Abstract

Although increased anxiety and cortisol reactivity can disrupt neural activity and impact cognition, little research has evaluated associations between anxiety, cortisol, and performance on neuropsychological instruments. The current study investigated the relationship between exogenous salivary cortisol activity and self-report state anxiety on measures tapping a variety of cognitive domains. Fifty-eight male participants were randomly assigned to either a control (no stress induction) or an experimental condition simulating testing anxiety. Self-report state anxiety measures and saliva samples were jointly collected on three occasions. The experimental group generally performed worse than controls on declarative memory and working memory tests. Cortisol and self-report anxiety were not correlated. Inverse relationships were demonstrated between self-report anxiety and neuropsychological test scores. Baseline levels of cortisol at session arrival were positively associated with facilitative memory effects, though there was little association between changes in cortisol and cognitive performance. This study highlights the importance of considering the impact of anxiety during neuropsychological evaluations.

Introduction

A number of studies have explored relationships between cortisol, a glucocorticoid stress hormone, and self-report anxiety (Kalpakjian, Farrell, Albright, Chiodo, & Young, 2009; Kirschbaum, Bartussek, & Strasburger, 1992; Polk, Cohen, Doyle, Skoner, & Kirschbaum, 2005). Despite the fact that anxiety and cortisol have been implicated in affecting cognition (Lee et al., 2007; McCormick, Lewis, Somley, & Kahan, 2007), virtually no studies have directly examined potential associations between these factors and performance on neuropsychological measures. Such knowledge could enhance interpretations in both clinical and research contexts. Cortisol release varies by the extent and level of stressor and, in the short term, can alter synapses and dendrites, as well as inhibit neuronal metabolism and glucose transport, although these effects can be reversible (McEwen & Magarinos, 1997). Although resting cortisol provides a general index of hypothalamus–pituitary adrenal axis activation, changes to cortisol production from stressors may be a more accurate measurement of stress reactivity (O'Leary et al., 2007) and would be expected to correlate more with measures sensitive to changes in anxiety rather than tonic levels of trait anxiety. Therefore, the current study was focused on cortisol reactivity and state anxiety in healthy individuals.

There is general consensus that cortisol fluctuations can negatively impact declarative memory, as the hippocampal region contains the greatest proliferation of corticosteroid receptors (Lupien & McEwen, 1997; McEwen & Sapolsky, 1995; Newcomer, Craft, Hershey, Askins, & Bardgett, 1994). Although executive functioning measures have been less frequently utilized in cortisol research (McCormick et al., 2007; Wingenfeld, Wolf, Krieg, & Lautenbacher, 2011), reactivity has been associated with impaired executive functioning, particularly working memory, as corticosteroid receptors are also seated in the prefrontal cortex (Lupien, Gillin, & Hauger, 1999; McAllister-Williams & Rugg, 2002; McCormick et al., 2007; Terfehr et al., 2010). Conversely, neuropsychological performance has been unimpaired on tasks targeting non-declarative memory (Kirschbaum, Wolf, May, Wippich, & Hellhammer, 1996; Lupien et al., 1997) and visuospatial functioning, with corresponding brain regions containing sparse numbers of glucocorticoid receptors (Driscoll, Hamilton, Yeo, Brooks, & Sutherland, 2005; McCormick et al., 2007; McCormick & Teillon, 2001; Zacks, Vettel, & Michelon, 2003).

Despite the aforementioned studies, inconsistent relationships between cortisol and cognition exist in the literature (Domes, Heinrichs, Reichwald, & Hautzinger, 2002; Hoffman & al'Absi, 2004). Cortisol has been found to differentially impact the memory processes of encoding, consolidation, and retrieval (Het, Ramlow, & Wolf, 2005). Method variance may in part contribute to the mixed findings, as cortisol variability can stem from a wide range of participant factors, level of stress, nature of the task, and so on (Het et al., 2005; Kudielka, Hellhammer, & Wüst, 2009; Starcke, Wolf, Markowitsch, & Brand, 2008).

Self-report state and trait anxiety questionnaires can be an important tool during neuropsychological assessments that may guide interpretation. Thoughts of worry have been posited to interfere with concurrent attentional control, diminishing working memory capacity (Eysenck & Derakshan, 2011; Eysenck, Derakshan, Santos, & Calvo, 2007). Elevated state anxiety has been negatively associated with digit-span performance (Darke, 1988), short-term memory capacity (Humphreys & Revelle, 1984), and working memory (Gass and Curiel, 2011), although studies do not always demonstrate relationships (Kizilbash, Vanderploeg, & Curtiss, 2002; Waldstein, Ryan, Jennings, Muldoon, & Manuck, 1997).

There is a need for basic research to further measure if and how stress and anxiety may negatively impact and skew cognitive test findings, as neuropsychologists seek to achieve valid and reliable test results to facilitate diagnostic impressions. Threat of evaluation is prominent for many individuals tested (e.g., loss of independent living, driving privileges, return to employment, etc.), and investigations targeting the effects of “state” anxiety during neuropsychological assessments are limited (Nagra, Skeel, & Penix-Sbraga, 2007). Social stressors, such as speeches, have been popular in the cortisol literature (e.g., Domes et al., 2002; Lupien et al., 1997). However, to enhance the ecological validity, as well as induce and maintain elevated anxiety similar to a testing situation, the present study employed a stress induction procedure that targeted cognitive performance throughout an entire testing battery. Specifically, the study evaluated self-report state anxiety through the State-Trait Anxiety Inventory (STAI)-State and physiological measures of cortisol reactivity through three collection periods during the battery. The STAI was administered three times in order to track self-reported anxiety across time to examine the impact of the stress induction. The overall goal was to evaluate the sensitivity of each method in differentiating stress between experimental conditions, as well as potential associations with cognitive performance corresponding to brain regions with proliferations of glucocorticoids.

Hypotheses

Relative to the control condition, state anxiety, quantified through self-report measures and cortisol activity, was predicted to increase following the stress induction procedure in the experimental condition. Specifically, in the experimental group, cortisol and self-report state anxiety were predicted to increase significantly between baseline time 1 and time 2, as time 2 measurements were 20 min following the stress induction, at peak cortisol activity. Anxiety levels were predicted to remain relatively stable between times 2 and 3 within both groups. The STAI-State and cortisol pairings at each time point were predicted to positively correlate with each other.

Compared with the control group, and in accord with much of the cortisol literature, the experimental condition was hypothesized to display specific deficits with declarative memory (i.e., learning, encoding, retrieval, retention) and more cognitively demanding working memory tasks (i.e., Paced Auditory Serial Addition Task [PASAT], Letter-Number Sequencing). The largest working memory decrements were predicted to occur with progressively more challenging tasks (e.g., later PASAT trials that have shorter time intervals between presented numbers). However, there were no predicted differences between groups on tasks of processing speed and attention, tests of inhibition, or visuospatial rotation, which served as control tasks based on previous literature.

Among the neuropsychological measures predicted to be impacted by anxiety, elevated self-report state anxiety ratings and cortisol reactivity were also predicted to correlate with worse performance on those measures.

Method

Participants

Due to cortisol variability during the menstrual cycle (Kirschbaum, Pirke, & Hellhammer, 1993; Kirschbaum, Kudielka, Gaab, Schommer, & Hellhammer, 1999), women were excluded from the study in order to reduce potential confounding variables. Six participants were excluded from analyses, as they either acknowledged deviation from the instruction protocol (n = 1), were experiencing abnormally high stressors prior to the study evidenced through pre-screening (n = 2), previously experienced loss of consciousness greater than 20 min (n = 2), or had cortisol levels that were outliers (>3 SD above the mean; n= 1). The final sample consisted of 58 male, college students recruited from a psychology subject pool through a mid-sized Midwestern University. Participants were randomly assigned to a stress induction condition (n= 31) or a control condition (n= 27). Ages ranged from 18 to 39, with an average age of 20.49 (SD = 3.44) and an average education of 14.02 (SD = 1.15). Ninety-two percent (n = 54) of the sample was Caucasian, 3% (n = 2) African American, 3% (n = 2) Asian, and the remaining 2% indicated “others.”

Materials

Salivary cortisol is considered a reliable and valid measure of unbound or free cortisol levels in plasma. Oral swabs (30 × 10 mm cylinder) and salivette storage tubes were purchased through Salimetrics and stored at room temperature prior to use. During collection, participants were instructed to place the swab under their tongue on the floor of the mouth for 1–2 min. Swabs were placed in salivettes and within an hour after collection were stored in a lab freezer below −20°C. Three batches of samples were shipped in Styrofoam containers within 24 h to Salimetrics Corporation located in University Park, PA. Each shipment occurred within 1 or 2 months after collection. Samples were assayed for afternoon, non-awakening salivary cortisol levels using a highly sensitive enzyme immunoassay. Duplicate assays were conducted for reliability and results contain the averaged values. The intra- and interassay coefficients of variance were less than 10% and 15%, respectively. There were no duplicate tests that varied by more than 5%, which necessitated repeat testing.

The following measures within the respective domain were administered in the neuropsychological battery: Memory, California Verbal Learning Test-Second Edition (CVLT-II; Delis, Kramer, Kaplan, & Ober, 2000); Attention and inhibition, Color-Word Interference Test from the Delis-Kaplan Executive Function System (D-KEFS; Delis, Kaplan, & Kramer, 2001); Working memory, Letter-Number Sequencing from the WAIS-IV (Weschler, 2008) and the PASAT (Gronwall & Sampson, 1974); Spatial rotation, Mental Rotations Test (MRT-A; Peters et al., 1995); State and trait anxiety, STAI (Spielberger, Gorsuch, & Lushene, 1972); Attention, motor speed, and set-shifting, Trail Making Tests (Reitan & Wolfson, 1985). In addition, all participants were administered a debriefing questionnaire measuring on a scale from 1 to 10 (1 = Not at all, 10 = Very much), the extent to which they believed the explained purpose of the study. Testing order was not randomized, so stress could be equated between measures across subjects.

Procedure

Participants were instructed to abstain from heavy alcohol consumption the night before, as well as to refrain from the following activities for 2 h prior to testing: Ingesting a large meal, chewing gum, brushing teeth, exercising, smoking, or drinking punch, lemonade, or caffeinated beverages. Participants were screened for eligibility at the outset of the experimental session. Those who had not followed the aforementioned protocol were excluded from participation. Participants reporting a significant past or present psychiatric illness were also deemed ineligible, since mood disorders can alter cortisol levels (Holsboer, 2000; Young, Abelson, & Cameron, 2004). Exclusionary criteria also included participants taking anabolic steroids, prescription medications, supplements known to alter cortisol levels (e.g., hydrocortisone creams, dehydroepiandrosterone supplements, use of inhalers), those experiencing abnormally high levels of life stress evidenced through STAI-Trait and/or cortisol levels >3 SD from the mean (i.e., death in family, recently served in Iraq), and individuals who were physically unhealthy at testing time. Lastly, due to cortisol levels being elevated and more variable in the morning (Susman, Dorn, Inoff-Germain, Nottelmann, & Chrousos, 1997), scheduling of participants occurred in the afternoon hours between 12:00 and 5:00 p.m. Because sleep schedules and wake times could not be controlled for participants, and cortisol was not collected throughout the diurnal cycle, investigators focused on cortisol reactivity between baseline and subsequent collections during the testing battery.

Participants were randomly assigned in a between-group design to either a control (no stress induction) or an experimental (stress induction) condition. Testing was completed in one session lasting 90 min. The administration timing of the neuropsychological and anxiety measures was synchronized in both groups. After consent and eligibility screening, baseline time #1 saliva samples and the STAI-State self-report were collected. Participants also completed a background demographic questionnaire before receiving either a stress-inducing or a neutral oral script describing the study. Participants in the experimental condition were read the following script:

Time efficiency is important during neuropsychological evaluations to decrease patient fatigue and such, and the current study is testing a new method for increasing time efficiency. Your performance will dictate which sets of tests will be administered. The tests are constructed so that you will receive fewer problems if you are performing well in order to stream-line the evaluation process; however, if you have difficulty you will receive additional problems to identify specific deficits. Since I can't score the tests as I am administering them, there will be a Ph.D. level neuropsychologist and two graduate level clinical psychology students behind this one-way mirror who will be scoring the tests and evaluating your performance as we go along. Towards the end of the session, they will tell me how you have been performing, which will determine whether you receive more tests. Before we begin, I have to now leave and bring the neuropsychologist and graduate students to the room next door, so they can observe your performance. We wanted to wait so they won't have any judgmental biases of you as an individual, which could affect their scoring.

The control condition received the following neutral instructions used in Nagra and colleagues (2007):

This study involves performing several tasks, such as solving some math problems, examining some pictures, repeating some digits and doing some motor tasks. We are testing some new tests to see how well they work. Therefore we are not concerned about your performance; so do not worry about whether you are doing well or not. Just relax, try to do your best, and follow the instructions provided to you.

Following the experimental script, the examiner exited the room to escort the fictitious judges behind the one-way mirror and returned after a few minutes to begin the neuropsychological battery. All participants were then administered an identical neuropsychological battery, both in measures and timing. The order and timing of the first portion of the neuropsychological battery beginning with the initial instructions was the following: Instructional procedures (0 min), Letter-Number Sequencing (5 min), MRT (10 min), STAI-State #2 and saliva sample #2 (20 min), CVLT-II (trials 1–5, list B, recall; 25 min), and the PASAT (35 min). The second anxiety measure sample was gathered ∼20 min following baseline collection, due to peak cortisol levels generally occurring 20–40 min following the onset of acute stressors in healthy individuals (Dickerson & Kemeny, 2004).

The PASAT concluded 55 min after initial procedural instructions. The examiner then informed participants in the experimental condition they were to now consult with the judges behind the one-way mirror regarding their performance to determine whether more tests would be administered. The examiner then left the room for a few minutes and returned saying, “Your test scores indicate you had difficulty on the tasks, so we are going to administer more tests to identify specific deficits.” Alternatively, following administration of the PASAT in the control condition, the examiner informed participants they realized a test that was to be later administered had been left in another room. This fictitious script was created to retain time consistency between conditions, and the examiner left the room for a few minutes and returned telling the participant the missing materials were obtained.

Participants in both conditions then received identical instructions and procedures for the remainder of the study. The second portion of the neuropsychological battery protocol order and time from the outset of instructional procedures was the following: CVLT-II recall and recognition (60 min), D-KEFS Color-Word Interference (65 min), Trails A and B (75 min), and STAI-State #3 and saliva sample #3 (80 min). The time #3 anxiety collection occurred ∼60 min following time #2 stress collection. Next, a debriefing questionnaire was administered, measuring on a scale from 1 to 10, the extent to which participants believed the explained purpose of the study. Participants were then verbally questioned what they believed the study was evaluating. Next, participants were debriefed regarding the true nature of the study and the rationale for the use of deception, and were given a list of psychological services in the event they felt distressed. Finally, the STAI-Trait measure was administered.

Statistical Analyses

All cognitive test raw scores were converted to z-scores. Analysis of variance (ANOVA) group analyses were conducted using the debriefing questionnaire ratings, cognitive tests, and cortisol and STAI-State data at each time point. Self-report STAI-State and STAI-Trait data were normally distributed and thus raw scores were analyzed. Due to excess kurtosis (>3) in the experimental group's time 2 cortisol sample (kurtosis= 3.7), all cortisol data were log-transformed. Cortisol and STAI-State data were evaluated with repeated measures within and between subjects over the three time gatherings. Post hoc analyses between groups were conducted with the STAI-State and cortisol levels at each time point. Pearson's correlation analyses were conducted between the cortisol and STAI-State ratings from times 1 to 3, as well as neuropsychological performance data.

Results

Debriefing Questionnaire

There were no significant differences in debriefing questionnaire (scale of 1–10) belief ratings of study purpose (e.g., 10 “very much believe”) between the control (mean = 7.26, SD = 2.01) and experimental (mean = 6.97, SD = 1.87) conditions, F(1, 56) = 0.33, p = .57, η2 = 0.01.

Cortisol Measurements

Using repeated-measures ANOVA, with the three cortisol measurements as the dependent variable (DV), there were no differences between groups, F(1, 47)= 0.95, p >.05, η2 = 0.040, main effects within subjects, F(2, 94)= 1.57, p >.05, η2 = 0.032, or interactions over time, F(2, 94)= 0.99, p <.05, η2 = 0.021. Tests of within-subjects contrasts demonstrated a significant quadratic relationship, F(1, 47)= 4.40, p <.05, η2 = 0.086. However, there were no significant linear or quadratic interactions, nor post hoc analyses (Table 1 and Fig. 1).

Table 1.

Raw cortisol assay (μg/dL) descriptive statistics

Measures Control condition (mean [SD]) Experimental condition (mean [SD]) 
Cortisol Time 1 0.20 (0.17) 0.17 (0.14) 
Cortisol Time 2 0.18 (0.13) 0.23 (0.18) 
Cortisol Time 3 0.16 (0.12) 0.19 (0.13) 
Measures Control condition (mean [SD]) Experimental condition (mean [SD]) 
Cortisol Time 1 0.20 (0.17) 0.17 (0.14) 
Cortisol Time 2 0.18 (0.13) 0.23 (0.18) 
Cortisol Time 3 0.16 (0.12) 0.19 (0.13) 

Notes: Control group: Time 1, n = 22; Time 2, n = 25; Time 3, n = 23. Experimental group: Time 1, n= 29; Time 2, n = 29; Time 3, n = 31.

n.s. between groups.

Fig. 1.

Times 1–3 mean raw cortisol assay (μg/dL) values by group.

Fig. 1.

Times 1–3 mean raw cortisol assay (μg/dL) values by group.

State Anxiety Self-Report Measures

Repeated-measures ANOVA, with STAI-State as the DV, revealed that sphericity was not violated. There was no main effect of self-report anxiety across gatherings between groups, F(1, 56) = 1.85, p = .18, η2 = .032. There was an STAI-State × condition interaction within subjects, F(2, 112) = 3.22, p < .05, η2 = 0.054, with the experimental group showing larger increases in anxiety over time. Post hoc analyses revealed the experimental condition endorsed significantly higher self-report anxiety at time 2 compared with controls (Table 2 and Fig. 2), F(1, 56) = 4.79, p <.05, η2 = 0.079. Within subjects, there was a significant linear main effect of STAI-State increasing over time, F(1, 56) = 49.02, p< .01, η2 = 0.467. There were no other group differences.

Table 2.

Raw STAI-State and STAI-Trait self-report descriptive statistics

Measures Control condition (mean [SD]) Experimental condition (mean [SD]) 
State self-report time 1 30.33 (7.07) 30.55 (6.15) 
State self-report time 2* 33.07 (9.12) 37.65 (8.11) 
State self-report time 3 35.89 (10.52) 39.23 (11.48) 
Trait self-report 33.70 (7.87) 35.58 (8.88) 
Measures Control condition (mean [SD]) Experimental condition (mean [SD]) 
State self-report time 1 30.33 (7.07) 30.55 (6.15) 
State self-report time 2* 33.07 (9.12) 37.65 (8.11) 
State self-report time 3 35.89 (10.52) 39.23 (11.48) 
Trait self-report 33.70 (7.87) 35.58 (8.88) 

Notes: Control, n = 27; experimental, n = 31.

*p< .05 between groups.

Fig. 2.

Times 1–3 mean raw STAI-State self-report by group.

Fig. 2.

Times 1–3 mean raw STAI-State self-report by group.

Self-Report Anxiety and Cortisol Correlations

Pearson's correlations revealed strong positive associations within the three STAI-State and STAI-Trait measurements, reflecting method variance (Table 3). Cortisol was not related to state or trait anxiety across time. Correlation analyses within separate conditions were similar to Table 3.

Table 3.

Correlations between self-report anxiety and log-transformed cortisol (μg/dL) assays

 Trait STAI #1 STAI #2 STAI #3 Cortisol #1 Cortisol #2 
STAI #1 .61*      
STAI #2 .42* .65*     
STAI #3 .49* .72* .75*    
Cortisol #1 −.12 −.04 .01 .03   
Cortisol #2 −.01 .07 .07 .06 .58*  
Cortisol #3 −.08 .05 −.08 −.06 .15 .51* 
 Trait STAI #1 STAI #2 STAI #3 Cortisol #1 Cortisol #2 
STAI #1 .61*      
STAI #2 .42* .65*     
STAI #3 .49* .72* .75*    
Cortisol #1 −.12 −.04 .01 .03   
Cortisol #2 −.01 .07 .07 .06 .58*  
Cortisol #3 −.08 .05 −.08 −.06 .15 .51* 

Notes: STAI #1 = State-Trait Anxiety Inventory Time #1 (N = 58); STAI #2 = State-Trait Anxiety Inventory Time #2 (N = 58); STAI #3 = State-Trait Anxiety Inventory Time #3 (N = 58); Trait = STAI Trait (N = 58); Cortisol #1 = Cortisol Assay (μg/dL) Time #1 (N = 51); Cortisol #2 = Cortisol Assay (μg/dL) Time #2 (N = 54); Cortisol #3 = Cortisol Assay (μg/dL)Time #3 (N = 54).

*p< .01.

Neuropsychological Performance Between Conditions

Consistent with hypotheses, one-way ANOVA revealed the experimental group performed significantly worse than controls on the Letter-Number Sequencing test, F(1, 56) = 7.55, p < .01, η2 = 0.119, and the progressively more challenging third, F(1, 56) = 4.59, p < .05, η2 = 0.076, and fourth, F(1, 56) = 4.31, p < .05, η2 = 0.071, PASAT working memory trials (Table 4). On the CVLT-2, compared with the control group, the experimental group freely recalled significantly fewer words over the five trials of word list repetition, F(1, 56) = 4.64, p < .05, η2 = 0.077, a measure of memory encoding. Compared with controls, the experimental condition performed significantly worse during the long-delay free recall of the repeated word list, F(1, 56) = 5.40, p < .05, η2 = 0.088, a measure of memory retrieval, and long-delay free recall was significantly worse than short-delay free recall (long- vs. short-delay free recall), F(1, 56) = 4.17, p < .05, η2 = 0.069. Compared with controls, the experimental group also recalled fewer words when given categorical cues after the short-, F(1, 56) = 4.18, p < .05, η2 = 0.069, and long-delay periods, F(1, 56) = 4.54, p < .05, η2 = 0.075. Performance on the MRT-A, Trails A and B, and D-KEFS Color-Word Interference test were not significantly different between conditions.

Table 4.

Neuropsychological battery descriptive statistics

 Condition
 
   
 Control (mean [SD]) Experimental (mean [SD]) F p η2 
Letter-Number Sequencing 00 (0.66) −0.47 (0.65) 7.55 .01** 0.119 
MRT-A −0.67 (0.83) −0.74 (1.04) 0.07 .80 0.001 
PASAT 
 Trial 1 −0.39 (1.32) −0.95 (1.70) 1.89 .17 0.033 
 Trial 2 −0.02 (0.81) −0.25 (1.12) 0.77 .38 0.014 
 Trial 3 0.13 (0.94) −0.48 (1.18) 4.59 .040.076 
 Trial 4 0.47 (1.21) −0.17 (1.14) 4.31 .040.071 
CW Trial 1 0.27 (0.70) 0.06 (0.65) 1.35 .25 0.024 
CW Trial 2 0.53 (0.70) 0.46 (0.63) 0.15 .70 0.003 
CW Trial 3 0.49 (0.55) 0.33 (0.87) 0.68 .41 0.012 
Trails A 0.27 (0.95) 0.14 (1.72) 0.13 .72 0.002 
Trails B 0.21 (0.67) −0.22 (1.29) 2.47 .12 0.042 
CVLT-2 
 Trials 1–5 0.60 (0.78) 0.05 (0.11) 4.64 .040.077 
 Short-delay free recall 0.15 (1.16) −0.11 (1.30) 0.65 .43 0.011 
 Short-delay cued recall 0.15 (0.91) −0.47 (1.32) 4.18 .050.069 
 Long-delay free recall 0.20 (0.96) −0.53 (1.38) 5.40 .020.088 
 Long-delay cued recall 0.02 (0.88) −0.63 (1.35) 4.54 .040.075 
 Total learning slope: Trials 1–5 −0.17 (0.78) −0.16 (0.88) 0.00 .98 0.000 
 Long- versus short-delay free recall 0.06 (0.70) −0.42 (1.02) 4.17 .050.069 
 Condition
 
   
 Control (mean [SD]) Experimental (mean [SD]) F p η2 
Letter-Number Sequencing 00 (0.66) −0.47 (0.65) 7.55 .01** 0.119 
MRT-A −0.67 (0.83) −0.74 (1.04) 0.07 .80 0.001 
PASAT 
 Trial 1 −0.39 (1.32) −0.95 (1.70) 1.89 .17 0.033 
 Trial 2 −0.02 (0.81) −0.25 (1.12) 0.77 .38 0.014 
 Trial 3 0.13 (0.94) −0.48 (1.18) 4.59 .040.076 
 Trial 4 0.47 (1.21) −0.17 (1.14) 4.31 .040.071 
CW Trial 1 0.27 (0.70) 0.06 (0.65) 1.35 .25 0.024 
CW Trial 2 0.53 (0.70) 0.46 (0.63) 0.15 .70 0.003 
CW Trial 3 0.49 (0.55) 0.33 (0.87) 0.68 .41 0.012 
Trails A 0.27 (0.95) 0.14 (1.72) 0.13 .72 0.002 
Trails B 0.21 (0.67) −0.22 (1.29) 2.47 .12 0.042 
CVLT-2 
 Trials 1–5 0.60 (0.78) 0.05 (0.11) 4.64 .040.077 
 Short-delay free recall 0.15 (1.16) −0.11 (1.30) 0.65 .43 0.011 
 Short-delay cued recall 0.15 (0.91) −0.47 (1.32) 4.18 .050.069 
 Long-delay free recall 0.20 (0.96) −0.53 (1.38) 5.40 .020.088 
 Long-delay cued recall 0.02 (0.88) −0.63 (1.35) 4.54 .040.075 
 Total learning slope: Trials 1–5 −0.17 (0.78) −0.16 (0.88) 0.00 .98 0.000 
 Long- versus short-delay free recall 0.06 (0.70) −0.42 (1.02) 4.17 .050.069 

Notes: Letter-Number Sequencing = Letter-Number Sequencing from the WAIS-IV; MRT-A = Mental Rotations Test; PASAT = Paced Auditory Serial Addition Task; CW Trial 1 = Color Naming subtest from the D-KEFS Color Word Interference Test; CW Trial 2 = Word Reading subtest from the D-KEFS Color-Word Interference Test; CW Trial 3 = Interference subtest from the D-KEFS Color-Word Interference Test; CVLT-2 = California Verbal Learning Test, Second Edition.

*p< .05.

**p< .01.

Neuropsychological Performance and Cortisol Correlations

As shown through Pearson's correlation, higher baseline cortisol was associated with a steeper learning slope over CVLT-2 trials 1–5 (r = .38, p < .01; Table 5). Time 1 baseline cortisol correlated positively with word reading trial 2 on the D-KEFS Color-Word Interference test (r= .29, p< .05).

Table 5.

Correlations between log-transformed cortisol (μg/dL) assays and standardized neuropsychological scores

Measures Time 1 Cortisol Time 2 Cortisol Time 3 Cortisol 
Letter-Number .24 .11 .01 
MRT-A −.07 −.07 .03 
PASAT 
 Trial 1 −.08 −.18 .00 
 Trial 2 .00 −.15 −.04 
 Trial 3 .05 −.15 −.07 
 Trial 4 .10 −.19 .03 
CW Trial 1 .23 −.01 .07 
CW Trial 2 .29* .06 .07 
CW Trial 3 −.09 .03 .11 
Trails A .03 −.03 .01 
Trails B .04 −.01 .22 
CVLT-II 
 Trials 1–5 .22 .06 .00 
 Short-delay free recall .25 .07 .16 
 Short-delay cued recall .24 .04 .02 
 Long-delay free recall .11 .01 .12 
 Long-delay cued recall .21 .00 .08 
 Total learning slope: Trials 1–5 .38** .10 .08 
 Long- versus short-delay free recall −.17 −.08 −.05 
Measures Time 1 Cortisol Time 2 Cortisol Time 3 Cortisol 
Letter-Number .24 .11 .01 
MRT-A −.07 −.07 .03 
PASAT 
 Trial 1 −.08 −.18 .00 
 Trial 2 .00 −.15 −.04 
 Trial 3 .05 −.15 −.07 
 Trial 4 .10 −.19 .03 
CW Trial 1 .23 −.01 .07 
CW Trial 2 .29* .06 .07 
CW Trial 3 −.09 .03 .11 
Trails A .03 −.03 .01 
Trails B .04 −.01 .22 
CVLT-II 
 Trials 1–5 .22 .06 .00 
 Short-delay free recall .25 .07 .16 
 Short-delay cued recall .24 .04 .02 
 Long-delay free recall .11 .01 .12 
 Long-delay cued recall .21 .00 .08 
 Total learning slope: Trials 1–5 .38** .10 .08 
 Long- versus short-delay free recall −.17 −.08 −.05 

Notes: Time 1, n = 51; Time 2, n = 54; Time 3, n = 54. Letter-Number Sequencing = Letter-Number Sequencing from the WAIS-IV; MRT-A = Mental Rotations Test; PASAT = Paced Auditory Serial Addition Task; CW Trial 1 = Color Naming subtest from the D-KEFS Color-Word Interference Test; CW Trial 2 = Word Reading subtest from the D-KEFS Color-Word Interference Test; CW Trial 3 = Interference subtest from the D-KEFS Color-Word Interference Test; and CVLT-2 = California Verbal Learning Test, Second Edition.

*p< .05.

**p< .01.

Neuropsychological Performance and Self-Report Anxiety Correlations

Data from experimental and control groups were combined to explore the overall relationship between anxiety and cognitive measures. Pearson's correlations were generally consistent with the findings of the between-group comparisons, with regard to the sensitivity of neuropsychological measures to anxiety. Time 2 and 3 STAI-State ratings were negatively related to the Letter-Number Sequencing test performance (Table 6). Time 1 and 2 STAI-State were negatively related to CVLT-2 short-delay free call. Higher self-report anxiety at time 2 related to the slower rate of learning on the CVLT-2, over the course of trials 1 through 5. Time 3 STAI-State was negatively related to the D-KEFS Color-Word Interference Color Naming.

Table 6.

Correlations between STAI-State and neuropsychological measures

Measures Time 1 STAI Time 2 STAI Time 3 STAI 
Letter-Number −.11 −.29* −.29* 
MRT −.05 −.23 −.10 
PASAT 
 Trial 1 −.09 −.08 −.18 
 Trial 2 .05 −.04 −.07 
 Trial 3 −.03 −.14 −.11 
 Trial 4 −.06 −.13 −.17 
CW Trial 1 −.25 −.15 −.31* 
CW Trial 2 −.25 −.17 −.19 
CW Trial 3 −.02 −.13 −.17 
Trails A .13 −.07 .02 
CVLT-2 
 Trials 1–5 −.10 −.18 −.11 
 Trial B −.10 −.18 −.14 
 Short-delay free recall −.27* −.30* −.25 
 Short-delay cued recall .00 −.14 −.12 
 Long-delay free recall −.03 −.16 −.19 
 Long-delay cued recall −.02 −.10 −.12 
 Total learning slope: Trials 1–5 −.23 −.26* −.24 
 Long- versus short-delay free recall .34** .18 .08 
Measures Time 1 STAI Time 2 STAI Time 3 STAI 
Letter-Number −.11 −.29* −.29* 
MRT −.05 −.23 −.10 
PASAT 
 Trial 1 −.09 −.08 −.18 
 Trial 2 .05 −.04 −.07 
 Trial 3 −.03 −.14 −.11 
 Trial 4 −.06 −.13 −.17 
CW Trial 1 −.25 −.15 −.31* 
CW Trial 2 −.25 −.17 −.19 
CW Trial 3 −.02 −.13 −.17 
Trails A .13 −.07 .02 
CVLT-2 
 Trials 1–5 −.10 −.18 −.11 
 Trial B −.10 −.18 −.14 
 Short-delay free recall −.27* −.30* −.25 
 Short-delay cued recall .00 −.14 −.12 
 Long-delay free recall −.03 −.16 −.19 
 Long-delay cued recall −.02 −.10 −.12 
 Total learning slope: Trials 1–5 −.23 −.26* −.24 
 Long- versus short-delay free recall .34** .18 .08 

Notes: Letter-Number Sequencing = Letter-Number Sequencing from the WAIS-IV; MRT-A = Mental Rotations Test; PASAT = Paced Auditory Serial Addition Task; CW Trial 1 = Color Naming subtest from the D-KEFS Color-Word Interference Test; CW Trial 2 = Word Reading subtest from the D-KEFS Color-Word Interference Test; CW Trial 3 = Interference subtest from the D-KEFS Color-Word Interference Test; CVLT-2 = California Verbal Learning Test, Second Edition.

*p< .05.

**p< .01.

Discussion

Research has demonstrated that elevated cortisol reactivity and state anxiety may have inverse associations with declarative memory and working memory performance, corresponding to brain regions containing increased glucocorticoid receptors. However, there have been limited investigations evaluating the utility of physiological and subjective measures of state anxiety in a single study targeting neuropsychological performance. The current study employed a stress induction procedure intended to simulate test anxiety in a clinical setting. The induction was effective, as there was an interaction in which the experimental group reported higher levels of anxiety during test administration than the control group. However, contrary to predictions, there were no differences between groups in cortisol over time. Thus, based on self-report, the stress instructions successfully elevated perceptions of anxiety, which leveled off toward the later portion of the battery. This finding appears to be specific to the experimental group, as the control group did not demonstrate similar changes across repeated measurements. There were no associations between cortisol and self-report anxiety, which adds to the literature demonstrating a lack of correspondence between subjective anxiety and physiological measures (Vedhara et al., 2003).

Consistent with the literature (Kirschbaum et al., 1996; Lee et al., 2007; Lupien et al., 1997, 1999; McCormick et al., 2007), the experimental group performed significantly worse than controls on measures of working memory (i.e., Letter-Number Sequencing, PASAT trials 3 and 4), as well as learning and memory (i.e., CVLT-2). In terms of memory performance, compared with controls, the experimental group had poorer memory encoding (trials 1–5 total), retrieval (long-delay free recall, short- and long-delay cued recall), and retention (long- vs. short-delay free recall). There were no group differences on control measures, such as the visuospatial rotation task. Although specific underlying mechanisms are unclear, researchers have proposed that negative self-talk can interfere with attentional control, depleting resources for working memory (Eysenck & Derakshan, 2011; Eysenck et al., 2007). Thus, when given explicit information that their behavior was being observed, individuals in the experimental group may have been more likely to experience more extensive self-talk that impacted performance. In sum, although the small battery limits generalizability to cognitive domains, it appears that anxiety may be an important factor to consider when interpreting test findings in a neuropsychological context. Therefore, it is important to employ valid and reliable measures for capturing anxiety during assessments.

A further finding, somewhat inconsistent with hypotheses, was that baseline cortisol elevations were related to a steeper learning slope during the CVLT-2. Individuals who were more invested in the task or more concerned about their performance from the outset may have presented with higher baseline cortisol levels. Albeit uncommon, cortisol elevations have been shown to at times enhance declarative memory (Domes et al., 2002; Zorawski, Cook, Kuhn, & LaBar, 2005) and memory formation (Shors, Weiss, & Thompson, 1992). Andreano and Cahill (2006) demonstrated a U-shaped relationship between cortisol activity and memory. Nonetheless, other non-physiological mechanisms (e.g., hyper-self-awareness, negative self-talk) could have contributed to performance decrements.

The correlations between self-report anxiety and cognitive performance were consistent with hypotheses and differences seen between the experimental and control groups. Self-report anxiety was negatively associated with Letter-Number Sequencing at times 2 and 3, short-delay memory free recall at times 1 and 2, and CVLT-2 trials 1- 5 learning slope at time 2. The negative relationship between time 3 STAI-State and Color-Word Interference color naming trial 1 was not predicted. Although unlikely, this lower order attentional process may have been impacted by stress. Additionally, the facilitative relationship between STAI-State time 1 and memory retention (long- vs. short-delay free recall) was unexpected. Those with higher baseline state anxiety may have been more motivated, although a greater number of positive relationships would have been expected if this were the case. Findings support that the STAI-State may be a useful measure of anxiety during neuropsychological assessments, although interpretations with cognition should be viewed cautiously.

There were several limitations to the current study. The generalizability of results is limited demographically, as only men were included in the study and the sample was primarily Caucasian. Repeated diurnal cortisol measurements could have also provided more information regarding individual variability in reactivity. Three cortisol and STAI-State measures were gathered during the experiment rather than prior to administration of each measure, making interpretation challenging between cortisol, anxiety with individual test scores, and anxiety measures. Studies may wish to analyze cortisol reactivity during each of the memory phases to better understand the possible facilitative effects. Future research should also take into account the equivocal relationship between physiological and subjective anxiety, especially since cortisol can at times have deleterious or at other times facilitative associations with cognition. In terms of applied recommendations, administering self-report anxiety measures as testing time unfolds may be more useful than completion prior to the battery.

These findings continue to reflect the complicated relationship between cortisol, self-report anxiety, and neuropsychological performance. Elucidating these relationships would have important implications in an applied neuropsychological setting, as threat of evaluation and anxiety are important factors to consider when interpreting test data. State anxiety could potentially have a negative impact on cognitive performance, and self-report measures may have utility in disentangling the effects of anxiety on neuropsychological test performance.

Funding

Central Michigan University dissertation award.

Conflict of Interest

None declared.

References

Andreano
J. M.
Cahill
L.
Glucocorticoid release and memory consolidation in men and women
Psychological Science
 , 
2006
, vol. 
17
 (pg. 
466
-
470
)
Darke
S.
Anxiety and working memory capacity
Cognition and Emotion
 , 
1988
, vol. 
2
 (pg. 
145
-
154
)
Delis
D. C.
Kaplan
E.
Kramer
J. H.
Delis-Kaplan Executive Function System (D- KEFS)
 , 
2001
San Antonio, TX
The Psychological Corp
Delis
D. C.
Kramer
J.
Kaplan
E.
Ober
B.
CVLT-II California Verbal Learning Test, Second Edition, Adult Version
 , 
2000
New York
Psychological Corporation, a Harcourt Assessment Company
Dickerson
S. S.
Kemeny
M. E.
Acute stressors and cortisol responses: A theoretical integration and synthesis of laboratory research
Psychological Bulletin
 , 
2004
, vol. 
130
 
3
(pg. 
355
-
391
)
Domes
G.
Heinrichs
M.
Reichwald
U.
Hautzinger
M.
Hypothalamic-pituitary-adrenal axis reactivity to psychological stress and memory in middle-aged women: High responders exhibit enhanced declarative memory performance
Psychoneuroendocrinology
 , 
2002
, vol. 
27
 (pg. 
843
-
853
)
Driscoll
I.
Hamilton
D. A.
Yeo
R. A.
Brooks
W. M.
Sutherland
R. J.
Virtual navigation in humans: The impact of age, sex, and hormones on place learning
Hormones and Behavior
 , 
2005
, vol. 
47
 (pg. 
326
-
335
)
Eysenck
M. W.
Derakshan
N.
New perspectives in attentional control
Personality and Individual Differences
 , 
2011
, vol. 
50
 (pg. 
955
-
960
)
Eysenck
M. W.
Derakshan
N.
Santos
R.
Calvo
M. G.
Anxiety and cognitive performance: Attentional control theory
Emotion
 , 
2007
, vol. 
7
 
2
(pg. 
336
-
353
)
Gass
C. S.
Curiel
R. E.
Test anxiety in relation to measures of cognitive and intellectual functioning
Archives of Clinical Neuropsychology
 , 
2011
, vol. 
26
 
5
(pg. 
396
-
404
)
Gronwall
D.
Sampson
H.
The psychological assessment
 , 
1974
Auckland, NZ
Auckland University Press/Oxford University Press
Het
S.
Ramlow
G.
Wolf
O. T.
A meta-analytic review of the effects of acute cortisol administration on human memory
Psychoneuroendocrinology
 , 
2005
, vol. 
30
 
8
(pg. 
771
-
784
)
Hoffman
R.
al'Absi
M.
The effect of acute stress on subsequent neuropsychological test performance
Archives of Clinical Neuropsychology
 , 
2004
, vol. 
19
 (pg. 
497
-
506
)
Holsboer
F.
The corticosteroid receptor hypothesis of depression
Neuropsychopharmacology
 , 
2000
, vol. 
23
 
5
(pg. 
477
-
501
)
Humphreys
M. S.
Revelle
W.
Personality, motivation, and performance: A theory of the relationship between individual differences and information processing
Psychological Review
 , 
1984
, vol. 
91
 (pg. 
153
-
184
)
Kalpakjian
C. Z.
Farrell
D. J.
Albright
K. J.
Chiodo
A.
Young
E. A.
Association of daily stressors and salivary cortisol in spinal cord injury
Rehabilitation Psychology
 , 
2009
, vol. 
54
 
3
(pg. 
288
-
298
)
Kirschbaum
C.
Bartussek
D.
Strasburger
C. J.
Cortisol responses to psychological stress and correlations with personality traits
Personality and Individual Differences
 , 
1992
, vol. 
13
 
12
(pg. 
1353
-
1357
)
Kirschbaum
C.
Kudielka
B. M.
Gaab
J.
Schommer
N. C.
Hellhammer
D. H.
Impact of gender, menstrual cycle phase, and oral contraceptives on the activity of the hypothalamus–pituitary–adrenal axis
Psychosomatic Medicine
 , 
1999
, vol. 
61
 (pg. 
154
-
162
)
Kirschbaum
C.
Pirke
K.
Hellhammer
D.
The ‘Trier Social Stress Test’—a tool for investigating psychobiological stress responses in a laboratory setting
Neuropsychobiology
 , 
1993
, vol. 
28
 (pg. 
76
-
81
)
Kirschbaum
C.
Wolf
O. T.
May
M.
Wippich
W.
Hellhammer
D. H.
Stress-and treatment-induced elevations of cortisol levels associated with impaired declarative memory in healthy adults
Life Science
 , 
1996
, vol. 
58
 (pg. 
1475
-
1483
)
Kizilbash
A. H.
Vanderploeg
R. D.
Curtiss
G.
The effects of depression and anxiety on memory performance
Archives of Clinical Neuropsychology
 , 
2002
, vol. 
17
 (pg. 
57
-
67
)
Kudielka
B. M.
Hellhammer
D. H.
Wüst
S.
Why do we respond so differently? Reviewing determinants of human salivary cortisol responses to challenge
Psychoneuroendocrinology
 , 
2009
, vol. 
34
 (pg. 
2
-
18
)
Lee
B. K.
Glass
T. A.
McAtee
M. J.
Wand
G. S.
Bandeen-Roche
K.
Bolla
K. I.
, et al.  . 
Associations of salivary cortisol with cognitive function in the Baltimore memory study
Archives of General Psychiatry
 , 
2007
, vol. 
64
 
7
(pg. 
810
-
818
)
Lupien
S. J.
Gaudreau
S.
Tchiteya
B. M.
Maheu
F.
Sharma
S.
Nair
N. P.
, et al.  . 
Stress-induced declarative memory impairment in healthy elderly subjects: Relationships to cortisol reactivity
Journal of Clinical Endocrinology and Metabolism
 , 
1997
, vol. 
82
 
7
(pg. 
2070
-
2075
)
Lupien
S. J.
Gillin
C. J.
Hauger
R. L.
Working memory is more sensitive than declarative memory to the acute effects of corticosteriods: A dose-response study in humans
Behavioral Neuroscience
 , 
1999
, vol. 
113
 
3
(pg. 
420
-
430
)
Lupien
S. J.
McEwen
B. S.
The acute effects of corticosteriods on cognition: Integration of animal and human model studies
Brain Research Reviews
 , 
1997
, vol. 
24
 (pg. 
1
-
27
)
McAllister-Williams
R. H.
Rugg
M. D.
Effects of repeated cortisol administration on brain potential correlates of episodic memory retrieval
Psychopharmacology
 , 
2002
, vol. 
160
 (pg. 
74
-
83
)
McCormick
C. M.
Lewis
E.
Somley
B.
Kahan
T. A.
Individual differences in cortisol levels and performance on a test of executive function in men and women
Physiology & Behavior
 , 
2007
, vol. 
91
 
1
(pg. 
87
-
94
)
McCormick
C. M.
Teillon
S. M.
Menstrual cycle variation in spatial ability: Relation to salivary cortisol levels
Hormones and Behavior
 , 
2001
, vol. 
39
 (pg. 
29
-
38
)
McEwen
B. S.
Magarinos
A. M.
Stress effects on morphology and function of the hippocampus
Annals of the New York Academy of Sciences
 , 
1997
, vol. 
821
 (pg. 
271
-
284
)
McEwen
B. S.
Sapolsky
R. M.
Stress and cognitive function
Current Opinion in Neurobiology
 , 
1995
, vol. 
5
 
2
(pg. 
205
-
216
)
Nagra
A.
Skeel
R. L.
Penix-Sbraga
T.
A pilot investigation of the effects of stress on neuropsychological performance in Asian-Indians in the United States
Cultural Diversity and Ethnic Minority Psychology
 , 
2007
, vol. 
13
 
1
(pg. 
54
-
63
)
Newcomer
J. W.
Craft
S.
Hershey
T.
Askins
K.
Bardgett
M. E.
Glucocorticoid-induced impairment in declarative memory performance in adult humans
Journal of Neuroscience
 , 
1994
, vol. 
14
 (pg. 
2047
-
2053
)
O'Leary
O. F.
Bechtholt
A. J.
Crowley
J. J.
Hill
T. E.
Page
M. E.
Lucki
I.
Depletion of serotonin and catecholamines block the acute behavioral response to different classes of antidepressant drugs in the mouse tail suspension test
Psychopharmacology
 , 
2007
, vol. 
192
 
3
(pg. 
357
-
371
)
Peters
M.
Laeng
B.
Latham
K.
Jackson
M.
Zaiyouna
R.
Richardson
C.
A redrawn Vandenberg and Kuse Mental Rotations Test: Different versions and factors that affect performance
Brain and Cognition
 , 
1995
, vol. 
28
 (pg. 
39
-
58
)
Polk
D. E.
Cohen
S.
Doyle
W. J.
Skoner
D. P.
Kirschbaum
C.
State and trait affect as predictors of salivary cortisol in healthy adults
Psychoneuroendocrinology
 , 
2005
, vol. 
30
 
3
(pg. 
261
-
272
)
Reitan
R. M.
Wolfson
D.
The Halstead-Reitan Neuropsychological Test Battery: Theory and clinical interpretation
 , 
1985
Tucson, AZ
Neuropsychology Press
Shors
T. J.
Weiss
C.
Thompson
R. F.
Stress-induced facilitation of classical conditioning
Science
 , 
1992
, vol. 
257
 (pg. 
537
-
539
)
Spielberger
C. D.
Gorsuch
R. L.
Lushene
R. E.
State-trait anxiety inventory
 , 
1972
Palo Alto, CA
Consulting Psychologists Press
Starcke
K.
Wolf
O. T.
Markowitsch
H. J.
Brand
M.
Anticipatory stress influences decision making under explicit risk conditions
Behavioral Neuroscience
 , 
2008
, vol. 
122
 
6
(pg. 
1352
-
1360
)
Susman
E. J.
Dorn
L. D.
Inoff-Germain
G.
Nottelmann
E. D.
Chrousos
G. P.
Cortisol reactivity, distress behavior, and behavioral and psychological problems in young adolescents: A longitudinal perspective
Journal of Research on Adolescence
 , 
1997
, vol. 
7
 (pg. 
81
-
105
)
Terfehr
K.
Wolf
O. T.
Schlosser
N.
Fernando
S. C.
Otte
C.
Muhtz
C.
, et al.  . 
Hydrocortisone impairs working memory in healthy humans, but not in patients with major depressive disorder
Psychopharmacology
 , 
2010
, vol. 
215
 (pg. 
71
-
79
)
Vedhara
K.
Miles
J.
Bennett
P.
Plummer
S.
Tallon
D.
Brooks
E.
An investigation into the relationship between salivary cortisol, stress, anxiety and depression
Biological Psychology
 , 
2003
, vol. 
62
 
2
(pg. 
89
-
96
)
Waldstein
S. R.
Ryan
C. M.
Jennings
J. R.
Muldoon
M. F.
Manuck
S. B.
Self- reported levels of anxiety do not predict neuropsychological performance in healthy men
Archives of Clinical Neuropsychology
 , 
1997
, vol. 
12
 
6
(pg. 
567
-
574
)
Weschler
D.
Wechsler Adult Intelligence Scale—Fourth Edition
 , 
2008
San Antonio, TX
Pearson Assessment
Wingenfeld
K.
Wolf
S.
Krieg
J.C.
Lautenbacher
S.
Working memory performance and cognitive flexibility after dexamethasone or hydrocortisone administration in healthy volunteers
Psychopharmacology
 , 
2011
, vol. 
217
 
3
(pg. 
323
-
329
)
Young
E. A.
Abelson
J. L.
Cameron
O. G.
Effect of comorbid anxiety disorders on the hypothalamic-pituitary-adrenal axis response to a social stressor in major depression
Biological Psychiatry
 , 
2004
, vol. 
56
 
2
(pg. 
113
-
120
)
Zacks
J. M.
Vettel
J. M.
Michelson
P.
Imagined viewer and object rotations dissociated with event-related FMRI
Journal of Cognitive Neuroscience
 , 
2003
, vol. 
15
 
7
(pg. 
1002
-
1018
)
Zorawski
M.
Cook
C. A.
Kuhn
C. M.
LaBar
K.
Sex, stress, and fear: Individual differences in conditioned learning
Cognitive, Affective, and Behavioral Neuroscience
 , 
2005
, vol. 
5
 
2
(pg. 
191
-
201
)