Fear conditioning in posttraumatic stress disorder: Evidence for delayed extinction of autonomic, experiential, and behavioural responses
Introduction
Posttraumatic stress disorder (PTSD) is a pervasive psychiatric condition characterised, inter alia, by symptoms of persistent re-experiencing of the traumatic event (DSM-IV, American Psychiatric Association, 1994). Contemporary theories of PTSD concur in assuming that memory and learning processes like perceptual priming and fear conditioning underlie these re-experiencing symptoms (Michael et al., 2005; Pitman, 1989; Rothbaum & Davis, 2003). According to the fear conditioning approach, the traumatic event (unconditioned stimulus, US) triggers an unconditioned response (UR) which is characterised by strong arousal and intense fear. This UR becomes associated with cues, such as smells, voices, or sights (conditioned stimuli, CSs) which were present during the traumatic event. As a result of this pairing, these cues can trigger similar responses (conditioned responses, CRs) even in the absence of the US. Thus, re-experiencing symptoms can be understood as CRs, which remain persistent, even a long time after the trauma.
As such, the fear conditioning account cannot explain why these symptoms disappear in the aftermath of a traumatic event in most individuals, but persist in those who develop PTSD. Within the conditioning framework, three accounts have been put forward to answer this question: enhanced conditionability, reduced conditioned inhibition, and reduced discrimination learning.
The concept of enhanced conditionability refers to a hypothetical trait predisposing to the development of stronger CRs to a traumatic event, and/or to a reduced ability to extinguish these CRs (Orr et al., 2000). Experimentally, conditionability is typically assessed in a differential fear conditioning paradigm in which one CS is paired with the US during the acquisition phase (the ) and another CS is not (the ). During a subsequent extinction phase, both CSs are presented without the US. The difference between reactions to the and the (also called differential or discriminative learning) during acquisition and/or extinction indexes conditionability. An individual high on conditionability is thought to be at risk for the development of PTSD subsequent to trauma exposure since particularly strong CRs develop and persist.
However, conditionability, as assessed by differential fear conditioning, actually confounds two processes: excitatory conditioning and inhibitory conditioning (assessed by responses to the and the , respectively) which each may be informative in its own right (Lissek et al., 2005). In fact, it has been suggested that the inability to inhibit fear in the presence of safety cues (i.e. the ) causes excessive fear responses in PTSD patients (Davis, Falls, & Gewirtz, 2000; Grillon & Morgan, 1999; Rothbaum & Davis, 2003). Thus, it is proposed that PTSD patients should differ from controls mainly because of poor inhibitory processes, i.e. they should show heightened responding to the . In the following we will refer to this account as conditioned inhibition account.1
In line with the enhanced conditionability account, Orr and coworkers found stronger differential responding during acquisition and extinction in PTSD in comparison to trauma exposed controls (Orr et al., 2000). Similarly, Peri and colleagues found enhanced differential responses during the extinction phase in PTSD patients in contrast to healthy and traumatised controls (Peri, Ben-Shakhar, Orr, & Shalev, 2000). However, Peri and colleagues also found heightened reactions in PTSD with respect to the during acquisition and extinction. Although these two studies interpreted their findings to support enhanced conditionability in PTSD, they are also partially consistent with the conditioned inhibition account.
In addition to these two accounts, a third conceptualisation of reduced discrimination learning has received support in the clinical conditioning literature. Investigating eye blink conditioning2 in combat veterans with and without PTSD and control participants, Ayers, White, and Powell (2003) found differential responding to the CSs only in control participants. They attributed this to impaired discriminative learning in combat veterans, possibly due to general memory deficits. Grillon and Morgan (1999) measured the fear potentiated startle reactions in a differential fear conditioning paradigm in two separate sessions separated by one week. In contrast to trauma exposed controls, PTSD patients failed to acquire differential conditioning during the first session. Only during the second conditioning session did they show differential startle responses similar to controls. In opposition to the enhanced conditionability account, the reduced discriminative learning account conceptualises this type of learning as a highly functional process by which participants learn to distinguish between threat and safety cues (Grillon, 2002a).
At this stage, research has yielded partial support for the enhanced conditionability account of PTSD. While some studies were supportive of this view (Orr et al., 2000; Peri et al., 2000) at least two others found equal (Orr et al., 2006; Vythilingam et al., 2006) or impaired discrimination learning (Ayers et al., 2003; Grillon & Morgan, 1999) in PTSD patients. This heterogeneity of findings clearly warrants further investigation. In addition, the conditioned inhibition account, predicting enhanced responding in PTSD patients to the , has not been explicitly addressed in previous fear conditioning studies of PTSD.
Previous conditioning studies in PTSD have focused primarily on implicit indicators of conditioning, such as skin conductance responses (SCRs) or the fear potentiated startle. However, this focus on implicit measurements unnecessarily confines the window of scientific inquiry and disregards the domains of verbal-cognitive and behavioural responses. Contemporary conditioning models highlight the role of cognitive processes (Davey, 1997; Lovibond, 2006). According to the expectancy model of fear conditioning (Lovibond, 2006; Reiss, 1991) individuals continuously and explicitly adjust their expectancies regarding the likelihood of the US when the and the are repeatedly presented. A growing number of studies have successfully included continuous “online” measures of US-expectancy (e.g., Lovibond, Davis, & O’Flaherty, 2000; Neumann, Lipp, & Cory, 2007). Another important process involved in human conditioning relates to conditioned changes in affective valence appraisals of the CSs, a process called evaluative conditioning (for review, see De Houwer, Thomas, & Baeyens, 2001). According to this theory, affective valence is transferred from the US to the CS as a result of paired presentations during conditioning.
In this study we examined differential fear conditioning in PTSD patients using a more comprehensive set of dependent measure which assessed autonomic (SCRs), affective (valence ratings), and cognitive (US-expectancy ratings) responses. As a subsidiary aim we explored if conditioned responding also generalises to the behavioural domain using a behavioural forced choice test (Michael, Blechert, & Vriends, unpublished data). In order to maximise the conclusiveness of between-group comparisons, we included two healthy control groups, one group with trauma exposure (TE group) and one group with no-trauma exposure (nTE group). Statistical analyses assessed differential conditioning (differences in responses to both CSs) but also included single CS analyses to evaluate the accounts of heightened conditionability, conditioned inhibition, and reduced discrimination learning. The enhanced conditionability account would predict larger differential reactions in the PTSD group compared to the other two groups while the account of reduced discrimination learning would predict smaller differential responses. The latter account would also predict impaired explicit learning of the CS–US contingency in the PTSD group. The conditioned inhibition account would predict enhanced responding to the in the PTSD group.
Section snippets
Participants
We recruited three study groups: the PTSD group consisted of 36 adults qualifying for a primary diagnosis of current chronic PTSD according to the DSM-IV (American Psychiatric Association, 1994), the TE group consisted of individuals who had been exposed to a traumatic event without developing PTSD (), and the nTE group consisted of healthy individuals, who had never been exposed a traumatic event (). Participants were included into the TE group if they fulfilled the A-criterion of the
Demographics, psychometrics and control variables
Table 1 shows demographic, psychometric, and control measures for the three groups. Groups did not differ in age and years of education. In accordance with the diagnostic categorisation, the PTSD group scored higher than the control groups on the PDS, DES, STAI, and the BDI. PTSD patients tended to select a lower US level than both control groups, but subjective ratings of US intensity did not differ between groups. PTSD patients also showed higher SCL during habituation and a trend to higher
Discussion
Test of theoretical accounts: Consistent with the enhanced conditionability account, we found stronger electrodermal differential conditioning during late acquisition and slowed extinction of responding to the in the PTSD group compared to the nTE group. This corresponds well with findings of Orr et al. (2000) and Peri et al. (2000) who both found delayed extinction of SCRs to the . In the PTSD vs. nTE group comparison, however, group differences were less convincing: only during the
Acknowledgements
This research was supported by Grants 105311-104038 and 105311-105850 from the Swiss National Science Foundation and by a grant from the Academic Society of Basel (Freiwillige Akademische Gesellschaft Basel). The authors would like to thank Marta Lajtman and Rebecca Frey for their help with the data collection and Shmuel Lissek for helpful comments.
References (58)
- et al.
Acquired affective evaluative value: Conservative but not unchangeable
Behaviour Research and Therapy
(1989) Context and ambiguity in the extinction of emotional learning: Implications for exposure therapy
Behaviour Research and Therapy
(1988)- et al.
Information processing and PTSD: A review of the empirical literature
Clinical Psychology Review
(2000) - et al.
Memory biases in the anxiety disorders: Current status
Clinical Psychology Review
(2002) - et al.
Counterconditioning in the treatment of spider phobia: effects on disgust, fear and valence
Behaviour Research and Therapy
(2000) - et al.
A cognitive model of posttraumatic stress disorder
Behaviour Research and Therapy
(2000) - et al.
Internal consistency and temporal stability of classically conditioned skin conductance responses
Biological Psychology
(1993) Associative learning deficits increase symptoms of anxiety in humans
Biological Psychiatry
(2002)Startle reactivity and anxiety disorders: Aversive conditioning, context, and neurobiology
Biological Psychiatry
(2002)- et al.
Reinstatement of fear responses in human aversive conditioning
Behaviour Research and Therapy
(2005)