The cognitive control of emotion Kevin N. Ochsner1 and James J. Gross2 1Department of Psychology, Columbia University, Schermerhorn Hall, 1190 Amsterdam Avenue, New York, NY 10027, USA 2Department of Psychology, Stanford University, Building 420, Stanford, CA 94305-2130, USA The capacity to control emotion is important for human adaptation. Questions about the neural bases of emotion regulation have recently taken on new import- ance, as functional imaging studies in humans have permitted direct investigation of control strategies that draw upon higher cognitive processes difficult to study in nonhumans. Such studies have examined (1) control- ling attention to, and (2) cognitively changing the meaning of, emotionally evocative stimuli. These two forms of emotion regulation depend upon interactions between prefrontal and cingulate control systems and cortical and subcortical emotion-generative systems. Taken together, the results suggest a functional archi- tecture for the cognitive control of emotion that dove- tails with findings from other human and nonhuman research on emotion. If you are distressed by anything external, the pain is not due to the thing itself, but to your estimate of it and this you have the power to revoke at any moment. Marcus Aurelius (Meditations) Introduction Conflicts, failures, and losses at times seem to conspire to ruin us. Yet, as Marcus Aurelius observed nearly two millennia ago, we humans have an extraordinary capacity to regulate the emotions occasioned by such travails. Importantly, these regulatory efforts largely determine the impact such difficulties will have on our mental and physical well-being [1���3]. Given its importance to adap- tive functioning, it is not surprising that research on emotion regulation has a long history (Box 1). Past work has investigated the cellular responses to stress, the behavioral consequences of adopting specific regulatory strategies, and the neural systems involved in simple forms of affective learning and social behavior in rodents and nonhuman primates [1,4���7]. In recent years, research on emotion regulation has entered a new phase as functional imaging studies of regulatory phenomena in humans have developed rapidly. This growth has facili- tated investigation of human analogs to affective beha- viors studied in animals, but, perhaps more importantly, has allowed study of the emotion regulatory power of higher cognitive control processes that are difficult to study in animal models. In so doing, current work on the ���hot��� control of emotion draws on rapidly developing cognitive neuroscience models of the ���cold��� control of attention and memory (e.g. [8,9]). The aim of this review is to evaluate recent imaging studies that, in the context of evidence from allied human and animal work, help to elucidate the functional architecture underlying the cognitive control of emotion. Emotion and emotion regulation An essential part of understanding emotion regulatory mechanisms is characterizing the processes that generate emotions. Current models posit that emotions are valenced responses to external stimuli and/or internal mental representations that (i) involve changes across multiple response systems (e.g. experiential, behavioral, peripheral physiological [10]), (ii) are distinct from moods, in that they often have identifiable objects or triggers, (iii) can be either unlearned responses to stimuli with intrinsic affective properties (e.g. an unconditioned response to an aversive shock) or learned responses to stimuli with acquired emotional value (e.g. a conditioned response or stimulus���reward association), (iv) and can involve multiple types of appraisal processes that assess the significance of stimuli to current goals [11], that (v) depend upon different neural systems [3,12,13]. Emotion regulation involves the initiation of new, or the alteration of ongoing, emotional responses through the Box 1. A brief history of psychological research on emotion regulation Study of the cognitive control of emotion has three major historical antecedents within psychology [1]. The first antecedent is the psychodynamic study of defense, which was initiated by Freud a century ago. This line of work has examined the regulation of anxiety and other negative emotions using clinical descriptions and individual difference studies of so-called perceptual defenses against processing negatively arousing stimuli, and specific defenses such as repressive coping [68,69]. The second antecedent is the stress and coping tradition that grew out of the psycho- dynamic approach in the 1960s. This line of work has focused on the management of situations that ���tax or exceed the resources of the person��� ([70], p. 141), and generated an early classic study of reappraisal showing that subjective and physiological responses decreased when a film of a potentially upsetting surgical procedure was viewed in analytical and detached terms [71]. The third antecedent is the developmental study of self-regulation, which had its roots in the study of socioemotional development. This work showed that children could obtain a preferred but delayed reward by thinking about available treats in abstract ways (e.g. putting a mental ���picture-frame��� around a cookie) that decreased their immediate impulse to eat them [72]. Contemporary research builds on this foundation using both behavioral and neuroscience methods to describe when, how, and with what consequences individuals regulate their emotions. Corresponding author: Ochsner, K.N. (ochsner@psych.columbia.edu). Available online 5 April 2005 Review TRENDS in Cognitive Sciences Vol.9 No.5 May 2005 www.sciencedirect.com 1364-6613/$ - see front matter Q 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.tics.2005.03.010

action of regulatory processes. Current work examines the processes that individuals use to influence which emotions they generate, when they do so, and how these emotions are experienced or expressed [1]. Several schemes have been proposed for organizing regulatory strategies (e.g. [14]). One distinction suggested by Gross and colleagues contrasts behavioral (e.g. suppressing expres- sive behavior) and cognitive (e.g. attending to or inter- preting emotion-eliciting situations in ways that limit emotional responding) regulation. Behavioral regulation of negative emotions might limit expressive action but does not dampen unpleasant experience, worsens memory, and increases sympathetic nervous system activation. By contrast, cognitive regulation neutralizes negative experi- ence without impairing memory and might decrease physiological arousal [15,16]. Individual differences in emotional responsivity and/or cognitive control capacity might be related to both normal and pathological variation in well-being and social behavior (Box 2). Recent imaging work has investigated two types of cognitive regulation, attentional control and cognitive change, which are the focus of this review. Figure 1 uses a hypothetical continuum to illustrate relationships between regulatory strategies tapping these two types of control. These strategies might differ in: (1) their targets ��� impacting different types of emotional appraisal processes and associated neural systems [17,18] (2) their effects ��� serving to initiate (amplify) or block (diminish) perception of our responses to stimuli (3) their relative reliance on the overlapping neural systems supporting attentional control and cognitive change, as indicated by their placement along the continuum and (4) whether emotion change is their explicit goal (���I want to feel better!���), or occurs as a by-product of pursuing some other learning or judgment-related goal (e.g. ���I want to learn which judg- ment is correct���). Attentional control Attention is often referred to as the selective aspect of information processing, enabling us to focus on goal- relevant (e.g. our writing) and ignore goal-irrelevant (e.g. loud music next door) information. In general, studies have indicated that behavioral and neural responses to attended as compared with unattended stimuli (or stimulus features) are either facilitated or inhibited, respectively (e.g. [19]). When responses to attended and unattended inputs do not differ, processing is considered to be relatively automatic. In the context of emotion, researchers have begun asking how paying less attention Box 2. From basic mechanisms to individual differences Characterizing the nature and operating characteristics of basic emotion regulatory mechanisms in healthy participants might help to establish a normative model for explaining the successful regulation of emotion. It might also lead to a greater under- standing of individual differences, clinical conditions and lifespan development, by describing them in terms of variation and change in the function of a basic functional architecture for the cognitive control of emotion. Among healthy adults, there is considerable variability in the nature and strength of emotional responses, and also in the capacity to regulate them. Behavioral studies have begun to explore the experiential and behavioral consequences of these differences [73], and characteristic patterns of resting and/or emotional stimulus- related neural activity in prefrontal and emotional appraisal systems are now being associated with gender, personality, negative affectivity [3,74] and regulatory ability. For example, Jackson et al. found that greater left PFC electrical activity at rest predicted dampened physiological reactivity to aversive stimuli, which might reflect automatic regulatory processes [75], and Ray et al. [76] found that the tendency to cognitively ruminate about emotional events pre- dicted enhanced ability to increase or decrease amygdala responses through reappraisal, which itself depends upon cognitively reexamin- ing the meaning of emotional events. Many forms of psychopathology revolve around failures to adaptively regulate emotional responses, with consequences ranging from personal distress to socially maladaptive and self-destructive behaviors [2,3,5]. Resting and symptom provocation studies have begun to identify abnormal patterns of neural response in psychiatric illness [3,6,13] and substance abuse (e.g. [77]) that might be related to emotion regulation failures. However, very few studies have examined directly the neural mechanisms mediating successful or unsuccessful regulation in clinical populations using methods like those described in this review (see, however, [78]). Building knowledge of dysregula- tory mechanisms from a basic model of effective regulation could elucidate the nature of these disorders and suggest avenues for cognitive and pharmacological treatment. Basic models of emotion regulation might also help to explain the development of regulatory capacities across the lifespan. It is possible, for example, that structural and functional changes in control and appraisal systems underlie normal and abnormal emotional responses in children [79], and the positivity of emotional experience in older adults [80]. TRENDS in Cognitive Sciences Attentional control Cognitive change Selective inattention to emotional stimuli [19���22,25] Performing distracting secondary task [31���35] Attention to and judgement of emotional vs. non-emotional stimulus attributes [23,24,26���28] Anticipatory/ expectancy- driven emotion [37���40,42���46] Top-down appraisal [17] Reappraisal [48���54] Placebo [55���57] S-R reversal/ extinction [58���65] Figure 1. Hypothetical continuum illustrating relationships among the forms of cognitive control of emotion described in this review. The left and right anchors for the continuum represent the exclusive use of attentional control or cognitive change, respectively, to modulate emotion perception and/or responses. Red and blue text denote strategies for controlled emotion generation and regulation, respectively. Relevant citations for each strategy are shown in brackets. This continuum is intended to serve a heuristic function, helping the reader to visualize relationships among control strategies (see text). Review TRENDS in Cognitive Sciences Vol.9 No.5 May 2005 243 www.sciencedirect.com