The Obsessive-Compulsive Style of Visual Attention

It has been suggested that individuals with obsessive-compulsive personalities tend to focus on small local details in their surroundings, whereas histrionic individuals are characterized by more global information processing. Using the global-local hierarchical-letters paradigm, we were able to provide support for the first but not the second hypothesis. Measures related to obsessive-compulsive personality disorder were associated with excessive visual attention to small details of the hierarchical letters. Specifically, the obsessive-compulsive cognitive style was associated with local interference, which reflects the effects of distraction by to-be-ignored small details on identification of global information. A fundamental assumption of cognitive approaches to personality and psychopathology is that individuals differ in their response to objectively similar situations because of differences in the way they process those situations. These differences can exist at many stages of cognitive processing, ranging from lowerlevel processes such as the initial detection of threat (e.g., Öhman, Lundqvist, & Esteves, 2001), to higher-level mental representations such as broad schemata of the self and of the world (e.g., Beck, 1976). Sometimes these individual differences in lower-level information processing occur primarily in the processing of affectively valenced information (for reviews, see Mineka, Rafaeli, & Yovel, 2003; Williams, Watts, MacLeod, & Mathews, 1997). However, personality also has important effects on lower-level cognitive processing of non-affectively valenced information (e.g., Witkin & Goodenough, 1977), although this topic has been less frequently studied (but see Granholm, Cadenhead, Shafer, & Filoteo, 2002). According to many theories of personality and cognition (e.g., Mischel & Shoda, 1999), individuals’ characteristic ways of perceiving and thinking mediate their cognitive, emotional, and social functioning; accordingly, the way individuals perform on certain cognitive tasks is associated with broader characteristics of their personality. Several early such theories (e.g., Gardner, Holzman, Klein, Linton, & Spence, 1959; Shapiro, 1965) were heavily influenced by psychodynamic theories of personality. For example, Shapiro (1965) described several patterns of perceiving and thinking that he argued to be typical of certain ‘‘neurotic styles’’ (e.g., ‘‘the obsessive-compulsive style,’’ ‘‘the hysterical style’’), and he suggested that these cognitive styles are the causal basis for the various cognitive, emotional, and behavioral characteristics that are associated with these pathological personality types. Shapiro (1965) argued that obsessive-compulsive individuals’ attention is active, intense, and sharply focused, particularly on small details. Consequently, these people perceive and apprehend their surroundings in a detailed and narrowed manner; for example, they ‘‘will notice a bit of dust or worry over some insignificant inaccuracy that, everything else aside, simply will not gain the attention of another person’’ (p. 27). In contrast, according to Shapiro, hysteric individuals’ cognition is global in that it is impressionistic, lacking in sharpness and details, and easily drawn to whatever is vivid and striking. Shapiro based his theory on clinical observations and on psychological tests such as the Rorschach. Later conceptualizations of related cognitive styles were similar (e.g., Witkin, Goodenough, & Oltman, 1979). Shapiro’s portrayals of hysterical and obsessive-compulsive neurotic styles roughly correspond to the contemporary histrionic and obsessive-compulsive Axis II personality disorders (HPD and OCPD, respectively) in recent versions of the American Psychiatric Association’s (APA’s) Diagnostic and Statistical Manual of Mental Disorders (DSM; APA, 1987, 1994). Indeed, according to the fourth edition of the DSM, perfectionism and preoccupation with details characterize OCPD, whereas excessively impressionistic speech that lacks in details is a criterion for HPD (APA, 1994). Address correspondence to Iftah Yovel, Department of Psychiatry, Massachusetts General Hospital and Harvard Medical School, 15 Parkman St., ACC 812, Boston, MA 02114-3117; e-mail: iyovel@ partners.org. It should be noted that OCPD and obsessive-compulsive disorder (OCD—an Axis I disorder) are separate and quite different disorders. Briefly, people with OCPD do not have true obsessions or compulsive rituals such as those seen in OCD, and the two disorders do not co-occur in a majority of cases. PSYCHOLOGICAL SCIENCE Volume 16—Number 2 123 Copyright r 2005 American Psychological Society The current study examined the basic features of attention described in Shapiro’s (1965) characterizations of the obsessive-compulsive and the histrionic cognitive styles. To do this, we used Navon’s (1977) now-classic global-local paradigm for the study of hierarchical visual attention, which seems to offer an appropriate operationalization of these cognitive characteristics (i.e., global vs. local visual attention). The basic assumption that underlies the global-local paradigm is that any visual scene can be thought of as composed of a hierarchical network of subscenes that are interrelated by spatial relationships (Kimchi, 1992). The globality of any of these subscenes depends on the place it occupies in this hierarchy. Thus, properties at higher levels are more global than those at lower levels, which are in turn more local. For example, a car has some global properties (e.g., shape, color), but it also has some local components (e.g., wheels, a license plate), each of which can be considered global in relation to its own component parts (e.g., the figures on the license plate). Navon’s (1977) global-local paradigm addresses this issue in an elegant way (for a review, see Kimchi, 1992). The stimuli in this paradigm are figures (e.g., letters) that are constructed by suitable displays of smaller figures. The two types of figures in these hierarchical stimuli are equally complicated and identifiable, and neither can be predicted by the characteristics of the other (see Fig. 1a). The bigger and smaller figures differ in their relative size, but more important, they differ in their globality, or their relative placement in the aforementioned hierarchy (Navon, 1977). Thus, the larger figure is the global level of the hierarchical stimulus, whereas the smaller figures are its local level. In the focused-attention global-local task used in the current study (based on Navon, 1977, Experiment 3), participants were asked to focus their attention on one of the levels (global or local) of each of a series of hierarchical letters and to decide on each trial which of the two target letters (H or T) was present in that prespecified level (e.g., the global), while ignoring the other level (e.g., the local). Thus, the task included two parts; for each part, only one level of the stimuli (i.e., the global or the local) was relevant. For example, in the global part of the task, the correct response for a big H made of small Ts was ‘‘H,’’ whereas in the local part, the correct response for this same stimulus was ‘‘T.’’ In each part, three types of stimuli were used (see Fig. 1a): a consistent type, in which the same target letter appeared at both the global and the local levels; an inconsistent type, in which different target letters appeared at the two levels; and a neutral type, in which the unattended level (e.g., local in the global part of the task) was not a target letter. Two well-established attentional effects are typically reported with this task (Kimchi, 1992; Fig. 1b illustrates these effects). First, participants identify the letter more rapidly when the global rather than the local level is relevant. This is the globalprecedence effect. Second, when the two levels of the hierarchical stimulus are inconsistent and the local level is relevant, a robust Stroop-like influence of the global on the local level is typically found (Kimchi, 1992). This is the global-interference effect. Both effects reflect that the global level is more salient than the local level—that the ‘‘forest’’ is more salient than the ‘‘trees.’’ Therefore, both effects seem directly related to Shapiro’s (1965) histrionic global cognitive style and inversely related to the obsessive-compulsive cognitive style (see Dickman, 1985; Maynard & Meyer, 1996). A third attentional effect that is sometimes reported in the global-local task is the local-interference effect, which is the Stroop-like effect that occurs when participants attend to the global level of the hierarchical stimuli but are distracted by the small inconsistent letters at the unattended local level. Although the magnitude of this effect depends on certain features of the task (e.g., the relative salience of the two levels; cf. Yovel, Yovel, & Levy, 2001), it is typically smaller than the magnitude of the global-precedence and global-interference effects. Fig. 1. Examples of the stimuli (a) and average response latencies (b) in the six conditions of the global-local task. The target letter was alwaysTor H. In the consistent and inconsistent conditions, the stimuli were the same for the global and local conditions. In the neutral condition, the stimuli for the global and local conditions differed, as shown. Stimuli are not drawn to scale. The three attentional effects of interest are labeled in the graph of reaction times. These indices were used in the correlational analyses examining the relation between the personality measures and attention to global and local visual information. 124 Volume 16—Number 2 The Obsessive-Compulsive Cognitive Style