- Email: [email protected]
No evidence for blocking the return of fear by disrupting reconsolidation prior to extinction learning
c o r t e x 7 9 ( 2 0 1 6 ) 1 1 2 e1 2 2
Available online at www.sciencedirect.com
ScienceDirect Journal homepage: www.elsevier.com/locate/cortex
Research report
No evidence for blocking the return of fear by disrupting reconsolidation prior to extinction learning Tim Klucken a,b,*, Onno Kruse a,b, Jan Schweckendiek a,b, Yvonne Kuepper c, Erik M. Mueller d, Juergen Hennig c and Rudolf Stark a,b a
Department of Psychotherapy and Systems Neuroscience, Justus Liebig University Giessen, Germany Bender Institute of Neuroimaging, Justus Liebig University Giessen, Germany c Personality Psychology and Individual Differences, Justus Liebig University Giessen, Germany d Department of Clinical Psychology, Justus Liebig University Giessen, Germany b
article info
abstract
Article history:
Fear extinction is a central model for the treatment of anxiety disorders. Initial research
Received 3 June 2015
has reported that the single presentation of a conditioned stimulus prior to extinction
Revised 4 October 2015
learning can permanently block the return of fear. However, only few studies have
Accepted 16 March 2016
explored this issue and could not always replicate the findings.
Reviewed 10 July 2015
The present study examined human fear extinction using a four-day design. On the first
Action editor Peter Kirsch
day, two neutral stimuli were paired with electrical stimulation (UCS), while a third
Published online 24 March 2016
stimulus (CS) was not. Twenty-four hours later, one conditioned stimulus (CSþrem) and the CS were reminded once, 10 min before extinction learning, while the other condi-
Keywords:
tioned stimulus (CSþnon-rem) was not presented prior to extinction learning. All stimuli
Extinction
were presented during extinction learning and during two re-extinction sessions (24 h and
Fear
6-months after extinction learning) without reinforcement. Blood oxygen level-dependent
Reconsolidation
(BOLD) responses and skin conductance responses (SCRs) to both CSþ and the CS were
Amygdala
explored during acquisition, extinction, and in both re-extinction sessions.
Fear conditioning COMT
Regarding SCRs, the results showed that a single presentation of a conditioned stimulus did not block the return of fear during re-extinction: Fear recovery during re-extinction (24 h and 6-months after extinction learning) was observed for both CSþ compared with the CS with no difference between CSþrem and CSþnon-rem. Regarding BOLD-responses, no significant differences between CSþrem and CSþnon-rem were found in region of interest (ROI)-analyses (amygdala, ventromedial prefrontal cortex) during extinction learning and both re-extinction sessions. Whole-brain analyses showed increased BOLD-responses to the CSþnon-rem as compared to the CSþrem in several regions (e.g., middle frontal gyrus) during extinction learning and re-extinction (24 h after extinction learning).
* Corresponding author. Department of Psychotherapy and Systems Neuroscience, Justus Liebig University Giessen, Otto-Behaghel-Str. 10 F, 35394 Giessen, Germany. E-mail address: [email protected] (T. Klucken). http://dx.doi.org/10.1016/j.cortex.2016.03.015 0010-9452/© 2016 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http:// creativecommons.org/licenses/by-nc-nd/4.0/).
c o r t e x 7 9 ( 2 0 1 6 ) 1 1 2 e1 2 2
113
The present findings suggest that the effect of preventing the return of fear by disrupting reconsolidation seems to be a more labile phenomenon than previously assumed. Possible boundary conditions and implications are discussed. © 2016 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
1.
Introduction
Fear conditioning and extinction are well-established models for the development, maintenance, and treatment of anxiety disorders (Goode & Maren, 2014; Milad & Quirk, 2012; Milad, Rosenbaum, & Simon, 2014; Vervliet, Craske, & Hermans, 2013). While fear associations can be rapidly acquired and persist over time, the extinction memory is susceptible to disruptions and cannot permanently block the initial fear memory (Myers & Davis, 2007). Consequently, treatments of psychiatric disorders based on extinction learning (e.g., exposure therapy) often produce effective short-term fear reductions, but relapses are not uncommon (Choy, Fyer, & Lipsitz, 2007; Lipsitz, Mannuzza, Klein, Ross, & Fyer, 1999; Sharma, Thennarasu, & Janardhan Reddy, 2014). Therefore, the identification of specific factors that may decrease the return of fear and the number of relapses are of high clinical interest. Differential fear conditioning paradigms typically consist of different phases (fear acquisition, extinction learning, and reextinction). During fear acquisition, one or two neutral stimuli (CSþ) are initially paired with electrical stimulation (UCS), while another stimulus (CS) is not. After a few trials, the CSþ elicits conditioned responses (CRs) such as increased skin conductance responses (SCRs), startle amplitude, or subjective € ratings (Hamm & Weike, 2005; Lang, Davis, & Ohman, 2000). After that, the CSþ and CS are repeatedly presented without the UCS (extinction learning), which finally results in a decrease of CRs in subjective and physiological responses (Milad & Quirk, 2012; Myers & Davis, 2007; Quirk & Mueller, 2008). During this time, the extinction memory is mainly modulated by the amygdala (Quirk & Mueller, 2008). After that, the CSþ and the CS are again presented without reinforcement (re-extinction), e.g., 24 h after extinction learning. The return of fear could be observed under a variety of conditions such as spontaneous recovery, reinstatement, and renewal (Bouton, 2002; Myers & Davis, 2002, 2007). Spontaneous recovery can be described as the reappearance of previously extinguished CRs after a delay following extinction learning without any further learning sessions due to the mere passage of time. Reinstatement refers to the reoccurrence of CRs after extinction learning through the presentation of an unpredictable UCS. Finally, renewal refers to the reactivation of CRs if a subsequent test session is conducted in a different context than the extinction phase. Many methods have been developed to analyze the reoccurrence of CRs of fear during re-extinction. While some studies compared the responses towards the CSþ and the CS during the first half or the first trial of re-extinction, others authors calculated different “fear-recovery indices” (e.g., first re-extinction trial minus last extinction learning trial; Schiller, Kanen, LeDoux, Monfils, & Phelps, 2013; Schiller et al., 2010).
Recently, animal and human studies have demonstrated that the re-occurrence of conditioned fear during reextinction can be prevented by different techniques, which presumably alter the initial fear memory (Agren, 2014; Agren, Furmark, Eriksson, & Fredrikson, 2012; Johnson & Casey, 2015; Kindt, Soeter, & Vervliet, 2009; Liu et al., 2014; Nader, Schafe, & Le Doux, 2000; Schiller et al., 2010, 2013; Warren et al., 2014). A frequently used technique in human studies is the presentation of a previously conditioned stimulus (CSþrem) 10 min prior to extinction learning without reinforcement, while the other conditioned stimulus (CSþnon-rem), also previously paired with the UCS during fear acquisition, is not presented prior to extinction learning (Schiller et al., 2010, 2013). It has been suggested that this single presentation of the CSþrem reactivates the original CSþ/UCS memory, which enables a new “CSþ/no-UCS” association during extinction learning to be permanently incorporated (Agren, 2014; Schiller et al., 2010). Influential studies have demonstrated successful blocking of the return of fear to the CSþrem as compared to the CSþnon-rem during re-extinction when using this procedure (Schiller et al., 2010, 2013). Regarding the underlying neural correlates, a previous study found increased activity in the ventromedial prefrontal cortex (vmPFC) and altered effective connectivity to the CSþnon-rem compared to the CSþrem during extinction (Schiller et al., 2013). Regarding re-extinction, increased amygdala responses to the CSþnon-rem were found compared to the CSþrem, which has been assumed as an indicator for fear responses (Agren, 2014; Schiller et al., 2013). However, other studies using similar paradigms could not replicate these promising findings (Golkar, Bellander, Olsson, & Ohman, 2012; Kindt & Soeter, 2013; Soeter & Kindt, 2011). They showed a return of fear to both CSþ and could not find any differences between the CSþrem and the CSþnon-rem. In a recent review, Agren (2014) hypothesized that these contrary results might be due to specific subgroups in which the blocking is effective. It was argued that the Val158Met-polymorphism in Catechol-O-Methyl-Transferase (COMT) is of special interest, because recent studies have been able to show an association of the COMT Val158Met-polymorphism with fear acquisition and extinction learning (Agren, Furmark et al., 2012; Lonsdorf et al., 2009; Wendt et al., 2014). For instance, Lonsdorf and colleagues showed deficits in extinction learning as well as a poorer treatment outcome in extinction-based therapy in Met/Met individuals (Lonsdorf et al., 2009, 2010, 2011). However, no study has investigated the association between the COMT Val158Met-polymorphism and delayed extinction recall or return of fear. Based on the above-mentioned findings, the present study aimed to investigate the following: First, we investigated potential SCR differences between CSþrem and CSþnon-rem during
114
c o r t e x 7 9 ( 2 0 1 6 ) 1 1 2 e1 2 2
extinction learning and re-extinction. Second, we explored the underlying neural correlates of processing the CSþrem compared with the CSþnon-rem. As a supplement, we also explored the potential association between the COMT Val158Met-polymorphism, fear acquisition, extinction learning, and the return of fear. SCRs and neural activity (blood oxygen level-dependent e BOLD signal change) were measured during all phases (fear acquisition, extinction learning, and reextinction). We hypothesized augmented SCRs to the CSþnon-rem compared to the CSþrem during re-extinction. Regarding the neural correlates, increased vmPFC activity was expected in the contrast CSþnon-rem versus CSþrem during extinction learning and increased amygdala responses to the CSþnon-rem compared to the CSþrem during re-extinction.
2.
Material and methods
2.1.
Participants
One hundred and thirty six Caucasian subjects (75 females; MAge ¼ 23.93; SDAge ¼ 4.15) participated in this study. All were native German speakers, right-handed, and had normal or corrected-to-normal vision. Subjects reporting current or past mental illnesses, chronic diseases, or a consumption of psychotropic drugs were excluded. All subjects gave written informed consent and received 50 V for their participation. We restricted our sample to individuals showing reliable SCRs (>.02 ms to one UCS-presentation) and increased SCRs to the CSþ as compared to the CS during fear acquisition (mean differential SCRs >.02 ms). This criterion reduced the sample size to 70 (23 Met/Met, 28 Met/Val, 19 Val/Val) participants. The exclusion ratio is comparable to that of other studies (Agren, Engman et al., 2012; Agren, Furmark et al., 2012; Schiller et al., 2013). As a supplement, we also conducted an additional analysis on the entire sample (without any SCR exclusion criteria). The additional analysis yielded comparable results. Eight subjects could not be recruited again for the fourth day after six months, leading to a sample size of 62 subjects for this day (19 Met/Met, 26 Met/Val, 17 Val/Val). There was no significant deviation from HardyeWeinberg equilibrium [c2(1) < .1.60; p > .2]. Subjects were also interviewed using a self-developed standardized interview for stressful life events and for psychiatric disorders in the previous months on the first experimental and on the last experimental day. Several power analyses were conducted to estimate the sample size necessary to adequately address the comparison between CSþrem and CSþnon-rem. A conservative threshold of 95% sensitivity (1beta ¼ .95) and a significance level of a ¼ .05 were set for the power analyses based on the results by Schiller et al. (2010). The required sample size was below 23. The study was conducted in accordance with the Declaration of Helsinki and was approved by the institutional ethics committee.
2.2.
Fear acquisition and extinction paradigm
The fear acquisition and extinction protocol consisted of four subsequent phases (day 1: fear acquisition; day 2: extinction
learning; day 3: 24 h re-extinction; day 4: 6-month reextinction). SCRs and BOLD-responses were measured during all phases. On the first day (fear acquisition), 16 trials of each CSþ (CSþrem; CSþnon-rem; blue and yellow squares) or 16 trials of the CS were presented. Each trial started with the presentation of a fixation cross with a variable duration (0e2.5 sec) to achieve a stimulus-onset-asynchrony. After that, a CS (CSþrem; CSþnon-rem; CS) was presented for 8 sec. The UCS (duration: 100 msec) was delivered 7.9 sec after the CSþ onset and co-terminated with the CSþ offset, while the CS was never associated with the UCS. In contrast to a 38% partial reinforcement rate, which was used by Schiller et al. (2010), we used a 50% reinforcement rate (cf. Golkar et al., 2012). A fixation cross was presented during the intertrial-interval (ITI). The duration of the ITI depended on the stimulus-onset-asynchrony (5.5e8 sec), because the length of a trial was always 16 sec. A black computer screen was presented at any other time of the experiment. The instructions were to pay attention to the computer screen and try to figure out the relationship between the CS and the UCS. On the second day (extinction learning), the CSþrem and the CS were presented once in a counterbalanced order without reinforcement 10 min before the extinction learning started, while the other CSþ (CSþnon-rem) was not reminded (Schiller et al., 2010). Subjects had to watch a movie about landscapes (“Colours of Earth e Faszination Natur 2”) during the 10min break. Extinction learning consisted of 11 presentations of the CSþnon-rem and 10 presentations of the CSþrem and the CS to ensure an equal number of presentations during extinction learning, because the CSþnon-rem had not been reminded before. On day three (24 h after extinction learning) and day four (6-months after extinction learning), a reinstatement procedure was conducted inside the scanner to reactivate the fear memory before the re-extinction session started. Accordingly, the time that elapsed between the last extinction learning trial and the first reinstatement trial was around 24 h for day three and around 6-months for day four. During reinstatement, five unsignalled UCS (US to US interval: 13.5e18.5 sec) were applied, while a black computer screen was presented. Around 2 min after the last UCS application of the reinstatement, the re-extinction sessions (10 trials of each CS) started without reinforcement. A pseudorandomized stimulus order was used on each experimental day, which ensured the presentation of all CS with equal frequency in the first and the second half, with no more than two presentations in succession and an equal frequency of all CS for the first trial. In addition, the UCS was presented equally often in the first and in the second half of day one. The UCS intensity was set individually using a gradually increasing procedure to achieve an “unpleasant but not painful” level of sensation. No UCS-recalibration was made on day 2, 3, and 4, to ensure the same number of UCS applications for each subject. A custom-made impulse-generator (833 Hz; 5 mA) provided transcutaneous electrical stimulation (UCS) through two Ag/AgCl electrodes (1 mm2 surface), which was triggered via an optic fiber cable. Electrodes were fixed to the middle of the left shin.
c o r t e x 7 9 ( 2 0 1 6 ) 1 1 2 e1 2 2
2.3.
Skin conductance measuring
SCRs were sampled using Ag/AgCl electrodes filled with isotonic (.05 M NaCl) electrolyte medium placed at the nondominant left hand. An SCR was defined as the highest phasic response following stimulus onset. Therefore, the largest difference between a minimum, which had to occur within 1e8 sec after onset of the CS, and the subsequent maximum was extracted using Ledalab 3.4.4 (Benedek & Kaernbach, 2010). SCRs were ln (mS þ 1), corrected for violation of normal distribution of the data. Outlier responses (SCRs ± 3 standard deviations) were excluded. Mean SCRs to CSþ and CS were analyzed via analysis of variance (ANOVA) with the factors CS-type (CSþrem, CSþnonrem, CS) number of trials (e.g., 16 in fear acquisition), followed by Bonferroni-corrected post-hoc tests, using SPSS 22 (SPSS Inc. Chicago, Illinois, USA). Furthermore, the fear recovery indexes (SCR of first re-extinction trial minus SCR of last extinction learning trial) of the CSþrem with the CSþnonrem were compared to analyze potential differences between both CSþ. In addition, we also investigated the first trial separately (SCR of the first trial of the CSþrem minus SCR of the first trial of the CSþnon-rem), because Schiller et al. (2013) found the greatest differences between both CSþ in this trial. Finally, in line with other studies (Golkar et al., 2012; Schiller et al., 2010, 2013), we also conducted these analyses with the first interval response only (1e4 sec after CS onset) but results were comparable.
2.4.
Magnetic resonance imaging
Subjects were scanned using a 1.5 T whole-body tomograph (Siemens Symphony with a quantum gradient system) with a standard head coil. 160 T1-weighted images (MPRage, 1 mm slice thickness) were acquired in sagittal orientation. Functional imaging consisted of 420 images in a T2*-weighted echo-planar imaging (EPI) sequence. For functional images, a T2*-weighted gradient EPI sequence was used with 25 slices covering the whole brain (slice thickness ¼ 5 mm; 1 mm gap; descending slice procedure; TR ¼ 2.5 sec; TE ¼ 55 msec; flip angle ¼ 90 ; field of view ¼ 192 192 mm; matrix size ¼ 64 64). The orientation of the axial slices was paralleled to the orbitofrontal cortex-bone transition in order to minimize susceptibility artefacts in prefrontal areas. Prior to all statistical analyses, data were preprocessed as described previously (Klucken, Kruse et al., 2015). The functional data were analyzed using a random-effects general linear model. For each day, separate first level models were calculated. In line with Schiller et al. (2013), separate boxcar regressors (CS duration) for each CS were calculated for each model of extinction learning and both re-extinction sessions (24 h and 6-months after extinction learning) and were divided into an early (first half) and a late (second half) phase, because Schiller et al. (2013) showed the greatest differences in the early, but not the late phase during extinction learning and re-extinction. We also conducted an alternative model using a stick function, but results were comparable. On the first day, regressors for the UCS and the non-UCS (time window corresponding to the UCS after the CS) were also introduced in the model (maximum dependency between
115
CSþ and UCS < .20). The six movement parameters estimated in the preprocessing realignment step were entered into all first level models. The voxel-based time series was filtered with a high pass filter (time constant ¼ 128 sec). On the group level, a 3 (CSþrem, CSþnon-rem, CS) 2 (early phase, late phase) ANOVA was computed in SPM8 to explore main effects of CS-type, phase, and CS-type by phase interaction effects. The threshold for whole-brain analyses (for extinction learning and re-extinction) was set to p < .001 and k > 10 to detect potential differences with a liberal criterion. We focused our region of interest (ROI) analyses on the amygdala and the vmPFC, because (1) both of these structures are crucially involved in fear acquisition, extinction learning, and re-extinction; and (2) previous studies found significant differences between CSþrem and CSþnon-rem in these two structures only (Agren, Engman et al., 2012; Schiller et al., 2013). ROI analyses were performed with a threshold of p < .05 (family-wise-error; FWE-corrected) and k > 5. The amygdala mask was taken from the “Harvard Oxford cortical and subcortical structural atlases” provided by the Harvard Center for Morphometric Analysis. The vmPFC mask was created with MARINA (Walter et al., 2003) and has been used in previous studies (e.g., Hermann, Keck, & Stark, 2014; Klucken, Schweckendiek et al., 2015; Klucken, Schweckendiek, Merz, Vaitl, & Stark, 2013). An additional analysis was conducted exploring differences between the CSþrem and the CSþnon-rem within the first trial during re-extinction. This procedure was chosen because the comparison of the first trial of the CSþrem and the CSþnonrem is of special interest, due to previous findings that showed differences between both CSþ in the first trial (parallel to the SCR-analyses of re-extinction; cf. Schiller et al., 2013). Therefore, we specified a new first level model for each subject including all trials as separate regressors (i.e., CSþrem_trial_1, CSþrem_trial_2,…, CSþnon-rem_trial_1, CSþnon-rem_trial_2,…). This results in 3 10 conditions (10 trials for the CSþrem, CSþnon-rem, and for the CS). Next, contrasts were created for each comparison of interest (e.g., CSþrem_trial_1 CSþnonrem_trial_1) and analyzed on the second level with the same whole-brain and ROI-analyses parameters as described above.
3.
Results
3.1.
Fear acquisition (day 1)
We will only briefly describe the fear acquisition results to ensure that CSþrem and CSþnon-rem did not differ, because the primary aim of the present study was to explore extinction learning and re-extinction. In addition, we did not expect significant differences, because an inclusion criterion for this study was successful conditioning on the first day.
3.1.1.
SCRs
The ANOVA revealed a main effect of CS-type [F(2, 68) ¼ 55.34; p < .001], trial [F(1, 69) ¼ 31.88; p < .001], as well as a CS-type x trial [F(2, 68) ¼ 13.29; p < .001] interaction effect. The post-hoc tests showed increased SCRs to the CSþrem and to the CSþnon-rem as compared to the CS (Fig. 1), but no significant differences between the two CSþ.
116
c o r t e x 7 9 ( 2 0 1 6 ) 1 1 2 e1 2 2
Fig. 1 e Mean skin conductance responses (ln (1 þ ms)) for the CSþrem, CSþnon-rem, and the CS¡ for each trial for the first day (fear conditioning) and the second day (extinction learning). *indicates significantly (p < .05) increased responses as compared to the CS¡ during extinction learning.
3.1.2.
Hemodynamic responses
We found a main effect of CS-type, showing increased amygdala responses to the CSþrem (x/y/z ¼ 30/4/26; z ¼ 4.07; p ¼ .002) and to the CSþnon-rem (x/y/z ¼ 30/4/26; z ¼ 3.74; p ¼ .006) as compared to the CS. We did not find increased left (z ¼ 2.15; pFWE-corr ¼ .24) or right (z ¼ 2.34; pFWE¼ .18) amygdala activations in the contrast corr CSþrem CSþnon-rem or vice versa (pFWE-corr > .50).
3.2.
Extinction learning (day 2)
3.2.1.
SCRs
The main research question was whether differences between the CSþrem and the CSþnon-rem would occur during extinction learning and during re-extinction. The ANOVA revealed a main effect of CS-type [F(2, 68) ¼ 10.14; p < .001] and trial [F(9, 61) ¼ 42.32; p < .001], showing increased SCRs to the CSþrem and to the CSþnon-rem as compared to the CS, but no differences between both CSþ (p > .20). In addition, we focused on the first trial of the CSþrem and the CSþnon-rem, because Schiller et al. (2013) found the greatest differences for the first trial. We found increased SCRs to both CSþ as compared to the CS (all p < .05), but no differences between the CSþrem and the CSþnon-rem (p ¼ .969; effect size: r ¼ .002; Fig. 1).
3.2.2.
Hemodynamic responses
Because the aim of the study was to explore potential differences between the CSþrem and the CSþnon-rem, we conducted whole-brain analyses as well as ROI-analyses and computed the contrasts CSþrem CSþnon-rem and CSþnon-rem CSþrem. In addition, because previous studies also found differences in the early half of extinction learning (Schiller et al., 2013), we also analyzed BOLD-responses during the early phase of extinction learning. Regarding ROI-analyses, we did not find significant BOLDresponses during the whole extinction phase for the contrast CSþrem CSþnon-rem in the amygdala (“peak voxel”: x/y/z ¼ 27/2/26; z ¼ 1.56; pFWE-corr ¼ .558) or in the vmPFC (x/y/ z ¼ 12/62/2; z ¼ 1.34; pFWE-corr ¼ .909). In addition, no significant activation was found for the opposite contrast (CSþnon-rem CSþrem) in the amygdala (x/y/z ¼ 12/7/17; z ¼ 1.45; pFWE-corr ¼ .585) or in the vmPFC (x/y/z ¼ 9/29/23;
z ¼ 2.34; pFWE-corr ¼ .449). Moreover, ROI-analyses showed no increased activation during the early phase of extinction learning for the contrast CSþrem CSþnon-rem in the amygdala (x/y/z ¼ 18/10/11; z ¼ 1.99; pFWE-corr ¼ .355), the vmPFC (x/y/ z ¼ 9/29/23; z ¼ 2.34; pFWE-corr ¼ .449), and for the opposite contrast (amygdala: x/y/z ¼ 24/7/20; z ¼ 1.96; pFWEcorr ¼ .350); vmPFC (x/y/z ¼ 30/7/20; z ¼ 1.19; pFWE-corr ¼ .698). Finally, ROI-analyses did not show significant differences in the contrasts CSþrem CS and CSþnon-rem CS. Whole-brain results did not show increased activations to the CSþrem compared with the CSþnon-rem (or vice versa) during the whole extinction phase (all puncorrected > .001), but increased BOLD-responses in several regions in the contrast CSþnon-rem CSþrem in the early half of extinction learning and increased activations of the CSþrem and the CSþnon-rem as compared with the CSe (Table 1).
3.3.
Re-extinction (day 3 and day 4)
3.3.1.
SCRs
Day 3 (24-h after extinction learning). The ANOVA showed a significant main effect of CS-type [F(2, 68) ¼ 6.35 p < .01] and trial [F(9, 61) ¼ 16.01; p < .001], with increased SCRs in the first trial to both CSþ, which diminished over time. We also calculated the recovery index and found increased recovery for the CSþrem and the CSþnon-rem (all p < .05). Notably, no significant differences could be found between the CSþrem and the CSþnon-rem in the first trial (p ¼ .269; effect size: r ¼ .07) or for the recovery index (p ¼ .46; effect size: r ¼ .04). Day 4 (6-months after extinction learning). The ANOVA revealed a significant main effect of time [F(9, 53) ¼ 14.03; p < .001] and a CS-type time trend [F(18, 44) ¼ 1.65; p ¼ .084]. In addition, increased SCRs to both CSþ as compared to the CS could also be found in the first re-extinction trial as well as an increased recovery index to both CSþ (all p < .05). Again, no significant differences were observed between the CSþrem and the CSþnon-rem in the first trial (p ¼ .230; effect size: r ¼ .068), or for the recovery index (p ¼ .461; effect size: r ¼ .08). In addition, we also investigated the differences between CSþrem and CSþnon-rem including all subjects, but results were comparable (see supplement for details).
117
c o r t e x 7 9 ( 2 0 1 6 ) 1 1 2 e1 2 2
Table 1 e Localization and statistics for whole-brain results for day 2 (extinction learning). Contrast Whole phase
CSþrem CSþnon-rem CSþnon-rem CSþrem CSþrem CS
CSþnon-rem CS Early phase
CSþrem CSþnon-rem CSþnon-rem CSþrem
CSþrem CS
CSþnon-rem CS
Structure
Side
Occipital cortex Supramarginal gyrus Insula Middle temporal gyrus Middle temporal gyrus Insula Occipital cortex Insula
L L L R L R L R
Temporal gyrus Middle frontal gyrus Orbitofrontal cortex Occipital cortex Insula Insula Middle temporal gyrus Supramarginal gyrus Limbic lobe Cerebellum Occipital cortex Orbitofrontal cortex Supramarginal gyrus Insula Limbic lobe Middle temporal gyrus Inferior temporal gyrus
L L R R R L R L L R L R L L R R L
k
x
y
No significant differences No significant differences 663 3 88 44 60 31 32 45 8 21 45 67 12 48 64 24 48 5 719 3 91 10 36 20 No significant differences 22 66 31 11 42 23 13 3 32 390 12 85 191 33 23 220 36 14 38 48 64 81 66 25 37 3 14 12 9 40 350 3 94 127 33 23 79 54 28 22 33 26 15 12 25 10 66 28 16 42 40
z
zmax
p
1 25 7 7 7 1 1 8
6.25 3.90 3.90 3.60 3.46 3.45 5.85 3.33
<.001 <.001 <.001 <.001 <.001 <.001 <.001 <.001
14 52 11 7 7 7 10 28 22 8 1 8 25 4 34 14 26
3.96 3.69 3.62 4.90 4.65 4.32 4.02 4.03 3.79 3.67 4.48 4.38 4.06 3.64 3.59 3.58 3.51
<.001 <.001 <.001 <.001 <.001 <.001 <.001 <.001 <.001 <.001 <.001 <.001 <.001 <.001 <.001 <.001 <.001
The threshold was p < .001 (uncorrected; whole-brain results according to SPM8). All coordinates are given in MNI space. L: left hemisphere, R: right hemisphere.
3.3.2.
Hemodynamic responses
Day 3 (24-h after extinction learning). Regarding ROI-analyses, no increased activation could be found for the contrast CSþrem CSþnon-rem during the whole phase in the amygdala or in the vmPFC (no suprathreshold clusters). Moreover, no increased ROI-activation was found for the opposite contrast (amygdala: x/y/z ¼ 18/13/14; z ¼ 2.73; pFWE-corr ¼ .097; vmPFC: x/y/z ¼ 3/59/23; z ¼ 3.12; pFWE-corr ¼ .105). In addition, no significant BOLD-responses could be found to the CSþrem compared to the CSþnon-rem during the early phase (amygdala: x/y/z ¼ 21/7/11; z ¼ .20; pFWE-corr ¼ .837; vmPFC: x/y/z ¼ 9/29/14; z ¼ 1.30; pFWE-corr ¼ .943). In addition, we did not find any significantly increased ROI-activations for the opposite contrast (amygdala: x/y/z ¼ 24/4/23; z ¼ 2.13; pFWE-corr ¼ .304; vmPFC: x/y/z ¼ 0/62/14; z ¼ 1.91; pFWEcorr ¼ .755). No significant differences were found in ROIanalyses in the contrasts CSþrem CS and CSþnonrem CS. Finally, no significant differences between CSþrem CSþnon-rem (or vice versa) were found in all ROI when only comparing activations in the first trial. Increased whole-brain results were found for the contrast CSþnon-rem CSþrem in the whole phase and in the first trial, but not for the early half of re-extinction, while the opposite contrast showed only significant (uncorrected) hippocampal activations in the first trial (see Table 2). Day 4 (6-months after extinction learning). ROI-analyses showed no increased activation for the contrast
CSþrem CSþnon-rem during the whole phase (amygdala: x/y/ z ¼ 30/1/20; z ¼ 1.02; pFWE-corr ¼ 745; vmPFC: x/y/z ¼ 6/44/ 1; z ¼ 3.20; pFWE-corr ¼ .081). In addition, we did not find any increased ROI-activation for the opposite contrast (amygdala: x/y/z ¼ 18/7/20; z ¼ .97; pFWE-corr ¼ .757; vmPFC: x/y/ z ¼ 12/68/2; z ¼ 1.25; pFWE-corr ¼ .931). Moreover, no significant BOLD-responses to the CSþrem compared to the CSþnonrem during the early phase could be observed in the amygdala (x/y/z ¼ 15/4/14; z ¼ .74; pFWE-corr ¼ .804) or in the vmPFC (x/y/z ¼ 9/44/8; z ¼ 2.83; pFWE-corr ¼ .197). In addition, we did not find any significantly increased ROI-activations for the opposite contrast (amygdala: x/y/z ¼ 12/7/17; z ¼ 2.02; pFWEcorr ¼ .352; vmPFC: x/y/z ¼ 15/29/17; z ¼ 1.58; pFWE-corr ¼ .874). Regarding the first trial, no significant differences between both CSþ were found in ROI-analyses. Finally, no significant ROI-activations were found in the contrasts CSþrem CS and CSþnon-rem CS. Whole-brain analyses showed increased BOLD-responses to the CSþrem as compared to the CSþnon-rem only in the precuneus in the whole phase and in the occipital cortex, precuneus, and supramarginal gyrus in the first trial and pronounced activation to the CSþrem as compared to the CS in the occipital cortex (Table 3; Fig. 2).
3.4.
COMT Val158Met-polymorphism
A detailed results section and a description of the genotyping and data analyses are included in the supplement. In short, regarding SCRs, neither main effects nor interactions with the
118
c o r t e x 7 9 ( 2 0 1 6 ) 1 1 2 e1 2 2
Table 2 e Localization and statistics for whole-brain results for day 3 (re-extinction). Contrast Whole phase
CSþrem CSþnon-rem CSþnon-rem CSþrem
CSþrem CS CSþnon-rem CS
Early phase
First trial
CSþrem CSþnon-rem CSþnon-rem CSþrem CSþrem CS CSþnon-rem CS CSþrem CSþnon-rem CSþnon-rem CSþrem
Structure
Side
Postcentral gyrus Precentral gyrus Middle temporal gyrus No distinct region Occipital cortex Occipital cortex Precuneus Inferior temporal gyrus
R R L R R R R R
Occipital cortex Precuneus Middle temporal gyrus Precentral gyrus Superior parietal gyrus Hippocampus
L L R R R L
k
x
Y
No significant differences 25 48 19 17 12 28 12 60 7 59 21 2 260 12 82 782 12 82 132 15 58 89 51 64 No significant differences No significant differences 283 6 82 16 21 49 24 48 67 18 27 19 23 27 58 14 27 19
z
zmax
p
34 73 26 16 7 10 22 5
3.76 3.72 3.67 3.90 4.56 5.08 4.09 4.00
<.001 <.001 <.001 <.001 <.001 <.001 <.001 <.001
7 4 16 70 64 17
4.96 3.58 3.55 3.69 3.65 3.81
<.001 <.001 <.001 <.001 <.001 <.001
The threshold was p < .001 (uncorrected; whole-brain results according to SPM8). All coordinates are given in MNI space. L: left hemisphere, R: right hemisphere.
Table 3 e Localization and statistics for whole-brain results for day 4 (re-extinction). Contrast Whole phase
Early phase
First trial
CSþrem CSþnon-rem CSþnon-rem CSþrem CSþrem CS CSþnon-rem CS CSþrem CSþnon-rem CSþnon-rem CSþrem CSþrem CS CSþnon-rem CS CSþrem CSþnon-rem
Structure
Side
k
Precuneus
L
Occipital cortex Occipital cortex Occipital cortex
L R R
Occipital cortex Occipital cortex
L R
Occipital cortex Precuneus Supramarginal gyrus
R R R
44 No 81 75 33 No No 31 21 No 45 81 17 No
CSþnon-rem CSþrem
x
y
3 61 significant differences 12 85 15 73 3 85 significant differences significant differences 9 85 21 73 significant differences 36 82 2 70 39 40 significant differences
z
zmax
p
19
3.66
<.001
4 13 4
4.33 4.18 3.72
<.001 <.001 <.001
7 13
3.53 3.66
<.001 <.001
22 31 43
4.96 3.91 3.76
<.001 <.001 <.001
The threshold was p < .001 (uncorrected; whole-brain results according to SPM8). All coordinates are given in MNI space. L: left hemisphere, R: right hemisphere.
COMT Val158Met-polymorphism were observed during fear acquisition, extinction learning, and re-extinction. BOLD-responses also showed no significant amygdala and vmPFC differences during fear acquisition, extinction learning, and re-extinction with respect to the COMT Val158Metpolymorphism.
4.
Discussion
The present study explored the impact of a single CSþ presentation prior to extinction learning on re-extinction. SCRs and BOLD-responses between the reminded and nonreminded CSþ were compared during all phases. The results revealed similarly increased SCRs to the CSþrem and to the CSþnonrem as compared to the CS during extinction learning and re-extinction in the present study. Consistent with the lack of CSþrem versus CSþnon-rem differences in SCRs, ROI-
analyses also showed no significant differences between the both CSþ during extinction learning and re-extinction. Moreover, no association with the COMT Val158Met-polymorphism was observed. The present findings could not replicate previous results (Agren, Engman et al., 2012; Agren, Furmark et al., 2012; Schiller et al., 2013; Schiller et al., 2010), which had suggested that a single presentation of a CSþ prior to extinction learning had a substantial impact on re-extinction. Yet, our results support other studies also showing no differences between the CSþrem and the CSþnon-rem during re-extinction (Golkar et al., 2012; Kindt & Soeter, 2013; Soeter & Kindt, 2011). In order to explain these inconsistent findings, different boundary conditions are discussed, which might have influenced the results. First, it could be assumed that differences in the experimental design such as CS duration, ITI, exact reinstatement procedure, UCS-recalibration, etc. might alter the effects
c o r t e x 7 9 ( 2 0 1 6 ) 1 1 2 e1 2 2
119
Fig. 2 e Mean skin conductance responses (ln (1 þ ms)) for the CSþrem, CSþnon-rem, and the CS¡ for each trial during reextinction (day 3 and day 4). Recovery index (below) for re-extinction for the CSþrem and the CSþnon-rem. *indicates significantly (p < .05) increased responses (return of fear) as compared to the CS¡.
(Agren, 2014; Golkar et al., 2012). For instance, Schiller et al. (2010) and the present study presented the CSþrem and the CS prior to extinction learning, while other studies presented the CSþrem before extinction learning only. In a recent review, Haaker, Golkar, Hermans, and Lonsdorf (2014) hypothesized that slight differences in the conditioning design could significantly influence the return of fear. Second, the exclusion ratios in the present study (~50%) as well as the fMRI-study (~73%) by Schiller et al. (2013) were higher compared to other extinction studies (e.g., ~32% in Golkar et al., 2012), which may also impact findings. However, the present results were comparable, if the entire sample was investigated (see Supplement). Nevertheless, independent replications are required to investigate this unexpected exclusion ratio in more detail. Third, it is also possible that different analysis strategies might have led to the heterogeneous findings. A variety of different approaches exist to analyze the return of fear index. While some studies calculated the recovery index based on the first trial of re-extinction and the last extinction learning trial (e.g., Schiller et al., 2013), other studies compared SCRs of the first trial (or the first half) during re-extinction of each CSþ with each other. To account for this, we conducted several analyses to explore differences between both CSþ during the early half only or at a single trial level, but did not find any significant SCR differences between the CSþrem and the CSþnon-rem. However, differences in the analysis methods as well as the comparison of CRs over different days could be problematic, because of habituation effects at the end of day 2 and large arousal effects on day 3 and day 4. In addition, it
should be noted that the used procedure does not allow a distinction of the exact mode of return of fear because reinstatement and spontaneous recovery effects are intermixed (Haaker et al., 2014). Moreover, it is also possible that differences in the ratios of contingency aware and contingency unaware subjects between the studies are responsible for the contrary findings. Previous studies showed that contingency awareness impacts on SCRs and neural activity, which were measured in the present study (Hamm & Weike, 2005; Klucken, Kagerer et al., 2009; Klucken, Tabbert et al., 2009; Mertens et al., 2015; Weike et al., 2005; Weike, Schupp, & Hamm, 2007). However, previous studies as well as the present study did not measure contingency awareness explicitly to prevent a repeated presentation of the CS, which might also have impacted extinction processes. Instead, we instructed subjects to pay attention to potential CS/UCS contingencies, which is in line with previous studies (Schiller et al., 2010, 2013). Finally, Agren (2014) hypothesized that the initial effects may occur in specific subgroups only, which differ in certain genotypes. For instance, an association between fear acquisition and extinction with genetic variation in the serotonin transporter gene (serotonergic transporter-linked polymorphic region; 5-HTTLPR, rs25531) and the Val158Met-polymorphism in the COMT was found, which are both closely related to fear learning, extinction, and updating of previously learned contingencies (Agren, Furmark et al., 2012; Cris‚an et al., 2009; Hermann et al., 2012; Klucken, Alexander et al., 2013; Klucken, Schweckendiek et al., 2015; Klumpers, Heitland, Oosting, Kenemans, & Baas, 2012;
120
c o r t e x 7 9 ( 2 0 1 6 ) 1 1 2 e1 2 2
Lonsdorf et al., 2009, 2011; Wendt, et al., 2014). As a supplement, we reanalyzed our results with respect to the COMT Val158Met-polymorphism, but did not find any significant associations with the COMT Val158Met-polymorphism. However, a previous study also found no relationship between the COMT Val158Met-polymorphism during extinction learning with SCRs, but with conditioned startle amplitude, suggesting that COMT Val158Met-polymorphism might be associated with specific response systems (Lonsdorf et al., 2009). A possible explanation for this selective effect is the assumption that conditioned startle amplitudes and SCRs are based, at least partly, on different neural circuits (Hamm & Weike, 2005; Weike et al., 2005). While conditioned startle amplitude is primarily mediated by amygdala activations (Davis & Whalen, 2001; Hamm & Weike, 2005), conditioned SCRs could be observed without an involvement of the amygdala (Weike et al., 2005; Tabbert, Stark, Kirsch, & Vaitl, 2006). Thus, the potential association of COMT Val158Metpolymorphism on extinction learning and re-extinction might be visible only in predominantly amygdaladependent response systems. Regarding (uncorrected significant) whole-brain results, we found differences between the CSþrem and the CSþnon-rem during extinction learning and re-extinction. For instance, increased BOLD-activations were found to the CSþnon-rem compared to the CSþrem in the orbitofrontal cortex, the middle frontal gyrus during the early phase of extinction learning. During extinction learning, it is necessary to adapt previously learned contingencies to new circumstances. The middle frontal gyrus has been linked to contingency awareness during fear acquisition (Carter, O'Doherty, Seymour, Koch, & Dolan, 2006). It seems possible that the process to change the CS/UCS relationship of the CSþnon-rem may require more explicit effort than the adaptation to the CSþrem, which may have led to the observed effects in this structure. In addition, increased activations to both CSþ in contrast to the CS were also found in prefrontal areas, the limbic lobe, the occipital cortex, and further structures during extinction learning and/ or re-extinction (see tables for detailed results). However, it should be pointed out that these whole-brain analyses are partly unexpected and the threshold for whole-brain analyses was set to a liberal criterion (uncorrected p < .001). Therefore, the results should be treated with caution until an independent replication is available. In conclusion, we would like to point out that the present results do not argue against the phenomenon of blocking the return of fear in general. Various animal and human studies found successful blocking of the return of fear (Agren, 2014; Johnson & Casey, 2015; Kindt et al., 2009; Liu et al., 2014; Nader et al., 2000; Schiller et al., 2010, 2013) and it is also possible that our results might be due to false-negative probability. Nevertheless, the unexpected (null) results may support the view that the reported effects may exist in specific constellations only and may require specific parameters, which are not entirely clear to date (Agren, 2014; Golkar et al., 2012). For instance, Schiller and colleagues also reported data of subjects without successfully blocked return of fear (Schiller et al., 2013). Therefore, future studies may help to give more insight into the impact and boundaries and the opportunity for blocking the return of fear more efficiently.
Nevertheless, the possibility of preventing the return of fear through disrupting reconsolidation is still fascinating and could someday be implemented in treatments for anxiety disorders (Shiban, Bru¨tting, Pauli, & Mu¨hlberger, 2015).
Financial disclosures All authors report no biomedical financial interests or other potential conflicts of interest.
Acknowledgments This study was supported by a research grant from the Deutsche Forschungsgemeinschaft German Research Foundation to the first author (Grant number: Klu 2500/1-1).
Supplementary data Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.cortex.2016.03.015.
references
Agren, T. (2014). Human reconsolidation: a reactivation and update. Brain Research Bulletin, 105, 70e82. € rkstrand, J., Larsson, E., Agren, T., Engman, J., Frick, A., Bjo Furmark, T., et al. (2012). Disruption of reconsolidation erases a fear memory trace in the human amygdala. Science, 337(6101), 1550e1552. Agren, T., Furmark, T., Eriksson, E., & Fredrikson, M. (2012). Human fear reconsolidation and allelic differences in serotonergic and dopaminergic genes. Translational Psychiatry, 2, e76. Benedek, M., & Kaernbach, C. (2010). A continuous measure of phasic electrodermal activity. Journal of Neuroscience Methods, 190(1), 80e91. Bouton, M. E. (2002). Context, ambiguity, and unlearning: sources of relapse after behavioral extinction. Biological Psychiatry, 52, 976e986. Carter, R. M., O'Doherty, J. P., Seymour, B., Koch, C., & Dolan, R. J. (2006). Contingency awareness in human aversive conditioning involves the middle frontal gyrus. NeuroImage, 29(3), 1007e1012. Choy, Y., Fyer, A. J., & Lipsitz, J. D. (2007). Treatment of specific phobia in adults. Clinical Psychology Review, 27(3), 266e286. Cris‚an, L. G., Pana, S., Vulturar, R., Heilman, R. M., Szekely, R., a, B., et al. (2009). Genetic contributions of the serotonin Drug transporter to social learning of fear and economic decision making. Social Cognitive and Affective Neuroscience, 4(4), 399e408. Davis, M., & Whalen, P. J. (2001). The amygdala: vigilance and emotion. Molecular Psychiatry, 6(1), 13e34. Golkar, A., Bellander, M., Olsson, A., & Ohman, A. (2012). Are fear memories erasable? Reconsolidation of learned fear with fearrelevant and fear-irrelevant stimuli. Frontiers in Behavioral Neuroscience, 6, 80. Goode, T. D., & Maren, S. (2014). Animal models of fear relapse. ILAR Journal/National Research Council, Institute of Laboratory Animal Resources, 55(2), 246e258.
c o r t e x 7 9 ( 2 0 1 6 ) 1 1 2 e1 2 2
Haaker, J., Golkar, A., Hermans, D., & Lonsdorf, T. B. (2014). A review on human reinstatement studies: an overview and methodological challenges. Learning and Memory, 21(9), 424e440. Hamm, A. O., & Weike, A. I. (2005). The neuropsychology of fear learning and fear regulation. International Journal of Psychophysiology, 57(1), 5e14. Hermann, A., Keck, T., & Stark, R. (2014). Dispositional cognitive reappraisal modulates the neural correlates of fear acquisition and extinction. Neurobiology of Learning and Memory, 113, 115e124. Hermann, A., Ku¨pper, Y., Schmitz, A., Walter, B., Vaitl, D., Hennig, J., et al. (2012). Functional gene polymorphisms in the serotonin system and traumatic life events modulate the neural basis of fear acquisition and extinction. PLoS One, 7(9), e44352. Johnson, D. C., & Casey, B. J. (2015). Extinction during memory reconsolidation blocks recovery of fear in adolescents. Scientific Reports, 5, 8863. Kindt, M., & Soeter, M. (2013). Reconsolidation in a human fear conditioning study: a test of extinction as updating mechanism. Biological Psychology, 92(1), 43e50. Kindt, M., Soeter, M., & Vervliet, B. (2009). Beyond extinction: erasing human fear responses and preventing the return of fear. Nature Neuroscience, 12(3), 256e258. Klucken, T., Alexander, N., Schweckendiek, J., Merz, C. J., Kagerer, S., Osinsky, R., et al. (2013). Individual differences in neural correlates of fear conditioning as a function of 5HTTLPR and stressful life events. Social Cognitive and Affective Neuroscience, 8(3), 318e325. Klucken, T., Kagerer, S., Schweckendiek, J., Tabbert, K., Vaitl, D., & Stark, R. (2009). Neural, electrodermal and behavioral response patterns in contingency aware and unaware subjects during a picture-picture conditioning paradigm. Neuroscience, 158(2), 721e731. Klucken, T., Kruse, O., Wehrum-Osinsky, S., Hennig, J., Schweckendiek, J., & Stark, R. (2015). Impact of COMT Val158Met-polymorphism on appetitive conditioning and amygdala/prefrontal effective connectivity. Human Brain Mapping, 36(3), 1093e1101. Klucken, T., Schweckendiek, J., Blecker, C., Walter, B., Kuepper, Y., Hennig, J., et al. (2015). The association between the 5-HTTLPR and neural correlates of fear conditioning and connectivity. Social Cognitive and Affective Neuroscience, 10(5), 700e707. Klucken, T., Schweckendiek, J., Merz, C. J., Vaitl, D., & Stark, R. (2013). Dissociation of neuronal, electrodermal, and evaluative responses in disgust extinction. Behavioral Neuroscience, 127(3), 380e386. Klucken, T., Tabbert, K., Schweckendiek, J., Merz, C. J., Kagerer, S., Vaitl, D., et al. (2009). Contingency learning in human fear conditioning involves the ventral striatum. Human Brain Mapping, 30(11), 3636e3644. Klumpers, F., Heitland, I., Oosting, R. S., Kenemans, J. L., & Baas, J. M. P. (2012). Genetic variation in serotonin transporter function affects human fear expression indexed by fearpotentiated startle. Biological Psychology, 89(2), 277e282. € Lang, P. J., Davis, M., & Ohman, A. (2000). Fear and anxiety: animal models and human cognitive psychophysiology. Journal of Affective Disorders, 61(3), 137e159. Lipsitz, J. D., Mannuzza, S., Klein, D. F., Ross, D. C., & Fyer, A. J. (1999). Specific phobia 10-16 years after treatment. Depression and Anxiety, 10(3), 105e111. Liu, J., Zhao, L., Xue, Y., Shi, J., Suo, L., Luo, Y., et al. (2014). An unconditioned stimulus retrieval extinction procedure to prevent the return of fear memory. Biological Psychiatry, 76(11), 895e901.
121
€ m, K. M., Fransson, P., Lonsdorf, T. B., Golkar, A., Lindsto Schalling, M., Ohman, A., et al. (2011). 5-HTTLPR and COMTval158met genotype gate amygdala reactivity and habituation. Biological Psychology, 87(1), 106e112. € m, J., Andersson, G., Ohman, A., Lonsdorf, T. B., Ru¨ck, C., Bergstro Lindefors, N., et al. (2010). The COMTval158met polymorphism is associated with symptom relief during exposure-based cognitive-behavioral treatment in panic disorder. BMC Psychiatry, 10, 99. Lonsdorf, T. B., Weike, A. I., Nikamo, P., Schalling, M., Hamm, A. O., & Ohman, A. (2009). Genetic gating of human fear learning and extinction: possible implications for geneenvironment interaction in anxiety disorder. Psychological Science, 20(2), 198e206. Mertens, G., Kuhn, M., Raes, A. K., Kalisch, R., de Houwer, J., & Lonsdorf, T. B. (2015). Fear expression and return of fear following threat instruction with or without direct contingency experience. Cognition and Emotion, 1e17. Milad, M. R., & Quirk, G. J. (2012). Fear extinction as a model for translational neuroscience: ten years of progress. Annual Review of Psychology, 63, 129e151. Milad, M. R., Rosenbaum, B. L., & Simon, N. M. (2014). Neuroscience of fear extinction: implications for assessment and treatment of fear-based and anxiety related disorders. Behaviour Research and Therapy, 62, 17e23. Myers, K. M., & Davis, M. (2002). Behavioral and neural analysis of extinction. Neuron, 36(4), 567e584. Myers, K. M., & Davis, M. (2007). Mechanisms of fear extinction. Molecular Psychiatry, 12(2), 120e150. Nader, K., Schafe, G. E., & Le Doux, J. E. (2000). Fear memories require protein synthesis in the amygdala for reconsolidation after retrieval. Nature, 406(6797), 722e726. Quirk, G. J., & Mueller, D. (2008). Neural mechanisms of extinction learning and retrieval. Neuropsychopharmacology Reviews, 33(1), 56e72. Schiller, D., Kanen, J. W., LeDoux, J. E., Monfils, M., & Phelps, E. A. (2013). Extinction during reconsolidation of threat memory diminishes prefrontal cortex involvement. Proceedings of the National Academy of Sciences of the United States of America, 110(50), 20040e20045. Schiller, D., Monfils, M., Raio, C. M., Johnson, D. C., LeDoux, J. E., & Phelps, E. A. (2010). Preventing the return of fear in humans using reconsolidation update mechanisms. Nature, 463(7277), 49e53. Sharma, E., Thennarasu, K., & Janardhan Reddy, Y. C. (2014). Long-term outcome of obsessive-compulsive disorder in adults: a meta-analysis. Journal of Clinical Psychiatry, 75(9), 1019e1027. Shiban, Y., Bru¨tting, J., Pauli, P., & Mu¨hlberger, A. (2015). Fear reactivation prior to exposure therapy: does it facilitate the effects of VR exposure in a randomized clinical sample? Journal of Behavior Therapy and Experimental Psychiatry, 46, 133e140. Soeter, M., & Kindt, M. (2011). Disrupting reconsolidation: pharmacological and behavioral manipulations. Learning and Memory, 18(6), 357e366. Tabbert, K., Stark, R., Kirsch, P., & Vaitl, D. (2006). Dissociation of neural responses and skin conductance reactions during fear conditioning with and without awareness of stimulus contingencies. NeuroImage, 32(2), 761e770. Vervliet, B., Craske, M. G., & Hermans, D. (2013). Fear extinction and relapse: state of the art. Annual Review of Clinical Psychology, 9, 215e248. Walter, B., Blecker, C., Kirsch, P., Sammer, G., Schienle, A., Stark, R., et al. (2003). MARINA: an easy to use tool for the creation of masks for region of interest analyses. In 9th
122
c o r t e x 7 9 ( 2 0 1 6 ) 1 1 2 e1 2 2
International Conference on Functional Mapping of the Human Brain (Vol. 19, No. 2). Available on CD-Rom in NeuroImage. Warren, V. T., Anderson, K. M., Kwon, C., Bosshardt, L., Jovanovic, T., Bradley, B., et al. (2014). Human fear extinction and return of fear using reconsolidation update mechanisms: the contribution of on-line expectancy ratings. Neurobiology of Learning and Memory, 113, 165e173. Weike, A. I., Hamm, A. O., Schupp, H. T., Runge, U., Schroeder, H. W. S., & Kessler, C. (2005). Fear conditioning following unilateral temporal lobectomy: dissociation of
conditioned startle potentiation and autonomic learning. The Journal of Neuroscience, 25(48), 11117e11124. Weike, A. I., Schupp, H. T., & Hamm, A. O. (2007). Fear acquisition requires awareness in trace but not delay conditioning. Psychophysiology, 44(1), 170e180. Wendt, J., Neubert, J., Lindner, K., Ernst, F. D., Homuth, G., Weike, A. I., et al. (2014). Genetic influences on the acquisition and inhibition of fear. International Journal of Psychophysiology. http://dx.doi.org/10.1016/j.ijpsycho.2014.10.007.
Available online at www.sciencedirect.com
ScienceDirect Journal homepage: www.elsevier.com/locate/cortex
Research report
No evidence for blocking the return of fear by disrupting reconsolidation prior to extinction learning Tim Klucken a,b,*, Onno Kruse a,b, Jan Schweckendiek a,b, Yvonne Kuepper c, Erik M. Mueller d, Juergen Hennig c and Rudolf Stark a,b a
Department of Psychotherapy and Systems Neuroscience, Justus Liebig University Giessen, Germany Bender Institute of Neuroimaging, Justus Liebig University Giessen, Germany c Personality Psychology and Individual Differences, Justus Liebig University Giessen, Germany d Department of Clinical Psychology, Justus Liebig University Giessen, Germany b
article info
abstract
Article history:
Fear extinction is a central model for the treatment of anxiety disorders. Initial research
Received 3 June 2015
has reported that the single presentation of a conditioned stimulus prior to extinction
Revised 4 October 2015
learning can permanently block the return of fear. However, only few studies have
Accepted 16 March 2016
explored this issue and could not always replicate the findings.
Reviewed 10 July 2015
The present study examined human fear extinction using a four-day design. On the first
Action editor Peter Kirsch
day, two neutral stimuli were paired with electrical stimulation (UCS), while a third
Published online 24 March 2016
stimulus (CS) was not. Twenty-four hours later, one conditioned stimulus (CSþrem) and the CS were reminded once, 10 min before extinction learning, while the other condi-
Keywords:
tioned stimulus (CSþnon-rem) was not presented prior to extinction learning. All stimuli
Extinction
were presented during extinction learning and during two re-extinction sessions (24 h and
Fear
6-months after extinction learning) without reinforcement. Blood oxygen level-dependent
Reconsolidation
(BOLD) responses and skin conductance responses (SCRs) to both CSþ and the CS were
Amygdala
explored during acquisition, extinction, and in both re-extinction sessions.
Fear conditioning COMT
Regarding SCRs, the results showed that a single presentation of a conditioned stimulus did not block the return of fear during re-extinction: Fear recovery during re-extinction (24 h and 6-months after extinction learning) was observed for both CSþ compared with the CS with no difference between CSþrem and CSþnon-rem. Regarding BOLD-responses, no significant differences between CSþrem and CSþnon-rem were found in region of interest (ROI)-analyses (amygdala, ventromedial prefrontal cortex) during extinction learning and both re-extinction sessions. Whole-brain analyses showed increased BOLD-responses to the CSþnon-rem as compared to the CSþrem in several regions (e.g., middle frontal gyrus) during extinction learning and re-extinction (24 h after extinction learning).
* Corresponding author. Department of Psychotherapy and Systems Neuroscience, Justus Liebig University Giessen, Otto-Behaghel-Str. 10 F, 35394 Giessen, Germany. E-mail address: [email protected] (T. Klucken). http://dx.doi.org/10.1016/j.cortex.2016.03.015 0010-9452/© 2016 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http:// creativecommons.org/licenses/by-nc-nd/4.0/).
c o r t e x 7 9 ( 2 0 1 6 ) 1 1 2 e1 2 2
113
The present findings suggest that the effect of preventing the return of fear by disrupting reconsolidation seems to be a more labile phenomenon than previously assumed. Possible boundary conditions and implications are discussed. © 2016 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
1.
Introduction
Fear conditioning and extinction are well-established models for the development, maintenance, and treatment of anxiety disorders (Goode & Maren, 2014; Milad & Quirk, 2012; Milad, Rosenbaum, & Simon, 2014; Vervliet, Craske, & Hermans, 2013). While fear associations can be rapidly acquired and persist over time, the extinction memory is susceptible to disruptions and cannot permanently block the initial fear memory (Myers & Davis, 2007). Consequently, treatments of psychiatric disorders based on extinction learning (e.g., exposure therapy) often produce effective short-term fear reductions, but relapses are not uncommon (Choy, Fyer, & Lipsitz, 2007; Lipsitz, Mannuzza, Klein, Ross, & Fyer, 1999; Sharma, Thennarasu, & Janardhan Reddy, 2014). Therefore, the identification of specific factors that may decrease the return of fear and the number of relapses are of high clinical interest. Differential fear conditioning paradigms typically consist of different phases (fear acquisition, extinction learning, and reextinction). During fear acquisition, one or two neutral stimuli (CSþ) are initially paired with electrical stimulation (UCS), while another stimulus (CS) is not. After a few trials, the CSþ elicits conditioned responses (CRs) such as increased skin conductance responses (SCRs), startle amplitude, or subjective € ratings (Hamm & Weike, 2005; Lang, Davis, & Ohman, 2000). After that, the CSþ and CS are repeatedly presented without the UCS (extinction learning), which finally results in a decrease of CRs in subjective and physiological responses (Milad & Quirk, 2012; Myers & Davis, 2007; Quirk & Mueller, 2008). During this time, the extinction memory is mainly modulated by the amygdala (Quirk & Mueller, 2008). After that, the CSþ and the CS are again presented without reinforcement (re-extinction), e.g., 24 h after extinction learning. The return of fear could be observed under a variety of conditions such as spontaneous recovery, reinstatement, and renewal (Bouton, 2002; Myers & Davis, 2002, 2007). Spontaneous recovery can be described as the reappearance of previously extinguished CRs after a delay following extinction learning without any further learning sessions due to the mere passage of time. Reinstatement refers to the reoccurrence of CRs after extinction learning through the presentation of an unpredictable UCS. Finally, renewal refers to the reactivation of CRs if a subsequent test session is conducted in a different context than the extinction phase. Many methods have been developed to analyze the reoccurrence of CRs of fear during re-extinction. While some studies compared the responses towards the CSþ and the CS during the first half or the first trial of re-extinction, others authors calculated different “fear-recovery indices” (e.g., first re-extinction trial minus last extinction learning trial; Schiller, Kanen, LeDoux, Monfils, & Phelps, 2013; Schiller et al., 2010).
Recently, animal and human studies have demonstrated that the re-occurrence of conditioned fear during reextinction can be prevented by different techniques, which presumably alter the initial fear memory (Agren, 2014; Agren, Furmark, Eriksson, & Fredrikson, 2012; Johnson & Casey, 2015; Kindt, Soeter, & Vervliet, 2009; Liu et al., 2014; Nader, Schafe, & Le Doux, 2000; Schiller et al., 2010, 2013; Warren et al., 2014). A frequently used technique in human studies is the presentation of a previously conditioned stimulus (CSþrem) 10 min prior to extinction learning without reinforcement, while the other conditioned stimulus (CSþnon-rem), also previously paired with the UCS during fear acquisition, is not presented prior to extinction learning (Schiller et al., 2010, 2013). It has been suggested that this single presentation of the CSþrem reactivates the original CSþ/UCS memory, which enables a new “CSþ/no-UCS” association during extinction learning to be permanently incorporated (Agren, 2014; Schiller et al., 2010). Influential studies have demonstrated successful blocking of the return of fear to the CSþrem as compared to the CSþnon-rem during re-extinction when using this procedure (Schiller et al., 2010, 2013). Regarding the underlying neural correlates, a previous study found increased activity in the ventromedial prefrontal cortex (vmPFC) and altered effective connectivity to the CSþnon-rem compared to the CSþrem during extinction (Schiller et al., 2013). Regarding re-extinction, increased amygdala responses to the CSþnon-rem were found compared to the CSþrem, which has been assumed as an indicator for fear responses (Agren, 2014; Schiller et al., 2013). However, other studies using similar paradigms could not replicate these promising findings (Golkar, Bellander, Olsson, & Ohman, 2012; Kindt & Soeter, 2013; Soeter & Kindt, 2011). They showed a return of fear to both CSþ and could not find any differences between the CSþrem and the CSþnon-rem. In a recent review, Agren (2014) hypothesized that these contrary results might be due to specific subgroups in which the blocking is effective. It was argued that the Val158Met-polymorphism in Catechol-O-Methyl-Transferase (COMT) is of special interest, because recent studies have been able to show an association of the COMT Val158Met-polymorphism with fear acquisition and extinction learning (Agren, Furmark et al., 2012; Lonsdorf et al., 2009; Wendt et al., 2014). For instance, Lonsdorf and colleagues showed deficits in extinction learning as well as a poorer treatment outcome in extinction-based therapy in Met/Met individuals (Lonsdorf et al., 2009, 2010, 2011). However, no study has investigated the association between the COMT Val158Met-polymorphism and delayed extinction recall or return of fear. Based on the above-mentioned findings, the present study aimed to investigate the following: First, we investigated potential SCR differences between CSþrem and CSþnon-rem during
114
c o r t e x 7 9 ( 2 0 1 6 ) 1 1 2 e1 2 2
extinction learning and re-extinction. Second, we explored the underlying neural correlates of processing the CSþrem compared with the CSþnon-rem. As a supplement, we also explored the potential association between the COMT Val158Met-polymorphism, fear acquisition, extinction learning, and the return of fear. SCRs and neural activity (blood oxygen level-dependent e BOLD signal change) were measured during all phases (fear acquisition, extinction learning, and reextinction). We hypothesized augmented SCRs to the CSþnon-rem compared to the CSþrem during re-extinction. Regarding the neural correlates, increased vmPFC activity was expected in the contrast CSþnon-rem versus CSþrem during extinction learning and increased amygdala responses to the CSþnon-rem compared to the CSþrem during re-extinction.
2.
Material and methods
2.1.
Participants
One hundred and thirty six Caucasian subjects (75 females; MAge ¼ 23.93; SDAge ¼ 4.15) participated in this study. All were native German speakers, right-handed, and had normal or corrected-to-normal vision. Subjects reporting current or past mental illnesses, chronic diseases, or a consumption of psychotropic drugs were excluded. All subjects gave written informed consent and received 50 V for their participation. We restricted our sample to individuals showing reliable SCRs (>.02 ms to one UCS-presentation) and increased SCRs to the CSþ as compared to the CS during fear acquisition (mean differential SCRs >.02 ms). This criterion reduced the sample size to 70 (23 Met/Met, 28 Met/Val, 19 Val/Val) participants. The exclusion ratio is comparable to that of other studies (Agren, Engman et al., 2012; Agren, Furmark et al., 2012; Schiller et al., 2013). As a supplement, we also conducted an additional analysis on the entire sample (without any SCR exclusion criteria). The additional analysis yielded comparable results. Eight subjects could not be recruited again for the fourth day after six months, leading to a sample size of 62 subjects for this day (19 Met/Met, 26 Met/Val, 17 Val/Val). There was no significant deviation from HardyeWeinberg equilibrium [c2(1) < .1.60; p > .2]. Subjects were also interviewed using a self-developed standardized interview for stressful life events and for psychiatric disorders in the previous months on the first experimental and on the last experimental day. Several power analyses were conducted to estimate the sample size necessary to adequately address the comparison between CSþrem and CSþnon-rem. A conservative threshold of 95% sensitivity (1beta ¼ .95) and a significance level of a ¼ .05 were set for the power analyses based on the results by Schiller et al. (2010). The required sample size was below 23. The study was conducted in accordance with the Declaration of Helsinki and was approved by the institutional ethics committee.
2.2.
Fear acquisition and extinction paradigm
The fear acquisition and extinction protocol consisted of four subsequent phases (day 1: fear acquisition; day 2: extinction
learning; day 3: 24 h re-extinction; day 4: 6-month reextinction). SCRs and BOLD-responses were measured during all phases. On the first day (fear acquisition), 16 trials of each CSþ (CSþrem; CSþnon-rem; blue and yellow squares) or 16 trials of the CS were presented. Each trial started with the presentation of a fixation cross with a variable duration (0e2.5 sec) to achieve a stimulus-onset-asynchrony. After that, a CS (CSþrem; CSþnon-rem; CS) was presented for 8 sec. The UCS (duration: 100 msec) was delivered 7.9 sec after the CSþ onset and co-terminated with the CSþ offset, while the CS was never associated with the UCS. In contrast to a 38% partial reinforcement rate, which was used by Schiller et al. (2010), we used a 50% reinforcement rate (cf. Golkar et al., 2012). A fixation cross was presented during the intertrial-interval (ITI). The duration of the ITI depended on the stimulus-onset-asynchrony (5.5e8 sec), because the length of a trial was always 16 sec. A black computer screen was presented at any other time of the experiment. The instructions were to pay attention to the computer screen and try to figure out the relationship between the CS and the UCS. On the second day (extinction learning), the CSþrem and the CS were presented once in a counterbalanced order without reinforcement 10 min before the extinction learning started, while the other CSþ (CSþnon-rem) was not reminded (Schiller et al., 2010). Subjects had to watch a movie about landscapes (“Colours of Earth e Faszination Natur 2”) during the 10min break. Extinction learning consisted of 11 presentations of the CSþnon-rem and 10 presentations of the CSþrem and the CS to ensure an equal number of presentations during extinction learning, because the CSþnon-rem had not been reminded before. On day three (24 h after extinction learning) and day four (6-months after extinction learning), a reinstatement procedure was conducted inside the scanner to reactivate the fear memory before the re-extinction session started. Accordingly, the time that elapsed between the last extinction learning trial and the first reinstatement trial was around 24 h for day three and around 6-months for day four. During reinstatement, five unsignalled UCS (US to US interval: 13.5e18.5 sec) were applied, while a black computer screen was presented. Around 2 min after the last UCS application of the reinstatement, the re-extinction sessions (10 trials of each CS) started without reinforcement. A pseudorandomized stimulus order was used on each experimental day, which ensured the presentation of all CS with equal frequency in the first and the second half, with no more than two presentations in succession and an equal frequency of all CS for the first trial. In addition, the UCS was presented equally often in the first and in the second half of day one. The UCS intensity was set individually using a gradually increasing procedure to achieve an “unpleasant but not painful” level of sensation. No UCS-recalibration was made on day 2, 3, and 4, to ensure the same number of UCS applications for each subject. A custom-made impulse-generator (833 Hz; 5 mA) provided transcutaneous electrical stimulation (UCS) through two Ag/AgCl electrodes (1 mm2 surface), which was triggered via an optic fiber cable. Electrodes were fixed to the middle of the left shin.
c o r t e x 7 9 ( 2 0 1 6 ) 1 1 2 e1 2 2
2.3.
Skin conductance measuring
SCRs were sampled using Ag/AgCl electrodes filled with isotonic (.05 M NaCl) electrolyte medium placed at the nondominant left hand. An SCR was defined as the highest phasic response following stimulus onset. Therefore, the largest difference between a minimum, which had to occur within 1e8 sec after onset of the CS, and the subsequent maximum was extracted using Ledalab 3.4.4 (Benedek & Kaernbach, 2010). SCRs were ln (mS þ 1), corrected for violation of normal distribution of the data. Outlier responses (SCRs ± 3 standard deviations) were excluded. Mean SCRs to CSþ and CS were analyzed via analysis of variance (ANOVA) with the factors CS-type (CSþrem, CSþnonrem, CS) number of trials (e.g., 16 in fear acquisition), followed by Bonferroni-corrected post-hoc tests, using SPSS 22 (SPSS Inc. Chicago, Illinois, USA). Furthermore, the fear recovery indexes (SCR of first re-extinction trial minus SCR of last extinction learning trial) of the CSþrem with the CSþnonrem were compared to analyze potential differences between both CSþ. In addition, we also investigated the first trial separately (SCR of the first trial of the CSþrem minus SCR of the first trial of the CSþnon-rem), because Schiller et al. (2013) found the greatest differences between both CSþ in this trial. Finally, in line with other studies (Golkar et al., 2012; Schiller et al., 2010, 2013), we also conducted these analyses with the first interval response only (1e4 sec after CS onset) but results were comparable.
2.4.
Magnetic resonance imaging
Subjects were scanned using a 1.5 T whole-body tomograph (Siemens Symphony with a quantum gradient system) with a standard head coil. 160 T1-weighted images (MPRage, 1 mm slice thickness) were acquired in sagittal orientation. Functional imaging consisted of 420 images in a T2*-weighted echo-planar imaging (EPI) sequence. For functional images, a T2*-weighted gradient EPI sequence was used with 25 slices covering the whole brain (slice thickness ¼ 5 mm; 1 mm gap; descending slice procedure; TR ¼ 2.5 sec; TE ¼ 55 msec; flip angle ¼ 90 ; field of view ¼ 192 192 mm; matrix size ¼ 64 64). The orientation of the axial slices was paralleled to the orbitofrontal cortex-bone transition in order to minimize susceptibility artefacts in prefrontal areas. Prior to all statistical analyses, data were preprocessed as described previously (Klucken, Kruse et al., 2015). The functional data were analyzed using a random-effects general linear model. For each day, separate first level models were calculated. In line with Schiller et al. (2013), separate boxcar regressors (CS duration) for each CS were calculated for each model of extinction learning and both re-extinction sessions (24 h and 6-months after extinction learning) and were divided into an early (first half) and a late (second half) phase, because Schiller et al. (2013) showed the greatest differences in the early, but not the late phase during extinction learning and re-extinction. We also conducted an alternative model using a stick function, but results were comparable. On the first day, regressors for the UCS and the non-UCS (time window corresponding to the UCS after the CS) were also introduced in the model (maximum dependency between
115
CSþ and UCS < .20). The six movement parameters estimated in the preprocessing realignment step were entered into all first level models. The voxel-based time series was filtered with a high pass filter (time constant ¼ 128 sec). On the group level, a 3 (CSþrem, CSþnon-rem, CS) 2 (early phase, late phase) ANOVA was computed in SPM8 to explore main effects of CS-type, phase, and CS-type by phase interaction effects. The threshold for whole-brain analyses (for extinction learning and re-extinction) was set to p < .001 and k > 10 to detect potential differences with a liberal criterion. We focused our region of interest (ROI) analyses on the amygdala and the vmPFC, because (1) both of these structures are crucially involved in fear acquisition, extinction learning, and re-extinction; and (2) previous studies found significant differences between CSþrem and CSþnon-rem in these two structures only (Agren, Engman et al., 2012; Schiller et al., 2013). ROI analyses were performed with a threshold of p < .05 (family-wise-error; FWE-corrected) and k > 5. The amygdala mask was taken from the “Harvard Oxford cortical and subcortical structural atlases” provided by the Harvard Center for Morphometric Analysis. The vmPFC mask was created with MARINA (Walter et al., 2003) and has been used in previous studies (e.g., Hermann, Keck, & Stark, 2014; Klucken, Schweckendiek et al., 2015; Klucken, Schweckendiek, Merz, Vaitl, & Stark, 2013). An additional analysis was conducted exploring differences between the CSþrem and the CSþnon-rem within the first trial during re-extinction. This procedure was chosen because the comparison of the first trial of the CSþrem and the CSþnonrem is of special interest, due to previous findings that showed differences between both CSþ in the first trial (parallel to the SCR-analyses of re-extinction; cf. Schiller et al., 2013). Therefore, we specified a new first level model for each subject including all trials as separate regressors (i.e., CSþrem_trial_1, CSþrem_trial_2,…, CSþnon-rem_trial_1, CSþnon-rem_trial_2,…). This results in 3 10 conditions (10 trials for the CSþrem, CSþnon-rem, and for the CS). Next, contrasts were created for each comparison of interest (e.g., CSþrem_trial_1 CSþnonrem_trial_1) and analyzed on the second level with the same whole-brain and ROI-analyses parameters as described above.
3.
Results
3.1.
Fear acquisition (day 1)
We will only briefly describe the fear acquisition results to ensure that CSþrem and CSþnon-rem did not differ, because the primary aim of the present study was to explore extinction learning and re-extinction. In addition, we did not expect significant differences, because an inclusion criterion for this study was successful conditioning on the first day.
3.1.1.
SCRs
The ANOVA revealed a main effect of CS-type [F(2, 68) ¼ 55.34; p < .001], trial [F(1, 69) ¼ 31.88; p < .001], as well as a CS-type x trial [F(2, 68) ¼ 13.29; p < .001] interaction effect. The post-hoc tests showed increased SCRs to the CSþrem and to the CSþnon-rem as compared to the CS (Fig. 1), but no significant differences between the two CSþ.
116
c o r t e x 7 9 ( 2 0 1 6 ) 1 1 2 e1 2 2
Fig. 1 e Mean skin conductance responses (ln (1 þ ms)) for the CSþrem, CSþnon-rem, and the CS¡ for each trial for the first day (fear conditioning) and the second day (extinction learning). *indicates significantly (p < .05) increased responses as compared to the CS¡ during extinction learning.
3.1.2.
Hemodynamic responses
We found a main effect of CS-type, showing increased amygdala responses to the CSþrem (x/y/z ¼ 30/4/26; z ¼ 4.07; p ¼ .002) and to the CSþnon-rem (x/y/z ¼ 30/4/26; z ¼ 3.74; p ¼ .006) as compared to the CS. We did not find increased left (z ¼ 2.15; pFWE-corr ¼ .24) or right (z ¼ 2.34; pFWE¼ .18) amygdala activations in the contrast corr CSþrem CSþnon-rem or vice versa (pFWE-corr > .50).
3.2.
Extinction learning (day 2)
3.2.1.
SCRs
The main research question was whether differences between the CSþrem and the CSþnon-rem would occur during extinction learning and during re-extinction. The ANOVA revealed a main effect of CS-type [F(2, 68) ¼ 10.14; p < .001] and trial [F(9, 61) ¼ 42.32; p < .001], showing increased SCRs to the CSþrem and to the CSþnon-rem as compared to the CS, but no differences between both CSþ (p > .20). In addition, we focused on the first trial of the CSþrem and the CSþnon-rem, because Schiller et al. (2013) found the greatest differences for the first trial. We found increased SCRs to both CSþ as compared to the CS (all p < .05), but no differences between the CSþrem and the CSþnon-rem (p ¼ .969; effect size: r ¼ .002; Fig. 1).
3.2.2.
Hemodynamic responses
Because the aim of the study was to explore potential differences between the CSþrem and the CSþnon-rem, we conducted whole-brain analyses as well as ROI-analyses and computed the contrasts CSþrem CSþnon-rem and CSþnon-rem CSþrem. In addition, because previous studies also found differences in the early half of extinction learning (Schiller et al., 2013), we also analyzed BOLD-responses during the early phase of extinction learning. Regarding ROI-analyses, we did not find significant BOLDresponses during the whole extinction phase for the contrast CSþrem CSþnon-rem in the amygdala (“peak voxel”: x/y/z ¼ 27/2/26; z ¼ 1.56; pFWE-corr ¼ .558) or in the vmPFC (x/y/ z ¼ 12/62/2; z ¼ 1.34; pFWE-corr ¼ .909). In addition, no significant activation was found for the opposite contrast (CSþnon-rem CSþrem) in the amygdala (x/y/z ¼ 12/7/17; z ¼ 1.45; pFWE-corr ¼ .585) or in the vmPFC (x/y/z ¼ 9/29/23;
z ¼ 2.34; pFWE-corr ¼ .449). Moreover, ROI-analyses showed no increased activation during the early phase of extinction learning for the contrast CSþrem CSþnon-rem in the amygdala (x/y/z ¼ 18/10/11; z ¼ 1.99; pFWE-corr ¼ .355), the vmPFC (x/y/ z ¼ 9/29/23; z ¼ 2.34; pFWE-corr ¼ .449), and for the opposite contrast (amygdala: x/y/z ¼ 24/7/20; z ¼ 1.96; pFWEcorr ¼ .350); vmPFC (x/y/z ¼ 30/7/20; z ¼ 1.19; pFWE-corr ¼ .698). Finally, ROI-analyses did not show significant differences in the contrasts CSþrem CS and CSþnon-rem CS. Whole-brain results did not show increased activations to the CSþrem compared with the CSþnon-rem (or vice versa) during the whole extinction phase (all puncorrected > .001), but increased BOLD-responses in several regions in the contrast CSþnon-rem CSþrem in the early half of extinction learning and increased activations of the CSþrem and the CSþnon-rem as compared with the CSe (Table 1).
3.3.
Re-extinction (day 3 and day 4)
3.3.1.
SCRs
Day 3 (24-h after extinction learning). The ANOVA showed a significant main effect of CS-type [F(2, 68) ¼ 6.35 p < .01] and trial [F(9, 61) ¼ 16.01; p < .001], with increased SCRs in the first trial to both CSþ, which diminished over time. We also calculated the recovery index and found increased recovery for the CSþrem and the CSþnon-rem (all p < .05). Notably, no significant differences could be found between the CSþrem and the CSþnon-rem in the first trial (p ¼ .269; effect size: r ¼ .07) or for the recovery index (p ¼ .46; effect size: r ¼ .04). Day 4 (6-months after extinction learning). The ANOVA revealed a significant main effect of time [F(9, 53) ¼ 14.03; p < .001] and a CS-type time trend [F(18, 44) ¼ 1.65; p ¼ .084]. In addition, increased SCRs to both CSþ as compared to the CS could also be found in the first re-extinction trial as well as an increased recovery index to both CSþ (all p < .05). Again, no significant differences were observed between the CSþrem and the CSþnon-rem in the first trial (p ¼ .230; effect size: r ¼ .068), or for the recovery index (p ¼ .461; effect size: r ¼ .08). In addition, we also investigated the differences between CSþrem and CSþnon-rem including all subjects, but results were comparable (see supplement for details).
117
c o r t e x 7 9 ( 2 0 1 6 ) 1 1 2 e1 2 2
Table 1 e Localization and statistics for whole-brain results for day 2 (extinction learning). Contrast Whole phase
CSþrem CSþnon-rem CSþnon-rem CSþrem CSþrem CS
CSþnon-rem CS Early phase
CSþrem CSþnon-rem CSþnon-rem CSþrem
CSþrem CS
CSþnon-rem CS
Structure
Side
Occipital cortex Supramarginal gyrus Insula Middle temporal gyrus Middle temporal gyrus Insula Occipital cortex Insula
L L L R L R L R
Temporal gyrus Middle frontal gyrus Orbitofrontal cortex Occipital cortex Insula Insula Middle temporal gyrus Supramarginal gyrus Limbic lobe Cerebellum Occipital cortex Orbitofrontal cortex Supramarginal gyrus Insula Limbic lobe Middle temporal gyrus Inferior temporal gyrus
L L R R R L R L L R L R L L R R L
k
x
y
No significant differences No significant differences 663 3 88 44 60 31 32 45 8 21 45 67 12 48 64 24 48 5 719 3 91 10 36 20 No significant differences 22 66 31 11 42 23 13 3 32 390 12 85 191 33 23 220 36 14 38 48 64 81 66 25 37 3 14 12 9 40 350 3 94 127 33 23 79 54 28 22 33 26 15 12 25 10 66 28 16 42 40
z
zmax
p
1 25 7 7 7 1 1 8
6.25 3.90 3.90 3.60 3.46 3.45 5.85 3.33
<.001 <.001 <.001 <.001 <.001 <.001 <.001 <.001
14 52 11 7 7 7 10 28 22 8 1 8 25 4 34 14 26
3.96 3.69 3.62 4.90 4.65 4.32 4.02 4.03 3.79 3.67 4.48 4.38 4.06 3.64 3.59 3.58 3.51
<.001 <.001 <.001 <.001 <.001 <.001 <.001 <.001 <.001 <.001 <.001 <.001 <.001 <.001 <.001 <.001 <.001
The threshold was p < .001 (uncorrected; whole-brain results according to SPM8). All coordinates are given in MNI space. L: left hemisphere, R: right hemisphere.
3.3.2.
Hemodynamic responses
Day 3 (24-h after extinction learning). Regarding ROI-analyses, no increased activation could be found for the contrast CSþrem CSþnon-rem during the whole phase in the amygdala or in the vmPFC (no suprathreshold clusters). Moreover, no increased ROI-activation was found for the opposite contrast (amygdala: x/y/z ¼ 18/13/14; z ¼ 2.73; pFWE-corr ¼ .097; vmPFC: x/y/z ¼ 3/59/23; z ¼ 3.12; pFWE-corr ¼ .105). In addition, no significant BOLD-responses could be found to the CSþrem compared to the CSþnon-rem during the early phase (amygdala: x/y/z ¼ 21/7/11; z ¼ .20; pFWE-corr ¼ .837; vmPFC: x/y/z ¼ 9/29/14; z ¼ 1.30; pFWE-corr ¼ .943). In addition, we did not find any significantly increased ROI-activations for the opposite contrast (amygdala: x/y/z ¼ 24/4/23; z ¼ 2.13; pFWE-corr ¼ .304; vmPFC: x/y/z ¼ 0/62/14; z ¼ 1.91; pFWEcorr ¼ .755). No significant differences were found in ROIanalyses in the contrasts CSþrem CS and CSþnonrem CS. Finally, no significant differences between CSþrem CSþnon-rem (or vice versa) were found in all ROI when only comparing activations in the first trial. Increased whole-brain results were found for the contrast CSþnon-rem CSþrem in the whole phase and in the first trial, but not for the early half of re-extinction, while the opposite contrast showed only significant (uncorrected) hippocampal activations in the first trial (see Table 2). Day 4 (6-months after extinction learning). ROI-analyses showed no increased activation for the contrast
CSþrem CSþnon-rem during the whole phase (amygdala: x/y/ z ¼ 30/1/20; z ¼ 1.02; pFWE-corr ¼ 745; vmPFC: x/y/z ¼ 6/44/ 1; z ¼ 3.20; pFWE-corr ¼ .081). In addition, we did not find any increased ROI-activation for the opposite contrast (amygdala: x/y/z ¼ 18/7/20; z ¼ .97; pFWE-corr ¼ .757; vmPFC: x/y/ z ¼ 12/68/2; z ¼ 1.25; pFWE-corr ¼ .931). Moreover, no significant BOLD-responses to the CSþrem compared to the CSþnonrem during the early phase could be observed in the amygdala (x/y/z ¼ 15/4/14; z ¼ .74; pFWE-corr ¼ .804) or in the vmPFC (x/y/z ¼ 9/44/8; z ¼ 2.83; pFWE-corr ¼ .197). In addition, we did not find any significantly increased ROI-activations for the opposite contrast (amygdala: x/y/z ¼ 12/7/17; z ¼ 2.02; pFWEcorr ¼ .352; vmPFC: x/y/z ¼ 15/29/17; z ¼ 1.58; pFWE-corr ¼ .874). Regarding the first trial, no significant differences between both CSþ were found in ROI-analyses. Finally, no significant ROI-activations were found in the contrasts CSþrem CS and CSþnon-rem CS. Whole-brain analyses showed increased BOLD-responses to the CSþrem as compared to the CSþnon-rem only in the precuneus in the whole phase and in the occipital cortex, precuneus, and supramarginal gyrus in the first trial and pronounced activation to the CSþrem as compared to the CS in the occipital cortex (Table 3; Fig. 2).
3.4.
COMT Val158Met-polymorphism
A detailed results section and a description of the genotyping and data analyses are included in the supplement. In short, regarding SCRs, neither main effects nor interactions with the
118
c o r t e x 7 9 ( 2 0 1 6 ) 1 1 2 e1 2 2
Table 2 e Localization and statistics for whole-brain results for day 3 (re-extinction). Contrast Whole phase
CSþrem CSþnon-rem CSþnon-rem CSþrem
CSþrem CS CSþnon-rem CS
Early phase
First trial
CSþrem CSþnon-rem CSþnon-rem CSþrem CSþrem CS CSþnon-rem CS CSþrem CSþnon-rem CSþnon-rem CSþrem
Structure
Side
Postcentral gyrus Precentral gyrus Middle temporal gyrus No distinct region Occipital cortex Occipital cortex Precuneus Inferior temporal gyrus
R R L R R R R R
Occipital cortex Precuneus Middle temporal gyrus Precentral gyrus Superior parietal gyrus Hippocampus
L L R R R L
k
x
Y
No significant differences 25 48 19 17 12 28 12 60 7 59 21 2 260 12 82 782 12 82 132 15 58 89 51 64 No significant differences No significant differences 283 6 82 16 21 49 24 48 67 18 27 19 23 27 58 14 27 19
z
zmax
p
34 73 26 16 7 10 22 5
3.76 3.72 3.67 3.90 4.56 5.08 4.09 4.00
<.001 <.001 <.001 <.001 <.001 <.001 <.001 <.001
7 4 16 70 64 17
4.96 3.58 3.55 3.69 3.65 3.81
<.001 <.001 <.001 <.001 <.001 <.001
The threshold was p < .001 (uncorrected; whole-brain results according to SPM8). All coordinates are given in MNI space. L: left hemisphere, R: right hemisphere.
Table 3 e Localization and statistics for whole-brain results for day 4 (re-extinction). Contrast Whole phase
Early phase
First trial
CSþrem CSþnon-rem CSþnon-rem CSþrem CSþrem CS CSþnon-rem CS CSþrem CSþnon-rem CSþnon-rem CSþrem CSþrem CS CSþnon-rem CS CSþrem CSþnon-rem
Structure
Side
k
Precuneus
L
Occipital cortex Occipital cortex Occipital cortex
L R R
Occipital cortex Occipital cortex
L R
Occipital cortex Precuneus Supramarginal gyrus
R R R
44 No 81 75 33 No No 31 21 No 45 81 17 No
CSþnon-rem CSþrem
x
y
3 61 significant differences 12 85 15 73 3 85 significant differences significant differences 9 85 21 73 significant differences 36 82 2 70 39 40 significant differences
z
zmax
p
19
3.66
<.001
4 13 4
4.33 4.18 3.72
<.001 <.001 <.001
7 13
3.53 3.66
<.001 <.001
22 31 43
4.96 3.91 3.76
<.001 <.001 <.001
The threshold was p < .001 (uncorrected; whole-brain results according to SPM8). All coordinates are given in MNI space. L: left hemisphere, R: right hemisphere.
COMT Val158Met-polymorphism were observed during fear acquisition, extinction learning, and re-extinction. BOLD-responses also showed no significant amygdala and vmPFC differences during fear acquisition, extinction learning, and re-extinction with respect to the COMT Val158Metpolymorphism.
4.
Discussion
The present study explored the impact of a single CSþ presentation prior to extinction learning on re-extinction. SCRs and BOLD-responses between the reminded and nonreminded CSþ were compared during all phases. The results revealed similarly increased SCRs to the CSþrem and to the CSþnonrem as compared to the CS during extinction learning and re-extinction in the present study. Consistent with the lack of CSþrem versus CSþnon-rem differences in SCRs, ROI-
analyses also showed no significant differences between the both CSþ during extinction learning and re-extinction. Moreover, no association with the COMT Val158Met-polymorphism was observed. The present findings could not replicate previous results (Agren, Engman et al., 2012; Agren, Furmark et al., 2012; Schiller et al., 2013; Schiller et al., 2010), which had suggested that a single presentation of a CSþ prior to extinction learning had a substantial impact on re-extinction. Yet, our results support other studies also showing no differences between the CSþrem and the CSþnon-rem during re-extinction (Golkar et al., 2012; Kindt & Soeter, 2013; Soeter & Kindt, 2011). In order to explain these inconsistent findings, different boundary conditions are discussed, which might have influenced the results. First, it could be assumed that differences in the experimental design such as CS duration, ITI, exact reinstatement procedure, UCS-recalibration, etc. might alter the effects
c o r t e x 7 9 ( 2 0 1 6 ) 1 1 2 e1 2 2
119
Fig. 2 e Mean skin conductance responses (ln (1 þ ms)) for the CSþrem, CSþnon-rem, and the CS¡ for each trial during reextinction (day 3 and day 4). Recovery index (below) for re-extinction for the CSþrem and the CSþnon-rem. *indicates significantly (p < .05) increased responses (return of fear) as compared to the CS¡.
(Agren, 2014; Golkar et al., 2012). For instance, Schiller et al. (2010) and the present study presented the CSþrem and the CS prior to extinction learning, while other studies presented the CSþrem before extinction learning only. In a recent review, Haaker, Golkar, Hermans, and Lonsdorf (2014) hypothesized that slight differences in the conditioning design could significantly influence the return of fear. Second, the exclusion ratios in the present study (~50%) as well as the fMRI-study (~73%) by Schiller et al. (2013) were higher compared to other extinction studies (e.g., ~32% in Golkar et al., 2012), which may also impact findings. However, the present results were comparable, if the entire sample was investigated (see Supplement). Nevertheless, independent replications are required to investigate this unexpected exclusion ratio in more detail. Third, it is also possible that different analysis strategies might have led to the heterogeneous findings. A variety of different approaches exist to analyze the return of fear index. While some studies calculated the recovery index based on the first trial of re-extinction and the last extinction learning trial (e.g., Schiller et al., 2013), other studies compared SCRs of the first trial (or the first half) during re-extinction of each CSþ with each other. To account for this, we conducted several analyses to explore differences between both CSþ during the early half only or at a single trial level, but did not find any significant SCR differences between the CSþrem and the CSþnon-rem. However, differences in the analysis methods as well as the comparison of CRs over different days could be problematic, because of habituation effects at the end of day 2 and large arousal effects on day 3 and day 4. In addition, it
should be noted that the used procedure does not allow a distinction of the exact mode of return of fear because reinstatement and spontaneous recovery effects are intermixed (Haaker et al., 2014). Moreover, it is also possible that differences in the ratios of contingency aware and contingency unaware subjects between the studies are responsible for the contrary findings. Previous studies showed that contingency awareness impacts on SCRs and neural activity, which were measured in the present study (Hamm & Weike, 2005; Klucken, Kagerer et al., 2009; Klucken, Tabbert et al., 2009; Mertens et al., 2015; Weike et al., 2005; Weike, Schupp, & Hamm, 2007). However, previous studies as well as the present study did not measure contingency awareness explicitly to prevent a repeated presentation of the CS, which might also have impacted extinction processes. Instead, we instructed subjects to pay attention to potential CS/UCS contingencies, which is in line with previous studies (Schiller et al., 2010, 2013). Finally, Agren (2014) hypothesized that the initial effects may occur in specific subgroups only, which differ in certain genotypes. For instance, an association between fear acquisition and extinction with genetic variation in the serotonin transporter gene (serotonergic transporter-linked polymorphic region; 5-HTTLPR, rs25531) and the Val158Met-polymorphism in the COMT was found, which are both closely related to fear learning, extinction, and updating of previously learned contingencies (Agren, Furmark et al., 2012; Cris‚an et al., 2009; Hermann et al., 2012; Klucken, Alexander et al., 2013; Klucken, Schweckendiek et al., 2015; Klumpers, Heitland, Oosting, Kenemans, & Baas, 2012;
120
c o r t e x 7 9 ( 2 0 1 6 ) 1 1 2 e1 2 2
Lonsdorf et al., 2009, 2011; Wendt, et al., 2014). As a supplement, we reanalyzed our results with respect to the COMT Val158Met-polymorphism, but did not find any significant associations with the COMT Val158Met-polymorphism. However, a previous study also found no relationship between the COMT Val158Met-polymorphism during extinction learning with SCRs, but with conditioned startle amplitude, suggesting that COMT Val158Met-polymorphism might be associated with specific response systems (Lonsdorf et al., 2009). A possible explanation for this selective effect is the assumption that conditioned startle amplitudes and SCRs are based, at least partly, on different neural circuits (Hamm & Weike, 2005; Weike et al., 2005). While conditioned startle amplitude is primarily mediated by amygdala activations (Davis & Whalen, 2001; Hamm & Weike, 2005), conditioned SCRs could be observed without an involvement of the amygdala (Weike et al., 2005; Tabbert, Stark, Kirsch, & Vaitl, 2006). Thus, the potential association of COMT Val158Metpolymorphism on extinction learning and re-extinction might be visible only in predominantly amygdaladependent response systems. Regarding (uncorrected significant) whole-brain results, we found differences between the CSþrem and the CSþnon-rem during extinction learning and re-extinction. For instance, increased BOLD-activations were found to the CSþnon-rem compared to the CSþrem in the orbitofrontal cortex, the middle frontal gyrus during the early phase of extinction learning. During extinction learning, it is necessary to adapt previously learned contingencies to new circumstances. The middle frontal gyrus has been linked to contingency awareness during fear acquisition (Carter, O'Doherty, Seymour, Koch, & Dolan, 2006). It seems possible that the process to change the CS/UCS relationship of the CSþnon-rem may require more explicit effort than the adaptation to the CSþrem, which may have led to the observed effects in this structure. In addition, increased activations to both CSþ in contrast to the CS were also found in prefrontal areas, the limbic lobe, the occipital cortex, and further structures during extinction learning and/ or re-extinction (see tables for detailed results). However, it should be pointed out that these whole-brain analyses are partly unexpected and the threshold for whole-brain analyses was set to a liberal criterion (uncorrected p < .001). Therefore, the results should be treated with caution until an independent replication is available. In conclusion, we would like to point out that the present results do not argue against the phenomenon of blocking the return of fear in general. Various animal and human studies found successful blocking of the return of fear (Agren, 2014; Johnson & Casey, 2015; Kindt et al., 2009; Liu et al., 2014; Nader et al., 2000; Schiller et al., 2010, 2013) and it is also possible that our results might be due to false-negative probability. Nevertheless, the unexpected (null) results may support the view that the reported effects may exist in specific constellations only and may require specific parameters, which are not entirely clear to date (Agren, 2014; Golkar et al., 2012). For instance, Schiller and colleagues also reported data of subjects without successfully blocked return of fear (Schiller et al., 2013). Therefore, future studies may help to give more insight into the impact and boundaries and the opportunity for blocking the return of fear more efficiently.
Nevertheless, the possibility of preventing the return of fear through disrupting reconsolidation is still fascinating and could someday be implemented in treatments for anxiety disorders (Shiban, Bru¨tting, Pauli, & Mu¨hlberger, 2015).
Financial disclosures All authors report no biomedical financial interests or other potential conflicts of interest.
Acknowledgments This study was supported by a research grant from the Deutsche Forschungsgemeinschaft German Research Foundation to the first author (Grant number: Klu 2500/1-1).
Supplementary data Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.cortex.2016.03.015.
references
Agren, T. (2014). Human reconsolidation: a reactivation and update. Brain Research Bulletin, 105, 70e82. € rkstrand, J., Larsson, E., Agren, T., Engman, J., Frick, A., Bjo Furmark, T., et al. (2012). Disruption of reconsolidation erases a fear memory trace in the human amygdala. Science, 337(6101), 1550e1552. Agren, T., Furmark, T., Eriksson, E., & Fredrikson, M. (2012). Human fear reconsolidation and allelic differences in serotonergic and dopaminergic genes. Translational Psychiatry, 2, e76. Benedek, M., & Kaernbach, C. (2010). A continuous measure of phasic electrodermal activity. Journal of Neuroscience Methods, 190(1), 80e91. Bouton, M. E. (2002). Context, ambiguity, and unlearning: sources of relapse after behavioral extinction. Biological Psychiatry, 52, 976e986. Carter, R. M., O'Doherty, J. P., Seymour, B., Koch, C., & Dolan, R. J. (2006). Contingency awareness in human aversive conditioning involves the middle frontal gyrus. NeuroImage, 29(3), 1007e1012. Choy, Y., Fyer, A. J., & Lipsitz, J. D. (2007). Treatment of specific phobia in adults. Clinical Psychology Review, 27(3), 266e286. Cris‚an, L. G., Pana, S., Vulturar, R., Heilman, R. M., Szekely, R., a, B., et al. (2009). Genetic contributions of the serotonin Drug transporter to social learning of fear and economic decision making. Social Cognitive and Affective Neuroscience, 4(4), 399e408. Davis, M., & Whalen, P. J. (2001). The amygdala: vigilance and emotion. Molecular Psychiatry, 6(1), 13e34. Golkar, A., Bellander, M., Olsson, A., & Ohman, A. (2012). Are fear memories erasable? Reconsolidation of learned fear with fearrelevant and fear-irrelevant stimuli. Frontiers in Behavioral Neuroscience, 6, 80. Goode, T. D., & Maren, S. (2014). Animal models of fear relapse. ILAR Journal/National Research Council, Institute of Laboratory Animal Resources, 55(2), 246e258.
c o r t e x 7 9 ( 2 0 1 6 ) 1 1 2 e1 2 2
Haaker, J., Golkar, A., Hermans, D., & Lonsdorf, T. B. (2014). A review on human reinstatement studies: an overview and methodological challenges. Learning and Memory, 21(9), 424e440. Hamm, A. O., & Weike, A. I. (2005). The neuropsychology of fear learning and fear regulation. International Journal of Psychophysiology, 57(1), 5e14. Hermann, A., Keck, T., & Stark, R. (2014). Dispositional cognitive reappraisal modulates the neural correlates of fear acquisition and extinction. Neurobiology of Learning and Memory, 113, 115e124. Hermann, A., Ku¨pper, Y., Schmitz, A., Walter, B., Vaitl, D., Hennig, J., et al. (2012). Functional gene polymorphisms in the serotonin system and traumatic life events modulate the neural basis of fear acquisition and extinction. PLoS One, 7(9), e44352. Johnson, D. C., & Casey, B. J. (2015). Extinction during memory reconsolidation blocks recovery of fear in adolescents. Scientific Reports, 5, 8863. Kindt, M., & Soeter, M. (2013). Reconsolidation in a human fear conditioning study: a test of extinction as updating mechanism. Biological Psychology, 92(1), 43e50. Kindt, M., Soeter, M., & Vervliet, B. (2009). Beyond extinction: erasing human fear responses and preventing the return of fear. Nature Neuroscience, 12(3), 256e258. Klucken, T., Alexander, N., Schweckendiek, J., Merz, C. J., Kagerer, S., Osinsky, R., et al. (2013). Individual differences in neural correlates of fear conditioning as a function of 5HTTLPR and stressful life events. Social Cognitive and Affective Neuroscience, 8(3), 318e325. Klucken, T., Kagerer, S., Schweckendiek, J., Tabbert, K., Vaitl, D., & Stark, R. (2009). Neural, electrodermal and behavioral response patterns in contingency aware and unaware subjects during a picture-picture conditioning paradigm. Neuroscience, 158(2), 721e731. Klucken, T., Kruse, O., Wehrum-Osinsky, S., Hennig, J., Schweckendiek, J., & Stark, R. (2015). Impact of COMT Val158Met-polymorphism on appetitive conditioning and amygdala/prefrontal effective connectivity. Human Brain Mapping, 36(3), 1093e1101. Klucken, T., Schweckendiek, J., Blecker, C., Walter, B., Kuepper, Y., Hennig, J., et al. (2015). The association between the 5-HTTLPR and neural correlates of fear conditioning and connectivity. Social Cognitive and Affective Neuroscience, 10(5), 700e707. Klucken, T., Schweckendiek, J., Merz, C. J., Vaitl, D., & Stark, R. (2013). Dissociation of neuronal, electrodermal, and evaluative responses in disgust extinction. Behavioral Neuroscience, 127(3), 380e386. Klucken, T., Tabbert, K., Schweckendiek, J., Merz, C. J., Kagerer, S., Vaitl, D., et al. (2009). Contingency learning in human fear conditioning involves the ventral striatum. Human Brain Mapping, 30(11), 3636e3644. Klumpers, F., Heitland, I., Oosting, R. S., Kenemans, J. L., & Baas, J. M. P. (2012). Genetic variation in serotonin transporter function affects human fear expression indexed by fearpotentiated startle. Biological Psychology, 89(2), 277e282. € Lang, P. J., Davis, M., & Ohman, A. (2000). Fear and anxiety: animal models and human cognitive psychophysiology. Journal of Affective Disorders, 61(3), 137e159. Lipsitz, J. D., Mannuzza, S., Klein, D. F., Ross, D. C., & Fyer, A. J. (1999). Specific phobia 10-16 years after treatment. Depression and Anxiety, 10(3), 105e111. Liu, J., Zhao, L., Xue, Y., Shi, J., Suo, L., Luo, Y., et al. (2014). An unconditioned stimulus retrieval extinction procedure to prevent the return of fear memory. Biological Psychiatry, 76(11), 895e901.
121
€ m, K. M., Fransson, P., Lonsdorf, T. B., Golkar, A., Lindsto Schalling, M., Ohman, A., et al. (2011). 5-HTTLPR and COMTval158met genotype gate amygdala reactivity and habituation. Biological Psychology, 87(1), 106e112. € m, J., Andersson, G., Ohman, A., Lonsdorf, T. B., Ru¨ck, C., Bergstro Lindefors, N., et al. (2010). The COMTval158met polymorphism is associated with symptom relief during exposure-based cognitive-behavioral treatment in panic disorder. BMC Psychiatry, 10, 99. Lonsdorf, T. B., Weike, A. I., Nikamo, P., Schalling, M., Hamm, A. O., & Ohman, A. (2009). Genetic gating of human fear learning and extinction: possible implications for geneenvironment interaction in anxiety disorder. Psychological Science, 20(2), 198e206. Mertens, G., Kuhn, M., Raes, A. K., Kalisch, R., de Houwer, J., & Lonsdorf, T. B. (2015). Fear expression and return of fear following threat instruction with or without direct contingency experience. Cognition and Emotion, 1e17. Milad, M. R., & Quirk, G. J. (2012). Fear extinction as a model for translational neuroscience: ten years of progress. Annual Review of Psychology, 63, 129e151. Milad, M. R., Rosenbaum, B. L., & Simon, N. M. (2014). Neuroscience of fear extinction: implications for assessment and treatment of fear-based and anxiety related disorders. Behaviour Research and Therapy, 62, 17e23. Myers, K. M., & Davis, M. (2002). Behavioral and neural analysis of extinction. Neuron, 36(4), 567e584. Myers, K. M., & Davis, M. (2007). Mechanisms of fear extinction. Molecular Psychiatry, 12(2), 120e150. Nader, K., Schafe, G. E., & Le Doux, J. E. (2000). Fear memories require protein synthesis in the amygdala for reconsolidation after retrieval. Nature, 406(6797), 722e726. Quirk, G. J., & Mueller, D. (2008). Neural mechanisms of extinction learning and retrieval. Neuropsychopharmacology Reviews, 33(1), 56e72. Schiller, D., Kanen, J. W., LeDoux, J. E., Monfils, M., & Phelps, E. A. (2013). Extinction during reconsolidation of threat memory diminishes prefrontal cortex involvement. Proceedings of the National Academy of Sciences of the United States of America, 110(50), 20040e20045. Schiller, D., Monfils, M., Raio, C. M., Johnson, D. C., LeDoux, J. E., & Phelps, E. A. (2010). Preventing the return of fear in humans using reconsolidation update mechanisms. Nature, 463(7277), 49e53. Sharma, E., Thennarasu, K., & Janardhan Reddy, Y. C. (2014). Long-term outcome of obsessive-compulsive disorder in adults: a meta-analysis. Journal of Clinical Psychiatry, 75(9), 1019e1027. Shiban, Y., Bru¨tting, J., Pauli, P., & Mu¨hlberger, A. (2015). Fear reactivation prior to exposure therapy: does it facilitate the effects of VR exposure in a randomized clinical sample? Journal of Behavior Therapy and Experimental Psychiatry, 46, 133e140. Soeter, M., & Kindt, M. (2011). Disrupting reconsolidation: pharmacological and behavioral manipulations. Learning and Memory, 18(6), 357e366. Tabbert, K., Stark, R., Kirsch, P., & Vaitl, D. (2006). Dissociation of neural responses and skin conductance reactions during fear conditioning with and without awareness of stimulus contingencies. NeuroImage, 32(2), 761e770. Vervliet, B., Craske, M. G., & Hermans, D. (2013). Fear extinction and relapse: state of the art. Annual Review of Clinical Psychology, 9, 215e248. Walter, B., Blecker, C., Kirsch, P., Sammer, G., Schienle, A., Stark, R., et al. (2003). MARINA: an easy to use tool for the creation of masks for region of interest analyses. In 9th
122
c o r t e x 7 9 ( 2 0 1 6 ) 1 1 2 e1 2 2
International Conference on Functional Mapping of the Human Brain (Vol. 19, No. 2). Available on CD-Rom in NeuroImage. Warren, V. T., Anderson, K. M., Kwon, C., Bosshardt, L., Jovanovic, T., Bradley, B., et al. (2014). Human fear extinction and return of fear using reconsolidation update mechanisms: the contribution of on-line expectancy ratings. Neurobiology of Learning and Memory, 113, 165e173. Weike, A. I., Hamm, A. O., Schupp, H. T., Runge, U., Schroeder, H. W. S., & Kessler, C. (2005). Fear conditioning following unilateral temporal lobectomy: dissociation of
conditioned startle potentiation and autonomic learning. The Journal of Neuroscience, 25(48), 11117e11124. Weike, A. I., Schupp, H. T., & Hamm, A. O. (2007). Fear acquisition requires awareness in trace but not delay conditioning. Psychophysiology, 44(1), 170e180. Wendt, J., Neubert, J., Lindner, K., Ernst, F. D., Homuth, G., Weike, A. I., et al. (2014). Genetic influences on the acquisition and inhibition of fear. International Journal of Psychophysiology. http://dx.doi.org/10.1016/j.ijpsycho.2014.10.007.