نقش استروژن در خاطرات مزاحم

Abstract Intrusive memories are highly vivid, emotional and involuntary recollections which cause significant distress across psychological disorders including posttraumatic disorder (PTSD). Recent evidence has potentially extended our understanding of the development of intrusive memories by identifying biological factors which significantly impact on memories for emotionally arousing stimuli. This study investigated the role of stress on the development of intrusions for negative and neutral images, and indexed the potential contributions of sex (estrogen and progesterone) and stress (noradrenaline and cortisol) hormones. Whilst viewing the images, half the participants underwent a cold pressor stress (CPS) procedure to induce stress while the control participants immersed their hands in warm water. Saliva samples were collected to index estrogen, progesterone and noradrenergic and cortisol response. Participants (55 university students, 26 men, 29 women) viewed a series of negatively arousing and neutral images. Participants completed recall and intrusions measures 2 days later. Negative images resulted in greater recall and more intrusions than neutral images. In the cold water condition females recalled fewer neutral memories than males. Cortisol increase predicted decreased recall of negative memories in males, and estrogen predicted increased intrusions of negative images in women. These findings are consistent with evidence that circulating levels of ovarian hormones influence memory for emotionally arousing events, and provides the first evidence of the influence of sex hormones on intrusive memories. These results provide one possible explanation for the higher incidence of anxiety disorders in women.

1. Introduction Intrusive memories are highly vivid, emotional, and involuntary recollections of events that are typically negative or traumatic in nature. They are a common feature of various psychological disorders, including posttraumatic stress disorder (PTSD; (Bryant, O’Donnell, Creamer, McFarlane, & Silove, 2011)), depression (Patel et al., 2007), bipolar disorder (Gregory, Brewin, Mansell, & Donaldson, 2010), obsessive–compulsive disorder (Lipton, Brewin, Linke, & Halperin, 2010), and social phobia (Hackmann, Clark, & McManus, 2000). Across the disorders, intrusive memories share common characteristics in that they are often unwanted, interfere with ongoing cognitive activity, are experienced as nonvolitional, and are difficult to control (Clark & Rhyno, 2005). Prevailing theories of intrusions posit that (a) these memories are not adequately embedded within a person’s autobiographical memory base, increasing the likelihood of triggering unwanted intrusive thoughts (Conway & Pleydell-Pearce, 2000), (b) these memories were encoded in highly sensory-perceptual details which lack contextual grounding, and therefore intrude into consciousness (Brewin, Gregory, Lipton, & Burgess, 2010), (c) lack of conceptual processing results in traumatic information remaining prone to automatic, cued activation (Ehlers & Clark, 2000), or that (d) memories are activated by cues associated with the original traumatic event (Foa, Steketee, & Rothbaum, 1989). A common theme across these models is that the reason the initial memory is encoded in a manner that leads to subsequent intrusive recollections is that the original encoding occurs during stress. The role of stress has been repeatedly demonstrated in emotional memories, insofar as emotionally arousing events are better remembered than neutral events (Cahill, 2000). The fast-acting sympathetic adrenomedullary system is responsible for the production of adrenaline and noradrenaline leading to increased blood pressure, heart rate, and availability of glucose to the muscles and brain. The slower-acting HPA system involves the production of cortisol by the adrenal cortex, which increases blood sugar levels and metabolism (Tsigos & Chrousos, 2002). Consistent with the notion that arousal plays a role in emotional memories, noradrenergic activation during encoding results in stronger memories for emotional stimuli (McGaugh and Roozendaal, 2009 and van Stegeren, 2005). Further, glucocorticoid activation plays a role in emotional memory in that hydrocortisone administration leads to superior memory of emotional, but not neutral, information (Abercrombie, Speck, & Monticelli, 2006). It appears that noradrenergic and glucocorticoid systems interact to enhance emotional memories (McGaugh & Roozendaal, 2009). The evidence that this interaction between arousal systems enhances emotional memories has been explained in terms of glucocorticoids being able to pass the blood–brain barrier, thereby facilitating noradrenergic effects in the amygdala (de Quervain et al., 2009 and Roozendaal et al., 2002). There is also accumulating evidence demonstrating that this stress-facilitated emotional memory is influenced by gender and sex hormones (Andreano & Cahill, 2009). Women have better recall of emotional information than men (Bloise and Johnson, 2007 and Canli et al., 2002), and females are more likely to develop disorders related to emotional memory, such as posttraumatic stress disorder, than males (Olff et al., 2007). Sex-related lateralization of amygdala activity has been shown to differentially mediate enhanced emotional memory in men and women (Cahill, Uncapher, Kilpatrick, Alkire, & Turner, 2004). There is also evidence that sex hormones influence emotional memory (Cahill, 2006). In rodent studies, administration of an estrogen antagonist or ovariectomy prevents the impairing effects of pre-training stress on eyeblink conditioning (Wood & Shors, 1998) and reduces sex differences in a fear conditioning task (Gupta, Sen, Diepenhorst, Rudick, & Maren, 2001). These findings suggest that levels of circulating estrogen may mediate the effects of stress in aversive training tasks. In humans, the luteal phase is associated with higher levels of circulating estrogen and progesterone, which in turn may heighten glucocorticoid release due to progesterone binding to receptor sites (Koubovec, Ronacher, Stubsrud, Louw, & Hapgood, 2005). The increase of glucocorticoid release at the time of encoding an emotional experience may consolidate the memory (Pitman, Shalev, & Orr, 2000). In this context it is noteworthy that the mid-luteal phase is associated with stronger recall of emotional memories in both healthy controls (Andreano, Arjomandi, & Cahill, 2008) and trauma-exposed (Bryant, Felmingham, et al., 2011) women. There is also evidence that progesterone levels at the time of encoding predict subsequent emotional memory for negative images (Ertman et al., 2011 and Felmingham et al., 2012). Few experimental studies have examined the role of sex hormones in intrusive memories. There is evidence that women experience more intrusive recollections of emotional stimuli than men (Ferree & Cahill, 2009), and females in the luteal phase are more likely to experience intrusions of experimentally-generated negative stimuli (Ferree, Kamat, & Cahill, 2011) and flashback memories if they experience trauma during the mid-luteal phase (Bryant et al., 2011). Taken together, sex hormones appear to play an important role in memory for emotional material and this issue needs to be further investigated in the development of intrusive emotional memories. A major goal of this study was to extend previous research by investigating the role of sex hormones on intrusive emotional memories. A further aim was to enhance the ecological validity of the experimental procedure by administering the stress manipulation at the same time as the presentation of the encoded material. Previous studies have administered a cold pressor stress procedure (Andreano et al., 2008) or epinephrine (Cahill & Alkire, 2003) immediately after presentation of the information. To more closely simulate occurrence of encoding during traumatic events, we required participants to immerse their arm in cold water during the viewing of the material. Timing of the stressor in relation to the encoding of information has been shown to be influential in terms of subsequent memory insofar as cortisol increase at the time of retrieval impairs memory ( de Quervain et al., 1998 and Kuhlmann and Wolf, 2005), but enhances emotional memory at the time of encoding or consolidation ( Andreano and Cahill, 2006 and Kuhlmann and Wolf, 2006). The role of cortisol in encoding appears to be complex, however, in that cortisol increase at consolidation can impair memory ( Diamond, Campbell, Park, Halonen, & Zoladz, 2007), and the effect may be modulated by attentional factors at the time of encoding/consolidation ( Preuss, Schoofs, & Wolf, 2009). Accordingly, to simulate the co-existence of the stressor at the time of encoding we randomized males and females to either a cold pressor stress or control condition during presentation of neutral and aversive images. Salivary samples of estrogen, progesterone, noradrenaline and cortisol were collected during the experiment. Two days later participants returned and completed a surprise free recall test and a measure of intrusive memories. We hypothesized that there would be more memories recalled and more intrusions for (a) negative than for neutral images, (b) high than low stress, (c) women than men, and (d) levels of estrogen and progesterone will be positively associated with intrusions for negative but not neutral images.

3. Results 3.1. Participant characteristics All statistical analyses were performed in SPSSv19. Mean participant characteristics are presented in Table 1. Separate 2 (Group) × 2 (Sex) analyses of variance (ANOVA) of participants’ ages and Depression, Anxiety, Stress (DASS) scores indicated no significant main or interaction effects, suggesting no systematic differences between groups at baseline. Separate 2 (Group) × 2 (Sex) ANOVAs of progesterone and estrogen indicated significant main effects for Sex [progesterone: F(1,51) = 6.57, p = .01; estrogen: F(1,51) = 8.20, p = .006], indicating that females had higher progesterone and estrogen levels than males. There were no differences between Groups on either sAA or cortisol (see Table 1) Table 1. Participant characteristics, memory and intrusions of negative and neutral images (N = 55). Cold water condition Warm water condition Male Female Male Female (n = 12) (n = 13) (n = 14) (n = 16) Age 18.92 (1.44) 18.77 (1.42) 19.36 (2.65) 18.94 (2.26) DASS-depression 4.67 (5.61) 4.77 (3.11) 4.43 (3.86) 4.50 (5.59) DASS-anxiety 3.17 (2.48) 3.85 (2.08) 4.00 (2.83) 5.63 (5.23) Estrogen 1.67 (0.82) 2.15 (0.96) 1.46 (0.47) 2.25 (0.89) Progesterone 26.71 (19.65) 51.52 (30.95) 32.49 (23.76) 43 79 (21.85) sAA baseline 139.14 (58.21) 107.77 (71.36) 104.53 (64.35) 154.76 (76.65) sAA post CPT 147.68 (80.44) 116.15 (74.42) 104.35 (63.92) 140.92 (66.11) Cortisol baseline .17 (.10) .28 (.21) .20 (.11) .14 (.07) Cortisol post CPT .26 (.21) .33 (.40) .14 (.09) .12 (.05) Negative recall 3.42 (1.62) 4.46 (1.66) 4.79 (1.37) 4.13 (1.78) Neutral recall 2.08 (1.08) .31 (.63) 1.14 (1.03) 1.00 (.82) Negative intrusion 1.75 (1.86) 3.85 (3.02) 3.00 (3.33) 2.63 (2.75) Neutral intrusions .50 (.80) 1.08 (1.26) .86 (1.96) 1.19 (1.68) Note: Standard deviations appear in parentheses. sAA measured in μ/ml. Cortisol measured in μg/dL. Estrogen measured in pmol/L. Progesterone measured in nmol/L. CPT = cold pressor task. Table options 3.2. Manipulation check sAA and cortisol levels prior to and following the CPS manipulation are presented in Fig. 1. To index the effect of the CPS on participants’ sAA and cortisol levels, a 2 (Group) × 2 (Gender) × 2 (Time) analysis of variance (ANOVA) of sAA indicated a significant (Group) × 2 (Gender) interaction effect, F(1,51) = 5.62, p = .02, η = .10. Whereas males in the Warm Water Group had lower sAA levels than females, males had higher sAA levels than females in the Cold Water Group (p < .05). A 2 (Group) × 2 (Gender) × 2 (Time) ANOVA of cortisol increase indicated a significant main effect for Group, F(1,51) = 6.30, p = .02, η = .11, and a Group × Time interaction, F(1,51) = 5.62, p = .02, η = .10. Participants in the Cold Water Group experienced a greater increase in cortisol than those in the Warm Water Group. Whereas groups did not differ in their cortisol levels at baseline, those in the Cold Water Group had higher cortisol levels than those in the Warm Water Group following the stress manipulation (p = .005). This pattern suggests that the stressor manipulation did increase the cortisol levels for participants in the Cold Water Group. Salivary sAA and cortisol levels prior to and following cold pressor task ... Fig. 1. Salivary sAA and cortisol levels prior to and following cold pressor task (N = 55). Figure options 3.3. Free recall The means of participants’ recall scores are presented in Table 1. A 2 (Group) × 2 (Gender) × 2 (Stimulus Type) ANOVA of recalled memories indicated a significant main effect for Stimulus Type [F(1,51) = 167.63, p = .001, η = .77] and a significant three-way Group × Gender × Stimulus Type interaction effect [F(1,51) = 12.44, p = 0.001, η = .20]. Participants recalled more negative than neutral memories. To clarify the three-way interaction, separate 2 (Group) × 2 (Gender) ANOVAs were conducted for negative and neutral stimuli, respectively. The 2 (Group) × 2 (Gender) ANOVA of negative memories indicated no significant effects. The 2 (Group) × 2 (Gender) ANOVA of neutral memories indicated a significant main effect for Gender [F(1,51) = 15.41, p < 0.01, η = .23] and a significant Group × Gender interaction effect [F(1,51) = 11.16, p < 0.002, η = .18]. Overall, fewer neutral pictures were recalled in the Cold Water than Warm Water Group. Females in the Cold Water Group reported fewer neutral memories than males [t(23) = 5.06, p < .001). 3.4. Intrusions The means of participants’ total scores on the intrusion questionnaire are presented in Table 1. A 2 (Group) × 2 (Gender) × 2 (Stimulus Type) × 2 (Questionnaire Order) ANOVA of intrusions indicated a significant main effect of Stimulus Type [F(1,51) = 35.73, p < 0.001, η = .41]. Participants reported more intrusions of the negative compared to the neutral pictures viewed. 3.5. Role of stress and sex hormones In terms of emotional memories for males, cortisol increase predicted decreased recall (accounting for 18% of the variance), and there were no significant predictors in females. In terms of neutral memories for males, cortisol increase predicted decreased recall (accounting for 17%1 of the variance), and there were no significant predictors in females(see Table 2). There were no predictors in males or females for recall of intrusions of neutral images. In terms of negative intrusions, there were no significant predictors for males, but estrogen predicted increased intrusions for females (accounting for 16% of variance) (see Table 3). Table 2. Summary of hierarchical regression models for recall of negative and neutral images (N = 55). B SEB β p Negative images Males Step 1: Arousal condition −1.39 .67 −.44 .05 Step 2: sAA increase .22 .87 .05 .80 Step 3: Cortisol increase 1.26 1.17 .21 .29 Step 4: Estrogen .10 .59 .05 .87 Step 5: Progesterone −.91 1.22 −.21 .47 Females Step 1: Arousal condition −.28 .76 −.08 .72 Step 2: sAA increase 1.04 .67 .40 .13 Step 3: Cortisol increase .86 .73 .28 .25 Step 4: Estrogen .16 .34 .10 .65 Step 5: Progesterone −.27 .37 −.72 .48 Note: Males: Step 1 R2 = 18, Δ R2 = .18. Step 2 R2 = .18, Δ R2 = .00. Step 3 R2 = .23, Δ R2 = .04, Step 4 R2 = .24, Δ R2 = .00. Step 5 R2 = .26, Δ R2 = .02. Females: Step 1 R2 = .01, Δ R2 = .01. Step 2 R2 = .04, Δ R2 = .03. Step 3 R2 = .10, Δ R2 = .07. Step 4 R2 = .10, Δ R2 = .00. Step 5 R2 = .12, Δ R2 = .02 B SEB β p Neutral images Males Step 1: Arousal condition .67 .47 .30 .17 Step 2: sAA increase −.22 .60 −.07 .72 Step 3: Cortisol increase −.07 .81 −.02 .93 Step 4: Estrogen .60 .42 .41 .16 Step 5: Progesterone −1.14 .85 −.38 .19 Females Step 1: Arousal condition −.44 .32 −.27 .19 Step 2: sAA increase −.50 .28 −.41 .09 Step 3: Cortisol increase −.26 .31 −.18 .40 Step 4: Estrogen −.07 .14 −.09 .66 Step 5: Progesterone .11 .16 .71 .48 Note: Males: Step 1 R2 = .19, Δ R2 = .19. Step 2 R2 = .24, Δ R2 = .05. Step 3 R2 = .27, Δ R2 = .03, Step 4 R2 = .27, Δ R2 = .00. Step 5 R2 = .29, Δ R2 = .02. Females: Step 1 R2 = .19, Δ R2 = .19. Step 2 R2 = .24, Δ R2 = .05. Step 3 R2 = .27, Δ R2 = .03. Step 4 R2 = .27, Δ R2 = .00. Step 5 R2 = .29, Δ R2 = .02 Table options Table 3. Summary of hierarchical regression models for intrusive memories of negative and neutral images (N = 55). B SEB β p Negative images Males Step 1: Arousal condition −.25 1.14 −.05 .83 Step 2: sAA increase 1.04 1.46 .14 .49 Step 3: Cortisol increase 3.70 1.97 .37 .08 Step 4: Estrogen −1.57 1.00 −.44 .14 Step 5: Progesterone 3.32 2.07 .45 .12 Females Step 1: Arousal condition 1.37 1.18 .24 .26 Step 2: sAA increase −.48 1.03 −.11 .65 Step 3: Cortisol increase −.19 1.13 −.04 .87 Step 4: Estrogen 1.23 .53 .48 .03 Step 5: Progesterone −.44 .59 −.18 .47 Note: Males: Step 1 R2 = .05, Δ R2 = .05. Step 2 R2 = .07, Δ R2 = .02. Step 3 R2 = .17, Δ R2 = .10, Step 4 R2 = .18, Δ R2 = .01. Step 5 R2 = .28, Δ R2 = .09. Females: Step 1 R2 = .05, Δ R2 = .05. Step 2 R2 = .08, Δ R2 = .00. Step 3 R2 = .08, Δ R2 = .00. Step 4 R2 = .25, Δ R2 = .16. Step 5 R2 = .26, Δ R2 = .02 B SEB β p Neutral images Males Step 1: Arousal condition −.04 .64 −.01 .95 Step 2: sAA increase −.08 .82 −.02 .92 Step 3: Cortisol increase 2.67 1.10 .49 .02 Step 4: Estrogen −.37 .56 −.19 .52 Step 5: Progesterone .29 1.15 .07 .81 Females Step 1: Arousal condition −.14 .62 −.05 .82 Step 2: sAA increase −.26 .54 −.12 .64 Step 3: Cortisol increase .24 .59 .09 .69 Step 4: Estrogen .54 .28 .42 .06 Step 5: Progesterone −.26 .31 −.21 .40 Note: Males: Step 1 R2 = .01, Δ R2 = .01. Step 2 R2 = .02, Δ R2 = .01. Step 3 R2 = .22, Δ R2 = .20, Step 4 R2 = .24, Δ R2 = .02. Step 5 R2 = .24, Δ R2 = .00. Females: Step 1 R2 = .00, Δ R2 = .00. Step 2 R2 = .07, Δ R2 = .07. Step 3 R2 = .09, Δ R2 = .02. Step 4 R2 = .20, Δ R2 = .10. Step 5 R2 = .22, Δ R2 = .02