Uncertainty during Anticipation Modulates Neural Responses to Aversion in Human Insula and Amygdala

Cerebral Cortex April 2010;20:929--940 doi:10.1093/cercor/bhp155 Advance Access publication August 13, 2009 Uncertainty during Anticipation Modulates Neural Responses to Aversion in Human Insula and Amygdala I. Sarinopoulos1, D. W. Grupe2, K. L. Mackiewicz3, J. D. Herrington4, M. Lor5, E. E. Steege6 and J. B. Nitschke2 Uncertainty about potential negative future outcomes can cause stress and is a central feature of anxiety disorders. The stress and anxiety associated with uncertain situations may lead individuals to overestimate the frequency with which uncertain cues are followed by negative outcomes, an example of covariation bias. Using functional magnetic resonance imaging, we found that uncertaintyrelated expectations modulated neural responses to aversion. Insula and amygdala responses to aversive pictures were larger after an uncertain cue (that preceded aversive or neutral pictures) than a certain cue (that always preceded aversive pictures). Anticipatory anterior cingulate cortex (ACC) activity elicited by the cues was inversely associated with the insula and amygdala responses to aversive pictures following the cues. Nearly 75% of subjects overestimated the frequency of aversive pictures following uncertain cues, and ACC and insula activity predicted this uncertainty-related covariation bias. Findings provide the first evidence of the brain mechanisms of covariation bias and highlight the temporal dynamics of ACC, insula, and amygdala recruitment for processing aversion in the context of uncertainty. investigated as it relates to phobias and other anxiety disorders (Tomarken et al. 1989; de Jong et al. 1992; Pauli et al. 1996, 1998; Amin and Lovibond 1997; Kennedy et al. 1997; Hermann et al. 2004). Biased estimates of covariance have also been identified in individuals without anxiety symptoms, given an adequately salient or fear-relevant cue (Tomarken et al. 1989; Pury and Mineka 1997; de Jong et al. 1998). Uncertainty about potential aversive stimuli is a fear-relevant cue (Freeston et al. 1994; Barlow 2002; Buhr and Dugas 2002) that would be expected to result in overestimates of the relationship between uncertain cues and aversive outcomes. Previous research has demonstrated that the anterior cingulate cortex (ACC), insula, and amygdala are recruited under conditions of uncertainty (Critchley et al. 2001; Davis and Whalen 2001; Volz et al. 2003; Hsu et al. 2005; Paulus 2005; Grinband et al. 2006; Krain et al. 2006; Rosen and Donley 2006; Belova et al. 2007; Dunsmoor et al. 2007; Hasler et al. 2007; Herry et al. 2007; Platt and Huettel 2008; Preuschoff et al. 2008). Importantly, these same broad regions have also been shown to be activated by a wide range of aversive stimuli and conditioning paradigms (Büchel et al. 1998, 1999; Ploghaus et al. 1999; Davis and Whalen 2001; Craig 2002, 2003; LeDoux 2002; Han et al. 2003; Phillips et al. 2003; Mackiewicz et al. 2006; Nitschke, Dixon, et al. 2006; Nitschke, Sarinopoulos, et al. 2006; Bissière et al. 2008). Uncertainty about the likelihood of an aversive event may serve to enhance responses to aversion when such events do occur, as illustrated by studies showing that aversive events are more stressful when associated with or preceded by uncertainty than certainty (Peeke and Grings 1968; Grings and Schell 1971; Lykken et al. 1972; Craske et al. 1995; Nader and Balleine 2007). Although further attention is needed to identify the specific regions within the ACC, insula, and amygdala showing overlap in the literatures on uncertainty and on aversion, these 3 brain areas are promising candidates for a modulatory neural network that can account for such uncertainty-enhanced responses to aversion. Indeed, a recent study found greater amygdala responses to an aversive noise (unconditioned stimulus) that was paired with a conditioned stimulus on 50% of trials than to the same noise paired with a conditioned stimulus on all trials (Dunsmoor et al. 2008). The current study investigated the temporal unfolding of activity in the ACC, insula, and amygdala that transpires between the introduction of uncertainty and subsequent aversive stimuli. Modifying a paradigm previously shown to activate the ACC, insula, and amygdala in anticipation of and response to aversive pictures (Mackiewicz et al. 2006; Introduction Knowledge about upcoming adverse circumstances can be helpful in terms of preparing for and potentially avoiding such events. However, there is often uncertainty about whether an upcoming aversive event will actually occur and how dangerous or negative it will be. Such uncertainty is central to worry and negative expectations about future events that can be debilitating in individuals suffering from anxiety disorders (Dugas et al. 1998; Barlow 2002; Lohr et al. 2007; Krain et al. 2008; Simmons et al. 2008). The resolution of uncertainty can be achieved through the detection of contingencies between environmental cues and subsequent aversive events. Such contingency detection allows individuals to explain past events and more appropriately prepare for the future (Alloy and Tabachnik 1984). This process, however, is subject to errors involved in the perception and interpretation of contingencies, and can result in ‘‘illusory correlations,’’ or the identification of relationships between cues and subsequent outcomes that in reality are not related (Chapman and Chapman 1967, 1969). Illusory correlation paradigms have been used to identify overestimates of the covariance of fear-relevant cues and subsequent aversive outcomes, known as covariation biases (Tomarken et al. 1989). The covariation bias has primarily been Ó The Author 2009. Published by Oxford University Press. All rights reserved. For permissions, please e-mail: Departments of Psychology and Radiology, Michigan State University, East Lansing, MI 48824, USA, 2Waisman Laboratory for Brain Imaging and Behavior, Departments of Psychiatry and Psychology, University of Wisconsin, Madison, WI 53705, USA, 3 Department of Psychology, University of Colorado, Boulder, CO 80309, USA, 4Center for Autism Research, Department of Pediatrics, Children’s Hospital of Philadelphia, Philadelphia, PA 19104, USA, 5Columbia University School of Social Work, New York City, NY 10027, USA and 6College of Physical Education and Sport Science, West Virginia University, Morgantown, WV 26506, USA Keywords: anterior cingulate cortex, covariation bias, emotion, expectancy, fMRI 1 Materials and Methods Subjects Forty right-handed healthy undergraduate students (18 women and 22 men; age M = 20.65, SD = 1.53) who responded to flyers posted in University of Wisconsin-Madison buildings participated in the study. Subjects reported no medical, neurological, or psychiatric problems and took no medications. Four subjects were dropped from analyses—2 women due to technical difficulties with functional magnetic resonance imaging (fMRI) data acquisition, one man due to excessive movement during fM (...truncated)