The Effect of Glucose on Hippocampal-Dependent Contextual Fear Conditioning

ARCHIVAL REPORT

The Effect of Glucose on Hippocampal-Dependent Contextual Fear Conditioning Daniel E. Glenn, Thomas R. Minor, Bram Vervliet, and Michelle G. Craske Background: The metabolic challenge of trauma disrupts hippocampal functioning, which is necessary for processing the complex cooccurring elements comprising the traumatic context. Poor contextual memory of trauma may subsequently contribute to intrusive memories and overgeneralization of fear. Glucose consumption following trauma may be a means to protect hippocampal functioning and contextual fear learning. This study experimentally examined the effect of glucose on hippocampal-dependent contextual learning versus cued fear learning in humans. Methods: Forty-two male participants underwent cued conditioning with an unconditional stimulus (US) (shock) paired with a discrete conditional stimulus (geometric shape) and context conditioning (requiring hippocampal processing) with a US unpredictably paired with a background context (picture of room). Participants were then blindly randomized to consume either a 25 g glucose or sweettasting placebo drink and returned for a test phase 24 hours later. Measures included acoustic startle response, US expectancy, blood glucose levels, and arousal ratings. Results: The glucose group showed superior retention of hippocampal-dependent contextual learning at test relative to the placebo group, as demonstrated by acoustic startle response and US expectancy ratings. Glucose and placebo groups did not differ on any measure of cued fear learning at test. Conclusions: This study provides experimental evidence that in mildly stressed humans postconditioning glucose consumption improves retention of hippocampal-dependent contextual learning but not cued learning. Ultimately, glucose consumption following trauma may be a means of improving learning about the traumatic context, thereby preventing subsequent development of symptoms of posttraumatic stress.

Key Words: Conditioning, context, cue, fear, glucose, hippocampus

T

raumatic experiences are common and can have severe and long-lasting biological and psychological consequences. Lifetime prevalence rates are estimated at over 50% for traumatic experiences and at roughly 6% to 7% for developing posttraumatic stress disorder (PTSD) (1). Posttraumatic stress disorder prevalence rates are elevated in populations with higher rates of trauma exposure (e.g., 12% to 18% in deployed military personnel) (2). Fear conditioning is assumed to play a central role in trauma exposure [e.g., (3)], as is extinction in exposure treatment for trauma memories. Recent research has targeted pharmacologic augmentation of extinction by d-cycloserine (4,5) and blockage of reconsolidation of trauma memories by propranolol (6,7). Thus far, however, little attention has been given to alteration of the unique biological mechanisms underlying the initial formation of trauma memories. Poor contextual control of the fear response appears to be a central component of PTSD. The spatial-temporal context for memories provides the when and where details that normally constrain a memory so that its recall is context appropriate (8). A dual representation model of PTSD proposes that poor contextual trauma memory results in subsequently deficient top-down contextual control over memory recall in response to internal

From the Department of Psychology (DEG, TRM, MGC), University of California, Los Angeles, California; and Department of Psychology (BV), Katholieke Universiteit Leuven, Leuven, Belgium. Address correspondence to Daniel E. Glenn, M.A., University of California, Los Angeles, Department of Psychology, 1285 Franz Hall, Box 951563, Los Angeles, CA 90095-1563; E-mail: [email protected]. Received Nov 28, 2012; revised Sep 17, 2013; accepted Sep 20, 2013.

0006-3223/$36.00 http://dx.doi.org/10.1016/j.biopsych.2013.09.022

and external trauma cues (9,10). Failure to appropriately contextualize trauma cues leads such stimuli to be experienced as indications of immediate threat (i.e., flashbacks) rather than as reminders of a past event. Posttraumatic stress disorder symptoms of forgetting specific traumatic details (e.g., dissociative amnesia), intrusive memories, and flashbacks triggered by vague traumatic reminders may all be due to lack of situationally specific memory. The hippocampus plays a central role in processing and consolidating the spatial-temporal context for memories (11– 13), and poor contextual memory of trauma may result from impaired hippocampal functioning. Reduced hippocampal volume has been observed in individuals with PTSD in the majority of studies (14–16), and impairment of contextual memory due to hippocampal dysfunction is believed to be central to the etiology of PTSD (17). One implication of deficient hippocampal functioning and contextual memory is impaired contextual fear conditioning. The definition of context in conditioning research has been broadly defined, generally based on at least one of the following qualities: 1) being an unpredictable predictor of the unconditional stimulus (US); 2) having longer duration than a typical conditioned stimulus (CS); and 3) being comprised of complex, multimodal features. Contexts have been operationalized in numerous distinctly different ways [i.e. (18–23)], and it has been argued that contextual learning may occur to physical environmental stimuli, internal emotional states, internal drug-induced states, and training trial number, as well as the passage of time (24). Learning about specific contextual details and complex stimuli may occur through conjunctive representation (memory that binds together numerous co-occurring elements of a context) or through elemental contextual representation (contextual features learned individually but not bound together). Conjunctive learning supporting context conditioning is associated with the

BIOL PSYCHIATRY 2014;75:847–854 Published by Elsevier Inc on behalf of Society of Biological Psychiatry

D.E. Glenn et al.

848 BIOL PSYCHIATRY 2014;75:847–854 hippocampus, while elemental learning supporting context conditioning is primarily associated with the amygdala [i.e., (25)]. Convergent results from distinctly different research paradigms [contextual fear conditioning: i.e., (25,26); appetitive discrimination learning (27)] suggest that conjunctive-hippocampal and elemental-amygdala learning processes compete over context learning and that under normal circumstances conjunctive/ hippocampal learning dominates. Contextual fear conditioning is impeded by hippocampal impairment (28,29) and has been demonstrated to require hippocampal function in humans. In a functional magnetic resonance imaging study, participants showed uniquely altered hippocampal activity, but not amygdala activity, during fear conditioning to similar but distinct background pictures of rooms (19). It remains debatable whether this methodology actually assessed context conditioning (i.e., does a photograph of a room constitute a context?) and whether the learning required conjunctive or elemental processing, but it demonstrated unique hippocampal-dependent processes in differentiating between complex pictures of rooms. Learning about contextual details during and following trauma may become disrupted due to hippocampal impairment resulting from substantial trauma-induced alterations in brain physiology and function that mediate changes in information processing and memory consolidation (11,12,30–35). Trauma strongly increases neural metabolic demand, but the brain does not store sufficient long-term supplies of blood-glucose or oxygen to maintain normal functioning during such circumstances (13,36,37). Simultaneously, the uptake transport of glucose at the blood-brain barrier is slowed by the release of cortisol, thereby impairing the anaerobic phase of respiration (13,36,37). This type of metabolic challenge seriously compromises normal cell function and related psychological processes. The hippocampus is particularly susceptible to impairment from glucose deprivation (13,34,36,37). Consequently, the metabolic challenge that occurs during traumatic experience functionally may take the hippocampus “offline,” thereby impeding formation and consolidation of representations of the traumatic context. Indeed, infusing glucocorticoids into the hippocampus of mice following fear conditioning produces PTSD-like impairments in their ability to learn about contextual prediction of threat (38). Sparing hippocampal function following trauma may mitigate impaired learning about the traumatic context, thereby improving the situational specificity of traumatic memories and potentially mitigating sequelae that lead to long-term emotional problems. One means to enhance hippocampal functioning and improve memory is through the ingestion of glucose (13). Elevated blood glucose, facilitated by the release of epinephrine, improves memory performance (39). Exogenous glucose administration before or after learning enhances memory performance in humans (40,41) and in rats (42–44). The decrease in hippocampal extracellular glucose typically observed in rats following a challenging maze task is reversed by administering a glucose injection, which additionally enhances memory performance (42). Furthermore, glucose ingestion immediately following uncontrollable traumatic stress in rats prevents development of the behavioral symptoms of anxiety, acute stress, and depression that are typically observed 24 hours later (30). One study examined the effect of carbohydrate consumption on recovery from stress-induced deficits in cognitive functioning (45), but no research in humans, thus far, has examined the effect of glucose on the hippocampus and contextual fear conditioning. The goal of the present study was to investigate www.sobp.org/journal

whether glucose administration uniquely enhances human hippocampal-dependent fear learning about complex contextual details. This research utilized fear conditioning methodology previously demonstrated to distinguish between hippocampaldependent learning (and possibly context conditioning) versus nonhippocampal-dependent cued conditioning (19). The results of this research will inform basic learning processes and may ultimately have implications for ameliorating the negative sequelae of trauma exposure.

Methods and Materials Participants Forty-two male undergraduate students (mean age ¼ 20.9 years, SD ¼ 5.8 years) served as study participants. The sample was restricted to male subjects to reduce variability in glucose metabolism (46,47). The ethnic distribution of the sample was 40.5% Caucasian, 35.6% Asian, 9.5% Hispanic, 4.8% African American, and 9.6% who classified themselves as multiethnic or other. Exclusion criteria and study recruitment details are included in Supplement 1. Overall Study Procedure The experiment consisted of 2 days. Day 1 included a stimulus selection phase, habituation phase, and acquisition phase, during which participants underwent both cued fear conditioning and context fear conditioning. [Throughout Methods and Materials and Results, the term context conditioning is used to refer to associative learning engendered by pairing a muscle stimulation with a background picture of a room. It should be acknowledged that it remains debatable whether associative learning about similar but distinct photographs truly represents context conditioning, but the term is used for the sake of clarity and for consistency with the wording originally used by Marscher et al. (16) to describe the procedures]. Participants were randomized to consume either a glucose drink or sweet-tasting placebo drink immediately following acquisition on day 1. On day 2 (24 hours later), participants returned to the laboratory and completed a test phase. Experimental Stimuli Two geometric shapes (purple pentagon, orange parallelogram) served as the discrete conditional stimuli (CS⫹, CS), and two background pictures of similar but distinguishable rooms served as the contexts (context⫹, context). During the habituation phase and extinction test phase, participants were also presented with a neutral CS (CSN) (long black rectangle) and neutral context (contextN). The US was an electrical shock to the bicep muscle delivered by a constant voltage stimulator (model STMISOL; BIOPAC Systems, Inc., Goleta, California). Details of the US parameters and workup procedure are included in Supplement 1. See Figure 1 for diagram of the presentation of conditioning stimuli across study phases. Physiological Measures Acoustic Startle Response. Surface electromyogram data were acquired with an analog/digital recording system (model MP150; BIOPAC Systems, Inc.). A total of four auditory startle probes were presented during each study trial: two probes occurred during context-alone periods, and two probes occurred while both context and CS were present. Probes were presented 6 to 9 seconds after onset of the context-alone and 3.5 to 6.5 seconds after CS onset. During the acquisition phase, probes were

D.E. Glenn et al.

BIOL PSYCHIATRY 2014;75:847–854 849 Figure 1. Conditioning stimulus presentation. Context, negative context; Context⫹, positive context; CS, conditioned stimulus; CS, negative conditioned stimulus; CS⫹, positive conditioned stimulus; US, unconditioned stimulus.

never presented within 3 seconds before or after the US. Acoustic startle reflexes (ASR) during CS and context-alone periods were calculated as the difference between mean electromyogram amplitude during the 20 milliseconds before onset and the maximum amplitude during the 52 to 65 milliseconds following probe onset. A full description of the parameters used in collecting and analyzing ASR are included in Supplement 1. Blood Glucose. Blood glucose level was measured using the AccuChek Compact Plus Blood Glucose Meter (Roche Diagnostics, Indianapolis, Indiana) (48) at three different points on day 1: immediately after informed consent, immediately following habituation, and immediately following acquisition (before drink consumption). Self-Report and Subjective Measures Questionnaires. Participants completed the following selfreport questionnaires on day 1 before beginning the experimental procedures: 1) the Beck Depression Inventory (49), a 21item questionnaire to assess symptoms of depressed mood; and 2) the Behavioral Inhibition Scale (50), a 16-item measure of anxiety proneness (51). These measures were collected to control for potential group differences at baseline in mood and anxiety proneness that might influence associative learning processes. US Expectancy. To assess explicit learning, participants continuously rated expectancy of the US on a sliding dial (BIOPAC model TSD115, AcqKnowledge version 3.7.3; BIOPAC Systems, Inc.). The US expectancy scale was labeled “expectancy of shock in the next few moments” and ranged from 0 (certain no shock) to 10 (certain shock) with uncertain as the midpoint. Unconditional stimulus expectancy was calculated as the mean rating during the .5 second before startle probe onset. Arousal. Conditional stimuli and contexts were rated on a 0 to 10 Likert scale (0 ¼ not at all, 5 ¼ moderately, 10 ¼ very) for arousal (“How fearful does this make you?”). Participants completed these ratings before acquisition, after acquisition, before test, and after test. Procedure Day 1. Participants were asked to refrain from eating or drinking anything other than water for 2 hours before both study sessions. After providing informed consent and completing

questionnaires, participants were seated in front of a computer monitor and fitted with electrodes for delivering the US and for recording physiological responses. Participants also put on headphones, which delivered the startle probe. The experimental procedures used for context and cued conditioning were similar to those used by Marschner et al. (19). The experimental procedures are described briefly below, with a full description included in Supplement 1. Following a 10-minute adaptation period, participants completed a habituation phase consisting of six 60-second trials, during each of which the CS⫹, CS, or CSN was presented twice (for 8 sec) during two time windows, superimposed over a background context (picture of similar but distinct rooms) present throughout the trial. In all study phases, there was a 30-second intertrial interval showing only a black screen. Participants then completed the acquisition phase, comprised of six 60-second trials alternating between cued conditioning and context conditioning. In the three cued conditioning trials, there were two 8-second CS⫹ paired with the US (co-terminating together), all superimposed over context. In the three context conditioning trials, the CS and US were each presented twice, with the US presented at random times before and after but never paired with the CS, all superimposed over context. Next, participants were randomized to consume either a glucose drink (25 g of glucose powder mixed with 300 mL of water) or placebo drink (5 tablets of saccharin mixed with 300 mL of water). A review of research using glucose to enhance learning in humans found the optimal amount of glucose to range from 25 g to 50 g (52). Day 2. One day later, participants returned to the laboratory completed a test phase consisting of six 60-second trials identical to those presented during the day 1 habituation phase.

Results Baseline Measures Independent samples t tests revealed no significant differences between the placebo and glucose groups on the Beck Depression Inventory, t40 ¼ .99, p ¼ .38, or Behavioral Inhibition Scale, t40 ¼ .51, p ¼ .96. Groups also did not differ in age, t40 ¼ .39, p ¼ .70, or ethnicity, χ25, n ¼ 42 ¼ 5.58, p ¼ .35. www.sobp.org/journal

D.E. Glenn et al.

850 BIOL PSYCHIATRY 2014;75:847–854 Table 1. Cued and Context Conditioning: US Expectancy (Mean, SD) Glucose CS⫹ Acquisition Acquisition Acquisition Acquisition Acquisition Acquisition Test 1 Test 2 Test 3

Acquisition Acquisition Acquisition Acquisition Acquisition Acquisition Test 1 Test 2 Test 3

1 2 3 4 5 6

1 2 3 4 5 6

5.65 6.59 6.45 6.15 6.67 5.74 4.29 4.11 3.49

(1.96) (1.98) (2.22) (2.75) (2.50) (3.14) (3.19) (3.18) (3.66)

Placebo CS

CS

(2.39) (2.51) (2.07) (2.34) (2.67) (2.99) (3.24) (3.11) (3.22)

3.46 (2.76) 5.08 (2.79) 5.09 (3.25) 5.66 (3.58) 5.5 (3.64) 4.97 (3.77) 3.84 (3.72) 3.40 (3.40) 2.78 (3.67)

4.05 4.45 3.19 3.04 2.28 2.86 2.89 2.22 2.72

Context⫹

Context

Context⫹

Context

5.47 5.99 5.69 6.05 5.57 5.47 3.60 3.00 2.02

5.05 6.06 5.64 6.19 5.76 5.93 2.05 2.50 2.58

3.70 4.95 4.30 5.34 5.13 4.54 3.66 4.55 3.45

2.90 4.61 3.56 2.90 1.87 2.42 3.13 3.43 2.18

(2.54) (2.29) (2.39) (2.71) (3.11) (2.86) (3.21) (3.47) (3.21)

4.83 5.08 5.52 6.27 6.25 5.19 4.09 3.75 3.26

CS⫹

(2.42) (2.16) (2.42) (2.49) (3.17) (3.12) (2.96) (3.36) (2.85)

(2.66) (3.41) (3.12) (3.67) (3.62) (3.89) (3.41) (2.75) (2.78)

(3.08) (2.84) (2.89) (2.98) (3.05) (2.96) (2.85) (2.60) (3.32)

(2.47) (2.91) (3.33) (3.35) (2.59) (2.95) (2.79) (2.91) (2.99)

Context, negative context; Context⫹, positive context; CS, negative conditioned stimulus; CS⫹, positive conditioned stimulus; US, unconditioned stimulus.

Blood Glucose Blood glucose level did not significantly correlate with any outcome measures and thus was not included as a covariate in additional analyses. CSN and ContextN Initially, analysis of variance (ANOVA) analyses were conducted with three types of CS (CS⫹, CS, CSN) and context (context⫹, context, contextN). All ANOVAs, which included three levels of CS and context, found no significant interactions or main effects of reinforcement. Thus, the remaining analyses conducted herein excluded the neutral stimuli (CSN and contextN) and included only stimuli that were included in the acquisition phase (CS⫹, CS, context⫹, and context). Acquisition: Cue Conditioning Mixed-design 2 (reinforcement; CS⫹, CS)  2 (group; glucose, placebo)  6 (trial; acquisition trials 1–6) ANOVAs were conducted for US expectancy and ASR. US Expectancy. There were no significant two- or three-way interactions, but there was a significant main effect of reinforcement, F1,37 ¼ 18.54, p ⬍ .000, ƞ2 ¼ .079, indicating higher US expectancy to the CS⫹ than the CS (Table 1). ASR. There were no significant two- or three-way interactions, but there was a significant main effect of reinforcement, F1,37 ¼ 4.43, p ¼ .042, ƞ2 ¼ .017), indicating higher startle reflex to the CS⫹ than the CS (Table 2). Arousal Ratings. A 2 (reinforcement; CS⫹, CS )  2 (group; glucose, placebo)  2 (trial; preacquisition, postacquisition) mixed-design ANOVA was conducted for arousal ratings. There were significant main effects of reinforcement (F1,40 ¼ 10.68, p ¼ .002, ƞ2 ¼ .114) and time (F1,40 ¼ 4.96, p ¼ .032, ƞ2 ¼ .024), indicating significantly higher arousal to the CS⫹ than CS and higher arousal to both CSs after acquisition (Table 3). www.sobp.org/journal

Acquisition: Context Conditioning Mixed-design 2 (reinforcement; context⫹, context)  2 (group; glucose, placebo)  6 (trial; acquisition trials 1–6) ANOVAs were conducted for US expectancy and ASR. US Expectancy. There was a significant group  reinforcement interaction (F1,37 ¼ 6.2, p ¼ .017, ƞ2 ¼ .036). Tests of simple main effects indicated significantly higher expectancy to context in the glucose group than the placebo group (F1,37 ¼ 14.37, p ¼ .001) but no group differences in expectancy to context⫹. ASR. There were no significant two- or three-way interactions, but there was a significant main effect of reinforcement (F1,37 ¼ 5.73, p ¼ .022, ƞ2 ¼ .007), indicating significantly stronger startle reflex to context⫹ than context. Arousal Ratings. A 2 (reinforcement; context⫹, context)  2 (group; glucose, placebo)  2 (trial; preacquisition, postacquisition) mixed-design ANOVA was conducted for arousal ratings. There were significant main effects of reinforcement (F1,40 ¼ 29.77, p ¼ ⬍.001, ƞ2 ¼ .063) and time (F1,40 ¼ 16.61, p ⬍ .001, ƞ2 ¼ .043), indicating significantly higher arousal to the context⫹ than context and higher arousal following acquisition. Test Phase: Cued Conditioning Mixed-design 2 (reinforcement; CS⫹, CS)  2 (group; glucose, placebo)  4 (trial; end acquisition trial, test trials 1–3) ANOVAs were conducted for US expectancy and ASR. US Expectancy. There were no significant two- or three-way interactions, but there were significant main effects of reinforcement (F1,37 ¼ 4.42, p ¼ .05, ƞ2 ¼ .02) and trial (F3,37 ¼ 5.14, p ¼ .002, ƞ2 ¼ .068), indicating significantly higher expectancy to CS⫹ than CS and significantly lower expectancy in later trials. ASR. There were no significant or main effects. Arousal Ratings. A 2 (reinforcement; CS⫹, CS)  2 (group; glucose, placebo)  2 (time; pretest, posttest) mixed-design

Table 2. Cued and Context Conditioning: Acoustic Startle Response (Mean, SD) Glucose CS⫹ Acquisition Acquisition Acquisition Acquisition Acquisition Acquisition Test 1 Test 2 Test 3

Acquisition Acquisition Acquisition Acquisition Acquisition Acquisition Test 1 Test 2 Test 3

1 2 3 4 5 6

1 2 3 4 5 6

.81 .66 .71 .59 .63 .17 .27 .32 .30

(1.00) (.72) (.74) (.85) (.71) (.74) (1.00) (1.07) (.99)

Placebo CS

.79 .45 .34 .44 .44 .37 .18 .26 .22

(.92) (.58) (.56) (.65) (.76) (.51) (1.22) (1.00) (1.11)

CS⫹ .72 .57 .70 .56 .49 .28 .24 .36 .22

(1.01) (.74) (.68) (.77) (.66) (.58) (.85) (.86) (.90)

CS .81 .39 .38 .10 .11 .26 .01 .18 .29

(.87) (.59) (.78) (.51) (.68) (.64) (.78) (1.00) (.58)

Context⫹

Context

Context⫹

Context

.75 .64 .19 .61 .59 .84 .20 .55 .57

.68 .80 .74 .60 .70 .85 .35 .28 .38

.67 .64 .44 .57 .70 .68 .04 .14 .43

.72 .67 .71 .82 .71 .64 .29 .25 .09

(.32) (.48) (1.51) (.40) (.60) (.57) (.79) (.86) (1.03)

(.41) (.30) (.40) (.60) (.37) (.40) (.64) (.46) (.40)

(.34) (.43) (1.29) (.53) (.40) (.40) (1.17) (.84) (.84)

(.41) (.39) (.40) (.42) (.35) (.54) (.30) (1.42) (1.33)

Context, negative context; Context⫹, positive context; CS, negative conditioned stimulus; CS⫹, positive conditioned stimulus.

D.E. Glenn et al.

BIOL PSYCHIATRY 2014;75:847–854 851

Table 3. Arousal Ratings Before and After Acquisition and Test (Mean, SD) Arousal Glucose CS⫹ Preacquisition Postacquisition Pretest Posttest

Preacquisition Postacquisition Pretest Posttest

2.41 3.73 2.68 2.73

Placebo CS

(2.91) (3.15) (2.90) (2.83)

2.23 3.09 2.68 2.18

CS⫹

(2.78) (2.94) (2.77) (2.48)

2.65 4.60 2.55 2.25

(2.60) (3.21) (2.73) (2.51)

CS 2.75 3.20 1.95 2.00

(2.77) (3.14) (2.63) (2.49)

Context⫹

Context

Context⫹

Context

2.05 3.68 2.36 2.23

1.77 3.36 2.59 2.41

2.15 (2.48) 4.6 (3.03) 2.85 (2.64) 2.25 (2.55)

1.45 3.10 2.40 2.20

(2.32) (3.12) (2.61) (2.87)

(2.64) (3.14) (3.02) (2.92)

(1.61) (3.07) (2.37) (2.71)

Context, negative context; Context⫹, positive context; CS, negative conditioned stimulus; CS⫹, positive conditioned stimulus.

ANOVA was conducted for arousal ratings. There were no significant two- or three-way interactions, but there was a significant main effect of time (F1,40 ¼ 4.12, p ¼ .049, ƞ2 ¼ .025), indicating a decrease in arousal from pretest to posttest.

Mean US Expectancy

Test Phase: Context Conditioning Mixed-design 2 (reinforcement; context⫹, context)  2 (group; glucose, placebo)  4 (trial; end acquisition trial, test trials 1–3) ANOVAs were conducted for US expectancy and ASR. US Expectancy. There was a significant three-way reinforcement  group  trial interaction (F3,37 ¼ 3.71, p ¼ .014, ƞ2 ¼ .032) but no significant two-way interactions. Tests of simple twoway interactions were not significant. Tests of simple main effects indicated significantly higher expectancy to the context⫹ than to context at the end of acquisition in the placebo group (F1,37 ¼ 4.58, p ¼ .039). In the glucose group, there was significantly higher expectancy to context⫹ than to context at test trial 2 (F1,37 ¼ 4.73, p ¼ .036) and test trial 3 (F1,37 ¼ 4.75, p ¼ .036) (Figure 2). Additional analyses indicated that the magnitude of context conditioning during acquisition, as indicated by US expectancy, did not moderate retention of context conditioning as indicated by US expectancy in the test phase (see Supplement 1 online material). ASR. There was a significant three-way reinforcement  group  trial interaction (F3,37 ¼ 3.73, p ¼ .013, ƞ2 ¼ .033) and a significant two-way reinforcement  group interaction (F3,37 ¼ 16.17, p ⬍ .001,

ƞ2 ¼ .054). Test of simple two-way interactions were not significant. Tests of simple main effects indicated significantly stronger startle reflex to context⫹ than to context at test trial 1 (F1,37 ¼ 5.08, p ¼ .030), trial 2 (F1,37 ¼ 10.07, p ¼ .003), and trial 3 (F1,37 ¼ 10.08, p ¼ .003) in the glucose group. Simple main effects were not significant in the placebo group (Fs ⬍ 2.5) (Figure 3). Arousal Ratings. A 2 (reinforcement; context⫹, context)  2 (group; glucose, placebo)  2 (time; pretest, posttest) mixeddesign ANOVA was conducted. There were no significant interactions or main effects.

Discussion This study utilized fear-conditioning procedures to investigate whether glucose consumption enhances human hippocampaldependent fear-conditioning processes to pictures of rooms (or contexts) relative to nonhippocampal-dependent cued fear conditioning. The results largely support the study hypotheses, with glucose consumption enhancing the retrieval of hippocampaldependent fear conditioning about complex contextual details and having little effect on recall of cued conditioning. As hypothesized, after undergoing procedures previously demonstrated to differentially alter hippocampus functioning via contextual fear conditioning, participants who consumed glucose showed superior specificity in the retention of contextual fear, but not cued, fear 24 hours after acquisition relative to individuals who consumed a placebo. This effect was observed in ratings of the expectancy of the aversive stimulus, an index of associative learning, as well as in the auditory startle reflex, an index of defensive responding. These findings indicate that glucose selectively enhances hippocampal-dependent contextual fear conditioning processes. Unexpectedly, participants who consumed glucose showed inferior initial acquisition of contextual fear, as demonstrated by US expectancy ratings, relative to those randomized to the placebo. This group difference during acquisition of fear reflects random error, since all participants completed identical procedures until randomization to the glucose or placebo drink that occurred following acquisition. Though the two groups did not differ on any baseline measures or demographics, they may have differed on an unmeasured factor that contributed to better initial conjunctive representation in the placebo group. However, with US expectancy ratings, acquisition of contextual fear did not moderate retention of contextual fear. Additionally, those who

7

7

6

6

5

5

4

4

3

3

2

2

Context+ Context-

1

1 End Acquisition

Test 1

Time

Test 2

Test 3

End Acquisition

Test 1

Test 2

Test 3

Time

Figure 2. Unconditioned stimulus (US) expectancy to context: end of acquisition to test phase. Shaded area represents day 1, nonshaded area represents day 2. Context, negative context; Context⫹, positive context.

www.sobp.org/journal

D.E. Glenn et al.

Standardized EMG Amplitude

852 BIOL PSYCHIATRY 2014;75:847–854 0.8

0.8

Context+

0.6

0.6

Context-

0.4

0.4

0.2

0.2

0

0

-0.2

-0.2

-0.4

-0.4

-0.6

-0.6

-0.8

-0.8 -1

-1 End Acquisition

Test 1

Test 2

Test 3

Time

End Acquisition

Test 1

Test 2

Test 3

Time

Figure 3. Acoustic startle reflex to context: end of acquisition to test phase. Shaded area represents day 1, nonshaded area represents day 2. Context, negative context; Context⫹, positive context; EMG, electromyogram.

consumed glucose did demonstrate fear acquisition to the context, as measured by the startle reflex, as well as ratings of CS arousal. Furthermore, the superior retention of contextual learning observed in the glucose group relative to the placebo group in the test phase was found in measures of startle reflex, on which the two groups did not differ during acquisition. These findings suggest that glucose consumption may improve normal hippocampal functioning and retention of contextual fear learning in mildly stressed participants. This study does not directly address whether glucose prevents hippocampal impairment following extreme stress or trauma, as it is unlikely that experimental fear conditioning produces neural deficits, but these results raise the possibility that glucose consumption is a simple yet powerful intervention that improves situational specificity of learning and might ultimately be applied to offset the development of overgeneralized contextual fear following trauma exposure. Hippocampal-dependent conjunctive representation tends to overshadow amygdala-dependent elemental learning during contextual fear conditioning, but when hippocampal dysfunction during conditioning impairs conjunctive representation, elemental learning occurs to specific contextual features (26). Trauma-induced fear may be more constrained when traumatic memory is processed in a hippocampal conjunctive representation whose recall is triggered by the combination of multiple contextual features necessary for pattern completion (e.g., late afternoon ⫹ sand ⫹ smoke ⫹ loud noises ¼ flashback) than when trauma recall is triggered by elemental representations of individual contextual features (e.g., late afternoon or sand or smoke or loud noises ¼ flashback). Current understanding of the mechanisms through which glucose modulates memory is incomplete, but there are several possibilities. Raised blood glucose levels may increase synthesis of acetylcholine (a neurotransmitter strongly associated with memory) (53,54), provide additional energy to specific neural components, and modulate neuronal excitability and neurotransmitter release (42). Glucose consumption following stress may cause a reduction in otherwise elevated cortisol levels, thereby bringing the hippocampus back online during encoding and consolidation of traumatic memories. Cortisol increases appetite for carbohydrates and fats, restoring metabolic homeostasis (55), and typically remains at high blood concentrations for an extended period following stress (56–58). Ingestion of glucose following extreme stress may satisfy this metabolic requirement, thereby downregulating cortisol concentrations and sparing normal hippocampal functioning. www.sobp.org/journal

This is the first demonstration of the effects of glucose on hippocampal-dependent contextual fear conditioning processes in humans. There are a number of steps to follow-up on these results. One important issue is the dosage of glucose. The 25 g glucose dosage used in this study was derived from a review of optimal dosage for enhancement of learning in general (52), but optimal dosage may vary for contextual fear learning in particular. Furthermore, extremely intense traumas or stressors that produce particularly high metabolic demand likely require larger amounts of glucose to maintain optimal neural functioning than is required for a laboratory fear-conditioning procedure. Future research should also examine temporal moderation of the effect of glucose on hippocampal-dependent learning processes (e.g., effect of glucose given immediately pretask versus immediately posttask versus 5 hours posttask). Several potential limitations of this research should be noted. First, differences in brain functioning are likely during intense trauma fear versus mild to moderate anxiety during laboratory fear conditioning. It is highly unlikely that experimental fear conditioning produces neural deficits and hippocampal dysfunction. Thus, the findings suggest that glucose may improve normal hippocampal functioning and retention of contextual fear learning, but they do not address whether glucose prevents hippocampal impairment following trauma. The investigation of the potential for glucose consumption to buffer the negative effects of trauma on hippocampal functioning may be best conducted in high-risk populations for trauma exposure, such as soldiers, first responders, and emergency room patients. Second, contextual learning in this study was operationalized by distinguishing between pictures of rooms, but it is debatable if these methods actually reflect context conditioning. As contextual learning occurs to numerous types of stimuli (24), future research should investigate the role of glucose on visuospatial as well as other types of contextual features. More broadly, the field of associative learning would benefit from further discussion and clarification of what constitutes contextual learning, given the numerous qualities used to define contexts (i.e., unpredictability, duration, complex multimodal features) and ways in which contextual learning has been operationalized [i.e., (18–23)]. These findings indicate the need for more research into the role of glucose on hippocampal-dependent contextual fear learning processes and on the potential for a mitigating role of glucose upon the long-term negative consequences of trauma exposure. Though the current findings are preliminary, the

D.E. Glenn et al. potential long-term implications of this line of research are encouraging. If found to be effective in improving contextual specificity of trauma memory, glucose could be provided as a novel means of secondary prevention for populations at high risk for trauma exposure and for recently traumatized individuals. We thank Michael Fanselow, Ph.D., and Moriel Zelikowsky, Ph.D., for consultation in regard to this work. We thank Mickey Rowe from BIOPAC Systems, Inc., for technical support. We also thank Nancy Smith, Chun Man David Yeung, and Miroslava Bernier for their assistance conducting study procedures. Mr. Glenn reports having received research funding from the National Institute of Mental Health. Dr. Craske reports having received royalties from both APA Books and Oxford University Press and research funding from the National Institute of Mental Health and is employed within the University of California system. Dr. Minor reports having received research funds from the Naval Health Research Center and the Stress & Motivated Behavior Institute and is employed within the University of California system. Dr. Vervliet reports no biomedical financial interests or potential conflicts of interest. Supplementary material cited in this article is available online at http://dx.doi.org/10.1016/j.biopsych.2013.09.022. 1. Kessler RC, Berglund P, Demler O, Jin R, Merikangas KR, Walters EE (2005): Lifetime prevalence and age-of-onset distributions of DSM-IV disorders in the National Comorbidity Survey Replication. Arch Gen Psychiatry 62:593–602. 2. Hoge CW, Castro CA, Messer SC, McGurk D, Cotting DI, Koffman RL (2004): Combat duty in Iraq and Afghanistan, mental health problems, and barriers to care. N Engl J Med 351:13–22. 3. Mahan AL, Ressler KJ (2012): Fear conditioning, synaptic plasticity and the amygdala: Implications for posttraumatic stress disorder. Trends Neurosci 35:24–35. 4. Yamamoto S, Morinobu S, Fuchikami M, Kurata A, Kozuru T, Yamawaki S (2008): Effects of single prolonged stress and D-cycloserine on contextual fear extinction and hippocampal NMDA receptor expression in a rat model of PTSD. Neuropsychopharmacology 33: 2108–2116. 5. Choi DC, Rothbaum BO, Gerardi M, Ressler KJ (2010): Pharmacological enhancement of behavioral therapy: Focus on posttraumatic stress disorder. Curr Top Behav Neurosci 2:279–299. 6. Brunet A, Orr SP, Tremblay J, Robertson K, Nader K, Pitman RK (2008): Effect of post-retrieval propranolol on psychophysiologic responding during subsequent script-driven traumatic imagery in post-traumatic stress disorder. J Psychiatr Res 42:503–506. 7. Hoge EA, Worthington JJ, Nagurney JT, Chang Y, Kay EB, Feterowski CM, et al. (2012): Effect of acute posttrauma propranolol on PTSD outcome and physiological responses during script-driven imagery. CNS Neurosci Ther 18:21–27. 8. Chun MM, Phelps EA (1999): Memory deficits for implicit contextual information in amnesic subjects with hippocampal damage. Nat Neurosci 2:844–847. 9. Brewin CR, Dalgleish T, Joseph S (1996): A dual representation theory of post-traumatic stress disorder. Psychol Rev 103:670–686. 10. Brewin CR, Gregory JD, Lipton M, Burgess N (2010): Intrusive images in psychological disorders: Characteristics, neural mechanisms, and treatment implications. Psychol Rev 117:210–222. 11. Sapolsky RM (1992): Stress in the Aging Brain, and the Mechanisms of Neuron Death. Cambridge, MA: MIT Press. 12. Jacobs WJ, Nadel L (1985): Stress-induced recovery of fears and phobias. Psychol Rev 92:512–531. 13. Gold PE (2004): Coordination of multiple memory systems. Neurobiol Learn Mem 82:230–242. 14. Kitayama N, Vaccarino V, Kutner M, Weiss P, Bremner JD (2005): Magnetic resonance imaging (MRI) measurement of hippocampal volume in posttraumatic stress disorder: A meta-analysis. J Affect Disord 88:79–86.

BIOL PSYCHIATRY 2014;75:847–854 853 15. Smith ME (2005): Bilateral hippocampal volume reduction in adults with post-traumatic stress disorder: A meta-analysis of structural MRI studies. Hippocampus 15:798–807. 16. Karl A, Schaefer M, Malta LS, Dörfel D, Rohleder N, Werner A (2006): A meta-analysis of structural brain abnormalities in PTSD. Neurosci Biobehav Rev 30:1004–1031. 17. Acheson DT, Gresack JE, Risbrough VB (2012): Hippocampal dysfunction effects on context memory: Possible etiology for posttraumatic stress disorder. Neuropharmacology 62:674–685. 18. Alvarez RP, Chen G, Bodurka J, Kaplan R, Grillon C (2011): Phasic and sustained fear in humans elicits distinct patterns of brain activity. Neuroimage 55:389–400. 19. Marschner A, Kalisch R, Vervliet B, Vansteenwegen D, Büchel C (2008): Dissociable roles for the hippocampus and the amygdala in human cued versus context fear conditioning. J Neurosci 28:9030–9036. 20. Armony JL, Dolan RJ (2001): Modulation of auditory neural responses by a visual context in human fear conditioning. Neuroreport 12: 3407–3411. 21. Grillon C (2002): Associative learning deficits increase symptoms of anxiety in humans. Biol Psychiatry 51:851–858. 22. Effting M, Kindt M (2007): Contextual control of human fear associations in a renewal paradigm. Behav Res Ther 45:2002–2018. 23. Neumann DL, Lipp OV, Cory SE (2007): Conducting extinction in multiple contexts does not necessarily attenuate the renewal of shock expectancy in a fear-conditioning procedure with humans. Behav Res Ther 45:385–394. 24. Bouton ME, Westbrook RF, Corcoran KA, Maren S (2006): Contextual and temporal modulation of extinction: Behavioral and biological mechanisms. Biol Psychiatry 60:352–360. 25. Rudy JW, Huff NC, Matus-Amat P (2004): Understanding contextual fear conditioning: Insights from a two-process model. Neurosci Biohehav Rev 28:675–685. 26. Maren S, Aharonov G, Fanselow MS (1997): Neurotoxic lesions of the dorsal hippocampus and Pavlovian fear conditioning in rats. Behav Brain Res 88:261–274. 27. Driscoll I, Howard SR, Prusky GT, Rudy JW, Sutherland RJ (2005): Seahorse wins all races: Hippocampus participates in both linear and non-linear visual discrimination learning. Behav Brain Res 164:29–35. 28. Rudy JW, Barrientos RM, O’Reilly RC (2002): Hippocampal formation supports conditioning to memory of a context. Behav Neurosci 116: 530–538. 29. Rudy JW, O’Reilly RC (1999): Contextual fear conditioning, conjunctive representations, pattern completion, and the hippocampus. Behav Neurosci 113:867–880. 30. Minor TR, Saade S (1997): Poststress glucose mitigates behavioral impairment in rats in the “learned helplessness” model of psychopathology. Biol Psychiatry 42:324–334. 31. Choi DW (1992): Excitotoxic cell death. J Neurobiol 23:1261–1276. 32. Clark DD, Sololoff L (1993): Circulation and energy metabolism in the brain. In: Siegel GL, Agranoff BW, Albers RW, Molinoff PB, editors. Brain Neurochemistry. New York: Raven Press, 645–680. 33. Plumb TN, Sterlace SA, Cavanaugh KA, Minor TR (2013): Stress, brain adenosine signaling, and fatigue-related behavioral processes. In: Masino SA, Boison D, editors. Adenosine: A Key Link between Metabolism and Brain Activity. New York: Springer. 34. Schwabe L, Schächinger H, de Kloet ER, Oitzl MS (2010): Stress impairs spatial but not early stimulus-response learning. Behav Brain Res 12: 50–55. 35. Shumake J, Edwards E, Gonzalez-Lima F (2002): Dissociation of septohippocampal metabolism in the congenitally helpless rat. Neuroscience 114:373–377. 36. Horner HC, Packan DR, Sapolsky RM (1990): Glucocorticoids inhibit glucose transport in cultured hippocampal neurons and glia. Neuroendocrinology 52:57–64. 37. Horner HC, Munck A, Leinhard GE (1987): Dexamethasone causes translocation of glucose transporters from the plasma membrane to an intracellular site in human fibroblasts. J Biol Chem 262: 17696–17702. 38. Kaouane N, Porte Y, Vallee M, Brayada-Bruno L, Mons N, Calandreu L, et al. (2012): Glucocorticoids can induce PTSD-like memory impairments in mice. Science 335:1510–1513. 39. Cahill L, Prins B, Weber M, McGaugh J (1994): β-Adrenergic activation and memory for emotional events. Nature 371:702–704.

www.sobp.org/journal

854 BIOL PSYCHIATRY 2014;75:847–854 40. Kennedy DO, Scholey AB (2000): Glucose administration, heart rate and cognitive performance: Effects of increasing mental effort. Psychopharmacology (Berl) 149:63–71. 41. Sünram-Lea SI, Foster JK, Durlach P, Perez C (2002): The effect of retrograde and anterograde glucose administration on memory performance in healthy young adults. Behav Brain Res 134:505–516. 42. McNay EC, Gold PE (2002): Food for thought: Fluctuations in brain extracellular glucose provide insight into the mechanisms of memory modulation. Behav Cogn Neurosci Rev 1:264–280. 43. Flint RW Jr, Riccio DC (1999): Post-training glucose administration attenuates forgetting of passive-avoidance conditioning in 18-day-old rats. Neurobiol Learn Mem 72:62–67. 44. Kopf SR, Baratti CM (1996): Effects of post-training administration of glucose on retention of a habituation response in mice: Participation of a central cholinergic mechanism. Neurobiol Learn Mem 65:253–260. 45. Morgan CA 3rd, Hazlett G, Southwick S, Rasmusson A, Lieberman HR (2009): Effect of carbohydrate administration on recovery from stressinduced deficits in cognitive function: A double-blind, placebocontrolled study of soldiers exposed to survival school stress. Mil Med 174:132–138. 46. Kawachi T, Ishii K, Sakamoto S, Matsui M, Mori T, Sasaki M (2002): Gender differences in cerebral glucose metabolism: A PET study. J Neurol Sci 199:79–83. 47. Janghorbani M, Amini M (2008): Effects of gender and height on the oral glucose tolerance test: The Isfahan diabetes prevention study. Rev Diabet Stud 5:163–170. 48. Gailliot MT, Baumeister RF, DeWall CN, Maner JK, Plant EA, Tice DM, et al. (2007): Self-control relies on glucose as a limited energy source: Willpower is more than a metaphor. J Pers Soc Psychol 92:325–336.

www.sobp.org/journal

D.E. Glenn et al. 49. Beck AT, Ward CH, Mendelson M, Mock J, Erbaugh J (1961): An inventory for measuring depression. Arch Gen Psychiatry 4:561–571. 50. Carver CS, White TL (1994): Behavioral inhibition, behavior activation, and affective responses to impending reward and punishment: The BIS/BAS scales. J Pers Soc Psychol 67:319–333. 51. Fowles DC (1987): Applications of a behavioral theory of motivation to the concepts of anxiety and impulsivity. J Res Pers 21:417–435. 52. Smith MA, Riby LM, van Eekelen JAM, Foster JK (2011): Glucose modulation of episodic memory: A comprehensive review of the glucose memory facilitation effect. Neurosci Biobehav Rev 35:770–783. 53. Durkin TP, Messier C, de Boer P, Westerink BH (1992): Raised glucose levels enhance scopolamine-induced acetylcholine overflow from the hippocampus: An in vivo microdialysis study in the rat. Behav Brain Res 49:181–188. 54. Pych JC, Chang Q, Colon-Rivera C, Gold PE (2005): Acetylcholine release in hippocampus and striatum during testing on a rewarded spontaneous alternation task. Neurobiol Learn Mem 84:93–101. 55. Pasiakos SM, Caruso CM, Kellogg MD, Kramer FM, Lieberman HR (2011): Appetite and endocrine regulators of energy balance after 2 days of energy restriction: Insulin, leptin, ghrelin, and DHEA-S. Obesity (Silver Spring) 19:1124–1130. 56. Haracz JL, Minor TR, Wilkins JN, Zommermann EG (1988): Learned helplessness: An experimental model of the DST in rats. Biol Psychiatry 15:388–396. 57. Swenson RM, Vogel WH (1983): Plasma catecholamine and corticosterone as well as brain catecholamine changes during coping in rats exposed to stressful footshock. Pharmacol Biochem Behav 18:689–693. 58. Weiss JM (1972): Psychological factors in stress and disease. Sci Am 226:104–113.