anxiety, panic, or depression stimuli (McNally et al., 1992, 1994; McNeil et al., 1999; Moritz, Jacobsen et al., 2004). As many patients do not experience fear/ panic but disgust, tension, ‘‘not just right’’ experiences (Coles, Heimberg, Frost, & Steketee, 2005) or an unspecified urge preceding compulsions, such material may not be best suited to capture an attentional bias. A prominent test to elicit an attentional bias is the emotional Stroop paradigm (MacLeod,1991). Whereas in a conventional Stroop task (Stroop, 1935), the print-colour of an incongruently typed colour word has to be named resulting in strong interference (i.e., reading is more automatized than the target response colour naming), the emotional Stroop uses disorder-related words. As personally or disorder-relevant items are attention-grabbing, the subject’s attention is diverted from the primary task towards the distractor, thus delaying the response. As an alternative account, rumination or a lowered threshold for disorder-relevant material has been
1102
S. Moritz et al. / Behaviour Research and Therapy 46 (2008) 1101–1104
postulated to explain this response pattern in mental disorders. Response latencies in the experimental condition are compared to a control condition (neutral stimuli, colour bars, or meaningless strings of characters). Usually, the emotional Stroop effect is weaker than interference effects elicited by colour words. For the present study, we compared OCD and healthy subjects on an emotional variant of the Stroop task involving OCD-relevant stimuli. We looked separately at items relevant for the two most prevalent OCD subtypes (checking and washing). This was done, as an attentional bias may not operate for all subtypes alike and to take into account the heterogeneous phenomenology of OCD (e.g., the word ‘‘dirt’’ is likely more arousing for washers than checkers). Stimuli were chosen via expert ratings to cover the most prevalent concerns for each subtype. Apart from subtype other prominent moderators for group differences in OCD apart from subtypes, such as severity of OCD as well as comorbid affective symptoms, were carefully considered. Finally, in view of recent studies (McKenna & Sharma, 2004; for a discussion see Phaf & Kan, 2007) that have detected slow but not fast emotional Stroop interference, we have looked at reaction times both for immediate (trial N) and subsequent trials (N þ 1). To the best of our knowledge, it has not been investigated whether interference effects need more time to evolve in OCD patients. Methods Participants Twenty-three patients meeting DSM-IV criteria for OCD were recruited from the Department of Psychiatry and Psychotherapy of the University Medical Center Hamburg/Germany (gender: 11 male, 12 female; age: M ¼ 35.57, SD ¼ 10.54; years of formal school education: M ¼ 11.04, SD ¼ 1.69). Presence of OCD was confirmed with the Mini International Neuropsychiatric Interview (MINI, Sheehan et al., 1998). Any history of psychotic and manic symptoms (grandiose or paranoid ideas, hallucinations of any modality) led to exclusion. Most patients were medicated with psychotropic drugs (15 patients were prescribed antidepressant drugs, six of these also with a neuroleptic agent). Eleven patients had a depressive episode secondary to OCD and eight patients were co-diagnosed with another anxiety disorder. Twenty-three subjects served as healthy controls who were recruited via an established subject pool or word-of-mouth (gender: 12 male, 11 female; age: M ¼ 31.39, SD ¼ 10.66; years of formal school education: M ¼ 11.74, SD ¼ 1.48). The MINI interview verified absence of any axis I psychiatric disorder in controls. None of the participants suffered from neurological disorders (e.g., stroke, epilepsy) including OCD spectrum disorders (e.g., Tourette’s syndrome). Written informed consent was obtained from all participants prior to baseline assessment. The German version of the YaleBrown Obsessive–Compulsive Scale (Y-BOCS) (Goodman et al., 1989; German translation by Hand & Bu¨ttner-Westphal, 1991) served as an index of OCD severity (total: M ¼ 22.73; SD ¼ 6.61). For assessing depression, the Hamilton depression rating scale (HDRS; 17 item version; Hamilton, 1960) was administered blind to neurocognitive status (total score: M ¼ 12.18, SD ¼ 7.08). Subscores for these scales were computed using algorithms derived from factor analytic studies (Moritz et al., 2002; Moritz, Meier, Hand, Schick, & Jahn, 2004). We administered the Hamburg Obsessional Compulsive Inventory (HOCI; Klepsch, Zaworka, Hand, Lu¨nenschloss, & Jauernig, 1991) to specify OCD subtypes. The HOCI is a self-rating instrument that assesses core obsessions (e.g., thoughts of doing harm to self/others) and compulsions (e.g., checking) along six scales. Seven of the patients displayed both checking and washing symptoms, while each three patients only showed checking or washing symptoms.
Experimental task The emotional Stroop task was constructed using SuperLabÒ software and was individually presented via a Macintosh computer. Following a short practice task with six items to familiarize subjects with the task requirements, 15 stimuli from each of the nine different conditions were presented in random order. Words for the non-OCD conditions were similar to stimuli used in a prior experiment (Moritz, Jacobsen et al., 2004) and compiled following consensus ratings by experienced clinicians. The OCD items were newly collected after an iterative reduction process guided by ratings from six clinical psychologists or psychiatrists with extensive experience with OCD patients. The final set of words was rated OCD-relevant by the majority of assessors. The conditions were similar according to word length, initial letters (plosive vs. non-plosive) and word frequency (all contrasts p > 0.1). The emotional word conditions were as follows: OCD-checking relevant (e.g., lock, accident), OCD-washing relevant (e.g., blood, dirt), anxiety (e.g., anxiety, panic), depression (e.g., weakness, loneliness), positive (e.g., beauty, success), and paranoia-relevant (e.g., spy, surveillance). There were two non-affective control conditions: neutral (e.g., bag, table) and Stroop interference (e.g., incongruently written colour words). Naming of horizontal colour bars served as baseline condition. As noted, the non-OCD conditions were similar to a previous experiment with one exception: in the neutral condition, two formerly presented items that could be considered OCDrelevant (soap, sink) were substituted for the present experiment. Colors were evenly distributed across all conditions: red (Ger¨ n), yellow (gelb), and blue (blau). Words were man: rot), green (gru presented in German language; all participants were German native speakers. Each trial consisted of the following sequence: a small fixation point (300 ms) was first followed by a 200 ms blank trial. Subsequently, the target stimulus (font 30, type Geneva) was shown until the individual triggered the voice-key via microphone. Participants were instructed to respond as fast as possible and to avoid making mistakes. Following each response, the experimenter pressed a button to denote correct (i.e., ‘‘C’’) and incorrect responses (i.e., ‘‘X’’) which initiated the next trial. Strategy of data analysis Reaction time latencies were recorded for correct trials only. Reaction times faster than 300 ms and slower than 4000 ms were excluded from analyses. As dependent variables, we computed reaction times from the median response latencies for target words (N) as well as subsequent words (trial N þ 1), as we were interested if an attentional bias persists over time and/or manifests after a delay due to, for example, rumination. A 2 9 mixed analysis of variance (ANOVA) was conducted with the experimental conditions as within-subject and Group (healthy, OCD) OCD as betweensubject factor. Since our hypothesis was undirected, for exploratory purposes we also computed single comparisons. Results Sociodemographic background variables and immediate interference Samples did not differ with respect to sociodemographic background variables (age, gender, school education; all p > .1). In the mixed ANOVA neither the effect of Group, F(1,44) ¼ 0.84, p > .3, h2partial ¼ .02, nor the interaction of Group Condition, F(8,352) ¼ 0.78, p > .3, h2partial ¼ .02, surpassed the level of significance. As can be seen in Fig. 1, the significant effect of Condition is mainly owing to discrepant reaction times in the control condition
S. Moritz et al. / Behaviour Research and Therapy 46 (2008) 1101–1104
reaction times in ms
850 healthy OCD
800 750 700 650 600
ty
na colo m r in g St ro op
xi e
n pr e de
an
ss io
ve si ti po
no id
l ut ra
ra pa
ne
as hi w
ch e
ck in
g
ng
550
Fig. 1. The two-way ANOVA did not yield any group differences regarding condition. However, post hoc analyses revealed a slight speeding of OCD patients for washingrelevant words in the colour bar naming condition but not in the neutral condition relative to controls.
colour naming versus the conventional Stroop interference condition, F(8,352) ¼ 28.50, p < .001, h2partial ¼ 0.39. When we subtracted reaction times in the two OCD and the Stroop interference condition from the neutral word and non-word control condition (colour naming), no evidence for greater interference emerged for OCD patients even before Bonferroni-correction for multiple comparisons. On the contrary, for washing-relevant material, the difference achieved significance (washing vs. colour bars, p ¼ .03) or trend level (washing vs. neutral word, p ¼ .09) indicating facilitation in patients. For checking-relevant items (p > .3) and conventional Stroop interference (p > .1) no such effects emerged. Delayed interference For trials following the critical trials (N þ 1) groups displayed a similar response pattern: the main effect of Group and the interaction did not produce significant effects (both p > 0.3, both h2partial < 0.028). Moreover, none of the post hoc comparisons (e.g., washing condition – neutral) yielded significance (p > 0.3). Subgroup analysis According to HOCI scores, the OCD sample was split into washers and checkers. None of the ANOVA results yielded significance except for the washing condition with colour naming as baseline, F(2,41) ¼ 3.34, p < .05. Post hoc comparisons showed that washers (M ¼ 48 ms) showed significantly less interference than healthy participants (M ¼ 124 ms). OCD-non-washers performed intermediate (M ¼ 92 ms). Finally, we investigated whether a subgroup of patients would display strong interference. To our surprise, for washing-relevant material six healthy subjects were among the 10 participants with the highest interference latencies, whereas for checking-relevant items both five healthy and OCD patients displayed highest interference. Discussion The present study revealed no evidence for greater interference on any of the experimental conditions of an emotional Stroop task in OCD. Specifically, no retardation was detected for OCD-relevant material – irrespective of subtype. The OCD stimuli were collected after a thorough iterative reduction process involving experts’ opinion to assure that central OCD concerns were covered. Moreover, special care was devoted to equate conditions for several important moderators such as words length and language frequency. Our findings unlikely reflect a lack of statistical power, as
1103
group differences were rather small and for the washing-related words even a counter-intuitive facilitation occurred: here, especially washers, displayed less interference! While it is thus tempting to infer that OCD is not associated with an attentional bias, particularly in view of many studies unable to detect such a cognitive preference in OCD (see introduction) and in line with claims made by Summerfeldt and Endler (1998), we should not jump to conclusions. First, although we gathered stimuli that deal with the prominent concerns of OCD washers and checkers, we cannot rule out that for a subgroup of patients sharing idiosyncratic and very isolated worries these items were not attention-grabbing enough thus attenuating group differences. The counter-intuitive facilitation for washing-related material could be owed to greater familiarity in patients with such stimuli acting against a pop-out effect of the distractor, which may only be found for the most fear-evoking stimuli. We would like to point out that Kyrios and Iob (1998) also found facilitation for OCD and other anxiety words in both OCD and healthy subjects. Interestingly, Lavy et al. (1994) reported greater interference for negatively valenced OCD words while for positively valenced OCD words the response pattern reversed in OCD patients. Nonetheless, future studies may benefit from greater consideration of individual concerns and may also employ paradigms that may tease apart the relative contribution of vigilance and disengagement problems in the emergence of Stroop interference (Amir, Elias, Klumpp, & Przeworski, 2003). If the attentional bias in OCD is existent, it is apparently more subtle than in PTSD patients or patients with simple phobias. Therefore, a verbal paradigm may not be sufficiently sensitive to elicit a heightened attentional bias, especially when involving random item administration precluding prolonged confrontation with disorder-related items. The worries of OCD patients are mostly triggered by visual cues or images (e.g., dirt on the table, key lock) and a visual paradigm may thus be fairer to put this hypothesis to test. Indeed, a recent study using neutral, anxiety and OCD pictures in the context of a modified inhibition of return (IOR) paradigm was able to show greater distractibility of OCD patients for OCD-relevant but not neutral and anxiety material relative to controls (Moritz et al., submitted for publication). Therefore, while we are safe to assume that any attentional bias in OCD is less pronounced than in other anxiety disorders, especially PTSD, researchers should continue to pursue this account with other paradigms before ultimately dismissing it. Moreover, subjective appraisal ratings may help to determine whether stimuli sufficiently cover patients’ concerns. In line with this, we found that washers but not checkers were slowed on a primed Stroop task for OCD-relevant targets when the sample was split in subjects endorsing more than one-third of the items as personally relevant versus those that did not meet this criterion (Fischer, 2007). So far, stimuli of past research were either rated by other subjects for relevance (Foa et al., 1993; Lavy et al., 1994; Unoki et al., 1999), experts (Moritz, Jacobsen et al., 2004, and present study) or not at all. None of these studies performed a subanalysis of items with special personal relevance. Finally, the blockwise administration of conditions may be more powerful to elicit effects (e.g., Lavy et al., 1994) if an attentional bias in OCD needs time to evolve and/or needs multiple triggers and prolonged exhibition (see Phaf & Kan, 2007), although the present analysis of delayed effects in trial N þ 1 did not provide evidence for this assumption.
References Amir, N., Elias, J., Klumpp, H., & Przeworski, A. (2003). Attentional bias to threat in social phobia: facilitated processing of threat or difficulty disengaging attention from threat? Behaviour Research and Therapy, 41, 1325–1335. Coles, M. E., Heimberg, R. G., Frost, R. O., & Steketee, G. (2005). Not just right experiences and obsessive–compulsive features: experimental and selfmonitoring perspectives. Behaviour Research and Therapy, 43, 153–167.
1104
S. Moritz et al. / Behaviour Research and Therapy 46 (2008) 1101–1104
Fischer, B.-K. (2007). Untersuchung kognitiver Einbahnstraßen bei Zwang unter Verwendung einer semantischen Priming-Aufgabe mit sto¨rungsrelevantem Material [Investigation of cognitive ‘‘one-way streets’’ in obsessive–compulsive disorder using a semantic priming-task with disorder-relevant material]. Master thesis, University of Hamburg, Hamburg. Foa, E. B., Ilai, D., McCarthy, P. R., Shoyer, B., & Murdock, T. (1993). Information processing in obsessive–compulsive disorder. Cognitive Therapy and Research, 17, 173–189. Goodman, W. K., Price, L. H., Rasmussen, S. A., Mazure, C., Fleischmann, R. L., Hill, C. L., et al. (1989). The Yale-Brown obsessive compulsive scale. I. Development, use, and reliability. Archives of General Psychiatry, 46, 1006–1011. Hamilton, M. (1960). A rating scale for depression. Journal of Neurology, Neurosurgery and Psychiatry, 23, 56–62. Hand, I., & Bu¨ttner-Westphal, H. (1991). Die Yale-Brown obsessive compulsive scale (Y-BOCS): Ein halbstrukturiertes Interview zur Beurteilung des Schweregrades von Denk- und Handlungszwa¨ngen. Verhaltenstherapie, 1, 223–225. Kampman, M., Keijsers, G. P., Verbraak, M. J., Naring, G., & Hoogduin, C. A. (2002). The emotional Stroop: a comparison of panic disorder patients, obsessive– compulsive patients, and normal controls, in two experiments. Journal of Anxiety Disorders, 16, 425–441. Klepsch, R., Zaworka, W., Hand, I., Lu¨nenschloss, K., & Jauernig, G. (1991). Derivation and validation of the Hamburg obsession/compulsion inventory-short form (HOCI-S): first results. Psychological Assessment, 3, 196–201. Kyrios, M., & Iob, M. A. (1998). Automatic and strategic processing in obsessive– compulsive disorder: attentional bias, cognitive avoidance or more complex phenomena? Journal of Anxiety Disorders, 12, 271–292. Lavy, E., van Oppen, P., & van den Hout, M. (1994). Selective processing of emotional information in obsessive compulsive disorder. Behaviour Research and Therapy, 32, 243–246. MacLeod, C. M. (1991). Half a century of research on the Stroop effect: an integrative review. Psychological Bulletin, 109, 163–203. Mathews, A., & MacLeod, C. (2005). Cognitive vulnerability to emotional disorders. Annual Review of Clinical Psychology, 1, 167–195. McKenna, F. P., & Sharma, D. (2004). Reversing the emotional Stroop effect reveals that it is not what it seems: the role of fast and slow components. Journal of Experimental Psychology: Learning, Memory, and Cognition, 30, 382–392. McNally, R. J., Amir, N., Louro, C. E., Lukach, B. M., Riemann, B. C., & Calamari, J. E. (1994). Cognitive processing of idiographic emotional information in panic disorder. Behaviour Research and Therapy, 32, 119–122.
McNally, R. J., Riemann, B. C., Louro, C. E., Lukach, B. M., & Kim, E. (1992). Cognitive processing of emotional information in panic disorder. Behaviour Research Therapy, 30, 143–149. McNeil, D. W., Tucker, P., Miranda, R., Jr., Lewin, M. R., & Nordgren, J. C. (1999). Response to depression and anxiety Stroop stimuli in posttraumatic stress disorder, obsessive–compulsive disorder, and major depressive disorder. Journal of Nervous and Mental Disease, 187, 512–516. Moritz, S., Jacobsen, D., Kloss, M., Fricke, S., Rufer, M., & Hand, I. (2004). Examination of emotional Stroop interference in obsessive–compulsive disorder. Behaviour Research and Therapy, 42, 671–682. Moritz, S., Kloss, M., Jacobsen, D., Wein, C., Fricke, S., & Hand, I. (2002). Dimensional structure of the Yale-Brown obsessive compulsive scale (Y-BOCS). Psychiatry Research, 109, 193–199. Moritz, S., Meier, B., Hand, I., Schick, M., & Jahn, H. (2004). Dimensional structure of the Hamilton depression rating scale in patients with obsessive–compulsive disorder. Psychiatry Research, 125, 171–180. Moritz, S., Mu¨hlenen, A. v., Jelinek, L., Randjbar, S., Ruhe, C., Fischer, B.-K., et al. Evidence for an attentional bias for washing and checking relevant stimuli in obsessive–compulsive disorder, submitted for publication. Moritz, S., & von Muhlenen, A. (2005). Inhibition of return in patients with obsessive–compulsive disorder. Journal of Anxiety Disorders, 19, 117–126. Moritz, S., & von Mu¨hlenen, A. (2008). Investigation of an attentional bias for fear-related material in obsessive–compulsive checkers. Depression and Anxiety, 25, 225–229. Phaf, R. H., & Kan, K. J. (2007). The automaticity of emotional Stroop: a metaanalysis. Journal of Behavior Therapy and Experimental Psychiatry, 38, 184–199. Sheehan, D. V., Lecrubier, Y., Sheehan, K. H., Amorim, P., Janavs, J., Weiller, E., et al. (1998). The MINI International Neuropsychiatric Interview (M.I.N.I.): The development and validation of a structured diagnostic psychiatric interview. Journal of Clinical Psychiatry, 59(Suppl. 20), 22–33. Stroop, J. R. (1935). Studies of interference in serial verbal reactions. Journal of Experimental Psychology, 18, 643–662. Summerfeldt, L. J., & Endler, N. S. (1998). Examining the evidence for anxietyrelated cognitive biases in obsessive–compulsive disorder. Journal of Anxiety Disorders, 12, 579–598. Unoki, K., Kasuga, T., Matsushima, E., & Ohta, K. (1999). Attentional processing of emotional information in obsessive–compulsive disorder. Psychiatry and Clinical Neuroscience, 53, 635–642. Williams, J. M. G., Meathews, A., & MacLeod, C. (1996). The emotional Stroop task and psychopathology. Psychological Bulletin, 120, 3–24.