Brain and Cognition 94 (2015) 60–67

Contents lists available at ScienceDirect

Brain and Cognition journal homepage: www.elsevier.com/locate/b&c

Anticipatory processes under academic stress: An ERP study Duan Hongxia a,b, Yuan Yiran a,b, Yang Can a,b, Zhang Liang a, Zhang Kan b, Wu Jianhui a,⇑ a b

Key Laboratory of Behavioral Science, Institute of Psychology, Chinese Academy of Sciences, Beijing, China University of Chinese Academy of Sciences, Beijing, China

a r t i c l e

i n f o

Article history: Accepted 10 January 2015 Available online 5 February 2015 Keywords: Long-term academic stress Anticipation State anxiety ERP CNV

a b s t r a c t It is well known that preparing for and taking high-stakes exams has a significant influence on the emotional and physiological wellbeing of exam-takers, but few studies have investigated the resulting cognitive changes. The current study examined the effect of examination-induced academic stress on anticipation in information processing. Anticipation was indexed using the contingent negative variation (CNV). Electroencephalograms (EEG) were collected from 42 participants using the classic S1–S2 paradigm. These participants were preparing for the Chinese National Postgraduate Entrance Exam (NPEE). EEGs were also collected from 21 age-matched, non-exam comparison participants. The levels of perceived stress and state anxiety were higher and both the initial CNV (iCNV) and the late CNV (lCNV) were more negative in the exam group than in the non-exam group. These results suggest that participants under academic stress experienced greater anticipation of upcoming events. More important, for the non-exam group, state anxiety was positively related to both the iCNV and lCNV amplitude, and this correlation existed when trait anxiety was controlled; however, there was no such relationship in the exam group. These results suggested that the cortical anticipatory activity in the high-stressed exam group reached the maximum ceiling, leaving little room for transient increases in state anxiety. Ó 2015 Elsevier Inc. All rights reserved.

1. Introduction Academic examinations have long been used to study how reallife stressors lead to psychological and physiological changes. Studies have shown that academic examinations can elicit both psychological responses, such as stress (Backovic, Zivojinovic, Maksimovic, & Maksimovic, 2012; Halamandaris & Power, 1999), anxiety (Borella et al., 1999; Conley & Lehman, 2012; Herbert, Moore, de la Riva, & Watts, 1986; Spangler, 1997) or depression (Kurokawa et al., 2011; Liu & Lu, 2012), and physiological responses, including changes in cardiovascular activity (Conley & Lehman, 2012; Papousek et al., 2010), hypothalamic–pituitary– adrenal axis activity (Bardi, Koone, Mewaldt, & O’Connor, 2011; Verschoor & Markus, 2011) and the immune system (Kamezaki, Katsuura, Kuwano, Tanahashi, & Rokutan, 2012; Koh et al., 2012). Studies exploring cognitive changes that occur during an examination period have found facilitatory and detrimental effects on higher cognitive functions, such as attention, working memory and executive function (Kofman, Meiran, Greenberg, Balas, & Cohen, 2006; Mogg, Bradley, & Hallowell, 1994; Vedhara, Hyde,

⇑ Corresponding author at: Institute of Psychology, Chinese Academy of Sciences, 16 Lincui Road, Chaoyang District, Beijing 100101, China. Fax: +86 10 64836047. E-mail address: [email protected] (J. Wu). http://dx.doi.org/10.1016/j.bandc.2015.01.002 0278-2626/Ó 2015 Elsevier Inc. All rights reserved.

Gilchrist, Tytherleigh, & Plummer, 2000). Liston, McEwen, and Casey (2009) examined the effect of long-term stress elicited by exam preparation on the attentional neural network. Their results suggested that exposure to one month of exam-related stress disrupts attention shifts and functional connectivity within the frontoparietal network. The predominant focus of these studies has been on neurocognitive response and behavioral output after stimulus presentation. Before a stimulus is presented, however, the brain has begun to anticipate and prepare for upcoming events. It is therefore possible that the anticipatory cognitive processes are also influenced by examination-induced academic stress. The anticipation of future events is an important adaption in human beings. ‘‘These anticipatory processes enable us to prepare for and consider the potential consequences of forthcoming events rather than respond to such events in a purely reactive manner’’ (Wynn, Horan, Kring, Simons, & Green, 2010). Contingent negative variation (CNV) is the most extensively studied anticipatory eventrelated potential (ERP). It is recorded from the scalp after a warning stimulus (S1) has been presented while the participant is waiting for an imperative stimulus (S2) and preparing to respond (Walter, 1967). According to Birbaumer, Elbert, Canavan, and Rockstroh (1990), CNV represents the activity of the dendritic trees of cortical pyramidal neurons and its amplitudes manifest the availability of resources of related neural systems. From the per-

H. Duan et al. / Brain and Cognition 94 (2015) 60–67

spective of psychological processes, CNV amplitude is related to anticipation, attention, motivation and motor preparation (for reviews, see Brunia & van Boxtel, 2001; van Boxtel & Böcker, 2004). When the interstimulus intervals are extended to 3 s, two CNV components can be distinguished: the initial wave (iCNV) and the late wave (lCNV) (Kropp, Kiewitt, Göbel, Vetter, & Gerber, 2000; Loveless, 1973). The iCNV amplitude is thought to be modulated by physiological arousal and anticipation of S2 (McCallum, 1988; Simons, Macmillan Iii, & Ireland, 1982). The lCNV is thought to reflect neural activities that prepare for goaldirected movement (McCallum, 1988; Sanquist, Beatty, & Lindsley, 1981). Both iCNV and lCNV have a scalp distribution over the front-central areas, with lCNV distributed relatively more posterior and broader than iCNV (Cui et al., 2000; Fischer, Langner, Diers, Brocke, & Birbaumer, 2010; Gaillard, 1976; Hamano et al., 1997; van den Bosch, 1983). The CNVs have a potential significance in clinical application, i.e., to evaluate the changes in cognitive functions occurring in various diseases and mental states. Previous CNV studies examined clinical samples with anticipatory deficits, such as schizophrenia (Wynn et al., 2010); posttraumatic stress disorder (Kimble, Ruddy, Deldin, & Kaufman, 2004) and migraine (Siniatchkin, Averkina, & Gerber, 2006; Siniatchkin, Gerber-von, Darabaneanu, Stephani, et al., 2011; Siniatchkin, Sandor, Schoenen, & Gerber, 2003); motor disorders, such as Parkinsonism (Oishi, Mochizuki, Du, & Takasu, 1995); or attention deficits, such as attention deficit hyperactivity disorder (Banaschewski et al., 2003). Electrophysiological studies also revealed that CNV can be modulated by stress and anxiety in otherwise healthy samples. Previous studies found that higher CNV was associated with relatively higher stress and arousal levels (Brown, Fenwick, & Howard, 1989; Nagai et al., 2004; Tecce, 1972). The results of a study by Siniatchkin et al. (2006) also showed that healthy women under experimental achievement stress had increased iCNV amplitude compared with those under control conditions. More interestingly, studies found that high-trait anxious individuals had a greater CNV than lowtrait anxious individuals who performed at comparable levels (Ansari & Derakshan, 2011; Glanzmann & Froehlich, 1984). They explained that anxious individuals use more processing resources to prevent decrements in performance at the expense of processing efficiency (Ansari & Derakshan, 2011). These studies utilized laboratory-induced acute stress/anxiety (Glanzmann & Froehlich, 1984; Siniatchkin et al., 2006) or specific anxiety population samples (Ansari & Derakshan, 2011), and relatively little is known about how natural, long-term psychosocial stressors modulate the anticipation step of information processing, as indexed by CNV. Some studies focused on the individual difference of trait anxiety in the relationship between CNV amplitude and stress/state anxiety. In a series experiments (Knott & Irwin, 1967; Knott & Irwin, 1968; Knott & Irwin, 1973; Van Veen, Peters, Knott, Miller, & Cohen, 1973), a larger CNV was found in low-trait anxiety subjects when the S2 was an electrical shock compared with the neutral stimulus; in high anxiety subjects, however, CNV was reduced. Similarly, McCallum and Papakostopoulos (1973) found that as a function of increasing task complexity, CNV increases in low-anxiety subjects and decreases in high-anxiety subjects. Glanzmann and Froehlich (1984) also suggested that the reduction of CNV is associated with an increase of state anxiety in high-anxiety subjects; no such relationship is found in low anxiety subjects. There have been two interpretations suggested for these results. Knott and Irwin (1967) suggested a ‘‘ceiling hypothesis,’’ meaning highanxiety persons operate this anticipation cognition from a higher baseline than low-anxiety persons, leaving little room for additional negativity. The ‘‘distraction hypothesis’’ proposed by Tecce (1971), Tecce (1972), however, suggest that high-anxiety persons might be more distracted by worry and preoccupation with current

61

task-irrelevant, self-focused thoughts when S2 is anticipated. These studies have usually differentiated participants by selfreport scales. It is still unclear how a natural, specific type and long-term stressor (e.g., academic stress) modulates the relationship between CNV amplitude and state anxiety. Therefore, the aim of our study was to improve the field’s understanding of the mechanisms involved in CNV under the conditions produced by real-life stressors. For this purpose, CNV was used within the classic S1–S2 paradigm to measure anticipatory processing in a homogeneous sample composed of 42 male students who had been preparing for an exam for a long time and 21 age- and education-matched non-exam male students. We collected electroencephalograms (EEG) 11–25 days before the examination from graduating male students who were preparing to participate in the National Postgraduate Entrance Exam (NPEE), which is a highly competitive and time-consuming event. According to the results of previous studies, we predicted that: (1) the exam group will produce iCNV and/or lCNV of greater amplitude than those of the non-exam group; (2) the relationship between state anxiety and CNV amplitude would be different between the groups, for example, there will be a positive correlation between CNV amplitude and state anxiety and/or stress level in the nonexam group, while there might be a negative or no correlation for the exam group, based on both the ceiling hypothesis and the distraction hypothesis. 2. Methods 2.1. Participants Because of the effects of gender on stress and electrophysiological activity (Crasson, Lembreghts, el Ahmadi, Legros, & TimsitBerthier, 2001; Knott & Peters, 1974), we recruited only male graduating participants. Sixty-three right-handed, healthy, male college students were recruited through advertisements posted at Wannan Medical College. All participants were medical students from the same university, and all of them had passed the same entrance requirements (e.g., the university entrance exam), which minimized the possibility that the observed group differences can be explained by major or intelligence related factors. Forty-two of these participants were planning to take the NPEE, while the remaining students did not plan to participate in any academic examinations or interviews one month before or after the experiment. None of the participants had a history of neurological or psychiatric disorders. All participants were screened using the Life Events Scale (LES) (Tennant & Andrews, 1976; Zhang et al., 1987) to exclude students who had experienced any other major stressor during the past month. Since there are evidences that the different personality characteristics have different CNV patterns (Brown et al., 1989; Werre, Mattie, & Berretty, 2001), we also measured their personality scores, as described in the questionnaires section, to ensure the homogeny of the two groups in personality characteristics. All of the students reported normal hearing and normal or corrected-to-normal vision, and all of them right-handed based on self-report. Three participants in the exam group were excluded from the analyses due to that the number of acceptable EEG trials was no less than five (Kropp et al., 2000). The final sample consisted of 39 exam-group participants and 21 non-exam group participants. The exam group and non-exam control group were matched for age (mean age of exam-group: 22.5 ± 1.0; non-exam group: 22.6 ± 1.1) and education level (17 years of education for all participants). All participants gave written informed consent and were paid for their participation. This experiment was approved by the Ethics Committee of Human Experimentation at the Institute of Psychology, Chinese Academy of Sciences.

62

H. Duan et al. / Brain and Cognition 94 (2015) 60–67

2.2. Procedure This study reports the results obtained from a larger study addressing the relationship between academic stress and cortisol response/cognition (Duan et al., 2013). All of the qualified participants underwent the experiment between December 12 and 27, 2011, a period between 11 and 25 days before the NPEE. Upon arrival in the lab, the participants completed questionnaires. They were then prepared for the EEG recording by being fit with an electrode cap, after which they were seated in a relaxed position on a comfortable chair in a sound-attenuated room. A slightly altered paradigm of Walter’s classical one (Walter, 1967) with a three-second interstimulus interval was used in our study, which has also been applied by many other studies (e.g., Bareš, Nestrašil, & Rektor, 2007; Kropp et al., 2000; Siniatchkin et al., 2011). The students completed a training session of five trials and a testing session of 40 trials. Each trial began by displaying a fixation point (a ‘‘+’’ sign) in the center of the screen for 1000 ms. Next, a medium frequency warning tone, S1 (frequency of 1000 Hz, duration of 50 ms, 5-ms rise and fall time), was presented via an earphone. This tone was followed exactly 3 s later by an imperative stimulus, S2 (frequency of 2000 Hz, duration of 1000 ms, 25-ms rise and fall time). The sound intensity was adjusted to a comfortable listening level of approximately 75 dB SPL using Adobe Audition (version 1.0). The participants were required to quickly press the ‘‘right’’ button on the keyboard with the index finger of their dominant hand when they heard the S2, and the trial was terminated as soon as the participants responded. The duration time from the onset of S2 to the key press of the participants was collected as reaction time (RT). Intertrial intervals varied randomly from 5 to 8 s. 2.3. Questionnaires Each participant completed the Cohen’s Perceived Stress Scale (10-item version) (PSS) (Cohen & Williamson, 1988), which is widely utilized to evaluate chronic stress levels (Liston et al., 2009; Tomiyama, Dallman, & Epel, 2011). The PSS has been demonstrated to be a reliable and valid instrument (Roberti, Harrington, & Storch, 2006), and its good reliability and validity has also been shown in many studies with different Chinese sample populations (Wang et al., 2011; Yu & Ho, 2010). The State-Trait Anxiety Inventory (Spielberger, 1983), which is one of the most universal scales used to measure anxiety (Gotlib & Cane, 1989), was also administered. We also collected the participants’ personality characteristics by administering the Mini International Personality Item Pool (the Mini-IPIP, including openness, conscientiousness, extraversion, agreeableness, and neuroticism) (Donnellan, Oswald, Baird, & Lucas, 2006). The information about the length of time the students had expended preparing for the exam before participating in the experiment (not including the time spent on purchasing review information or completing school applications) and the intensity of their study habits (the time spent reviewing every day) were also collected. To explore whether a higher state of alertness would translate into better exam performance, the students in the exam group were contacted after the exam to collect information on performance. Considering the variability of specialty tests taken among the participants, we report only the total public test scores of English and Political Science (see the Section 3).

Impedance was below 5 KX. The EEG was amplified with a bandpass filter of 0.05–100 Hz and digitized using a sampling rate of 500 Hz. Due to the interference of ocular potentials, horizontal eye movements were monitored from electrodes placed at the outer canthus of each eye, and vertical movements were monitored separately from electrodes located above and below the left eye. The EEG data were processed offline using Neuroscan 4.3 software. Ocular artifacts were removed using a regression procedure built into the Neuroscan software. The data were low-pass filtered using a cutoff frequency of 30 Hz, and epochs of 4000 ms in duration (including 500 ms before S1, which served as a baseline for baseline correction, and 500 ms after S2) were extracted. An artifact was rejected automatically if the signal amplitude exceeded ±100 lV.

2.5. Analysis FC3, FCz and FC4 were selected to measure iCNV and lCNV. The iCNV was calculated as the mean value obtained from the 500-ms time-window from 450 to 950 ms after the S1 presentation, and the lCNV, from the 500 ms time window just before the S2 presentation (i.e., from 2500 to 3000 ms after the S1 presentation). These sites and time windows were chosen in agreement with the existing literature (Cui et al., 2000; Fischer et al., 2010; Funderud et al., 2013; Gaillard, 1976; Hamano et al., 1997; van den Bosch, 1983) and the visual inspection of the data conducted to determine where and when CNVs were maximal (see Fig. 2). A repeated-measures ANOVA with a between-subject factor of group (exam vs. non-exam group) and within-subject factor of the site (FC3, FCz and FC4) was used to analyze the iCNV and lCNV data. Trait anxiety may have an effect on CNV amplitude (Ansari & Derakshan, 2011); thus, we also applied the covariance analysis with trait anxiety score as a covariate. Responses that were behaviorally too slow or too quick (±3 standard deviation) were excluded from the analysis. The Greenhouse–Geisser correction was used to adjust for sphericity violations. The Bonferroni correction was applied for multiple comparisons. Pearson correlation analyses were also conducted to assess the relationship between CNV amplitudes and subjective stress measures. To control the multiple comparison problems, we used the FCz to explore the relationship between questionnaire scores and the iCNV amplitude, and FC3 was used to explore the relationship between questionnaire scores and the lCNV amplitude, where

2.4. EEG recording and preprocessing EEG data were continuously recorded by 64 cap-mounted Ag/ AgCl electrodes arranged according to the international 10–20 system (Neuroscan Inc., USA), with an on-line reference to the left mastoid and an offline algebraic re-reference to the linked mastoid.

Fig. 1. The mean values of trait anxiety, PSS and state anxiety, with error bars representing the standard error of the mean. Notes: PSS: Scores of the Perceived Stress Scale. ⁄p < .05, ⁄⁄p < .01.

H. Duan et al. / Brain and Cognition 94 (2015) 60–67

63

Fig. 2. Grand averaged ERPs and their scalp distributions. Upper: averaged contingent negative variation (CNV) recorded at FC3, FCz and FC4 for both groups. The gray areas represent the time windows of the measured mean amplitude of iCNV (450–950 ms) and lCNV (2500–3000 ms, i.e., 500 ms before S2). Lower: the scalp distributions of the measured iCNV/lCNV for both groups. The black dots represent these three analyzed electrodes (FC3, FCz and FC4).

there were maximal amplitudes. To clarify the relationship between state stress and CNV, we also applied the partial correlation between CNV amplitudes and state stress level (i.e., state anxiety and PSS) with trait anxiety controlled. Before ANOVAs and Pearson correlations were conducted, we tested the homogeneity of variances between the two groups, and the results showed they had constant variances (Fs < 1.83, p > 0.1). Thus, we assumed that the unequal sample sizes might not affect the ANOVAs and Pearson correlations in this study. Comparisons of the psychological measurements between the exam and non-exam groups were made using independent sample ttests. All reported p values are two-tailed.

There were no significant differences of the mean RTs (mean ± S.D.: 228.05 ± 25.71 ms vs. 228.51 ± 21.18 ms, t(58) = .07, p > .1) and accuracy rate (mean ± S.D.: 0.97 ± 0.03 vs. 0.98 ± 0.03, t(58) = .38, p > .1) between the exam and non-exam groups. The review duration and intensity results showed that participants in the exam group spent an average of 6.15 ± 2.44 months and 9.56 ± 1.56 h per day preparing for the exam. As for the exam performance in the exam group, the total scores were 113.57 ± 14.45 (range: 75–136; the minimum and maximum possible scores were 0 and 200, respectively) based on thirty-five student’s reports.

3. Results

3.2. CNV data

3.1. Descriptive statistics

The CNV waveforms obtained from the exam and non-exam groups are shown in Fig. 2. The iCNV amplitude of the exam group was significantly larger than that of the non-exam group ( 2.63 ± 3.45 vs. 0.55 ± 3.96; F(1, 58) = 4.45, p < .05), and no other significant main effect or interaction effect was found. The lCNV amplitude of the exam group was significantly larger than that of the non-exam group ( 0.95 ± 5.36 vs. 2.18 ± 5.77; F(1, 58) = 4.42, p < .05), and there was a main effect of sites (F(2) = 6.40, p < .01). Pairwise comparisons revealed that both FC3 and FC4 had significantly more negative lCNVs than FCz (ps < 0.05), but the difference between FC3 and FC4 did not achieve significance. When the trait anxiety was controlled, the difference in iCNV amplitude between the two groups did not achieve significance (F(1, 57) = 2.126, p > .1)

The results of the psychological measures are shown in Fig. 1. The exam group was found to have higher trait anxiety (t(43) = 3.49, p < .01), higher state anxiety scores (t(55) = 3.43, p < .01) and higher levels of perceived stress (t(27) = 2.35, p < .05) than the non-exam group. There were no significant differences between the exam group and non-exam group on the five subscales of the Big Five Inventory (|t|s < 1.66, ps > .1) (extraversion: 7.82 ± 2.54 vs. 7.25 ± 2.79; agreeableness: 11.90 ± 1.89 vs. 11.60 ± 1.43; conscientiousness: 11.51 ± 2.19 vs. 10.85 ± 2.08; neuroticism: 7.08 ± 2.08 vs. 6.35 ± 1.27; openness: 10.18 ± 2.58 vs. 10.70 ± 2.16).

64

H. Duan et al. / Brain and Cognition 94 (2015) 60–67

and the difference in lCNV amplitude between the two groups became marginally significant (F(1, 57) = 3.47, p < .1). 3.3. Correlation analysis For the non-exam group, the state anxiety level was positively related to the iCNV amplitude (r = 0.508, p < .05), but not the lCNV amplitude (r = 0.294, p > .1) (Fig. 3, upper panel). PSS and trait anxiety level had no significant correlation with iCNV/lCNV amplitude (|r|s 6 0.339, ps > .1). Furthermore, when trait anxiety level was partialled out, state anxiety showed a significantly positive correlation with both iCNV (r = 0.523, p < .05) and lCNV amplitude (r = 0.457, p < .05), but the correlation between PSS and iCNV/lCNV amplitude did not achieve significance (|r|s 6 0.220, ps > .1). For the exam group, there was a significant positive correlation between PSS and lCNV amplitude (r = 0.338, p < .05), but when trait anxiety level was controlled, this correlation did not achieve significance (r = 0.231, p > .1). There were no other significant correlation in this group (|r|s 6 0.254, ps > .1) (Fig. 3, lower panel). The exam performance did not significantly relate to the amplitude of iCNV or lCNV (|r|s 6 0.165, ps > .1). 4. Discussion In the present study, we tested the anticipatory processes under academic stress by measuring CNVs. To pass the NPEE, participants in this study spent an average of 9.6 h each day for approximately

6 months preparing. The data showed that the exam group has a greater level of trait anxiety and experienced greater stress and state anxiety levels when compared with the non-exam group. This was accompanied by more negative shifts in iCNV/lCNV in the exam group, but no clear difference in RT to S2 between the two groups. Furthermore, the iCNV and lCNV amplitude were positively related to state anxiety in the non-exam group but not the exam group. The data showed that the exam group had more negative iCNV/ lCNV than the non-exam group. This result is in accordance with that found by Siniatchkin et al. (2006), who revealed that experimental achievement stress increased iCNV and lCNV amplitude in healthy women. To avoid the potential confounding factor of hormone changes during the menstrual cycle in women, in the present study, we chose not to study women and used a sample of healthy male graduating students. In addition, a long-term real-life stressor rather than laboratory-induced acute stress was also applied to explore the external validity of CNV under real-life stress. The iCNV has been interpreted as a cortical component of anticipation, and its amplitude is thought to be modulated by physiological arousal and expectancy-related attention (McCallum, 1988; Simons et al., 1982). The lCNV is believed to be dependent on the cognitive preparation of a motor response to the imperative stimulus (Leynes, Allen, & Marsh, 1998; Ruchkin, Sutton, Mahaffey, & Glaser, 1986). The results of our study suggested that participants in the exam group might be more alert to and may anticipate more the upcoming events in their environment and also be more prepared to respond. These alterations in

Fig. 3. Scatter plots of the correlation between the state anxiety and iCNV/lCNV amplitude in the non-exam (upper) and exam group (lower). Note: Because CNV is a negative component, the negative correlation trend, in fact, denotes a positive correlation.

H. Duan et al. / Brain and Cognition 94 (2015) 60–67

both components of anticipatory processing may happen even under response to a relatively neutral, sustained, and repetitive stimulus. Nevertheless, we did not find altered behavioral performance in the exam group, suggesting that increased alertness does not lead to better behavioral performance. This result was consistent with a report by Ansari and Derakshan (2011), who found that high-trait anxious individuals had a greater CNV than low-trait anxious individuals while performing at comparable levels in the anti- and prosaccade tasks. They explained it as impaired processing efficiency in the high-trait anxiety individual. Haier et al. (1988), Haier, Siegel, Tang, Abel, and Buchsbaum (1992) showed that glucose consumption in the brain in higher intelligent individuals is lower when compared to less intelligent individuals, and they explained this as the neural efficiency hypothesis. Our study suggested that individuals experiencing higher academic stressor consume more neural resources (as indexed by larger iCNV and lCNV) and have lower neural efficiency under a simple auditory detection task by pressing a button compared with individuals at a relatively normal stress level. More importantly, we found that there was a difference between the exam and non-exam group in the relationship between state anxiety and CNV amplitude. For the non-exam group, there was a positive correlation between the students’ iCNV/lCNV amplitude and the state anxiety score. This result was consistent with literature that reports that a higher CNV was associated with a relatively higher state stress and arousal level (Brown et al., 1989; Nagai et al., 2004; Tecce, 1972), suggesting that transient increases in state anxiety led to heightened cortical excitability during anticipatory processing. Most interestingly, there was no such correlation in the exam group. The literature has reported the individual difference of trait anxiety in the relationship between CNV amplitude and state anxiety (Glanzmann & Froehlich, 1984; Knott & Irwin, 1967; Knott & Irwin, 1968; Knott & Irwin, 1973; McCallum & Papakostopoulos, 1973; Van Veen et al., 1973). Both the ceiling hypothesis (Knott & Irwin, 1967) and the distraction hypothesis (Tecce, 1971; Tecce, 1972) have been proposed for these results. Our results (i.e., the exam group having higher amplitudes of CNVs when compared with the non-exam individuals, but the amplitude in exam individuals not increasing as a function of state anxiety) favor the ceiling hypothesis. Individuals under academic stress may have already reached maximum cortical activity during anticipatory processing; thus, transient increases in state anxiety become less effective in modulating CNV magnitude. Our findings have some important implications for individuals experiencing long-term major stressor, such as students facing major academic stressors. Exposure to a long-term stressor increases the vulnerability to adverse outcomes, including mental problems, physical diseases and cognitive decrement (Schmidt, Sterlemann, & Muller, 2008). Scientists have long been interested in understanding the process how long-term stress ‘‘gets under the skin’’ to affect the brain and health outcomes. Before getting on the skin, long-term stress may change our brain function by impairing neural efficiency and leading to ceiling effect of cortical activity. Some means of early intervention might be reasonable for individuals who are otherwise healthy, but experiencing academic stressors to break this progression before the coming of visual outcomes brought by long-term major stress. Our results also revealed that the exam group has a significantly higher trait anxiety level than the non-exam group, which confounds the results in the sense that any group differences might not be due to stress elicited by the exam. The difference in the iCNV/lCNV amplitude between the groups reduced from significant to insignificant or marginally significant when the trait anxiety was controlled, suggesting these group differences were contributed to by trait anxiety to some extent, thus limiting our theoretical inter-

65

pretation for CNV amplitude differences between the groups. This issue, however, probably did not hinder our conclusion on academic stress. One possible fact is that both the trait and state characteristics may contribute to this high academic stress and it may be difficult to disentangle them from one another. Higher trait anxiety would reinforce the motivational system (Burton, 1971; Calapoglu et al., 2011), and individuals with higher trait anxiety might have more achievement motivation to take the NPEE. The literature has suggested that high trait anxiety is associated with an exaggerated stress response (Landgraf & Wigger, 2002; Watson & Clark, 1984). More importantly, results of correlation analysis between CNV amplitude and state anxiety were achieved when the trait anxiety level was controlled; thus, conclusions from the correlation analysis were not confounded by trait anxiety. Our study had several other limitations. First, regarding generalizability, our study focuses on an examination stressor in young male students only. Therefore, it is possible that the effects of academic stress observed in this study may not be generalizable to other samples. Second, we only measured the emotional and cortical response during the examination preparation period; however, how these psychological and electrophysiological responses change over time is still unknown. Thus, a longitudinal study should be conducted. In conclusion, our data suggest that preparing for an academic examination leads to negative emotional responses in students, such as stress and anxiety. This academic stress increases the students’ anticipation of and cortical excitability for the upcoming stimulus, as indexed by increases in both the iCNV and lCNV, but without any behavioral benefit for the students. Furthermore, within the non-exam group, increased anticipation for upcoming events may be a function of a transient increase in state anxiety, whereas there was no such relationship in the exam group, suggesting that the cortical anticipatory activity in individuals with major academic stress have reached the maximum ceiling, leaving little room for transient increases in state anxiety. Acknowledgments This work was supported by the National Natural Science Foundation of China (91124003, 81371203 and 31100734) and the Excellent Young Scientists Fund, IP, CAS (YICX63S03). References Ansari, T. L., & Derakshan, N. (2011). The neural correlates of cognitive effort in anxiety: Effects on processing efficiency. Biological Psychology, 86, 337–348. Backovic, D. V., Zivojinovic, J. I., Maksimovic, J., & Maksimovic, M. (2012). Gender differences in academic stress and burnout among medical students in final years of education. Psychiatria Danubina, 24, 175–181. Banaschewski, T., Brandeis, D., Heinrich, H., Albrecht, B., Brunner, E., & Rothenberger, A. (2003). Association of ADHD and conduct disorder–brain electrical evidence for the existence of a distinct subtype. Journal of Child Psychology and Psychiatry, 44, 356–376. Bardi, M., Koone, T., Mewaldt, S., & O’Connor, K. (2011). Behavioral and physiological correlates of stress related to examination performance in college chemistry students. Stress, 14, 557–566. Bareš, M., Nestrašil, I., & Rektor, I. (2007). The effect of response type (motor output versus mental counting) on the intracerebral distribution of the slow cortical potentials in an externally cued (CNV) paradigm. Brain research bulletin, 71, 428–435. Birbaumer, N., Elbert, T., Canavan, A. G., & Rockstroh, B. (1990). Slow potentials of the cerebral cortex and behavior. Physiological Reviews, 70, 1–41. Borella, P., Bargellini, A., Rovesti, S., Pinelli, M., Vivoli, R., Solfrini, V., et al. (1999). Emotional stability, anxiety, and natural killer activity under examination stress. Psychoneuroendocrinology, 24, 613–627. Brown, D., Fenwick, P., & Howard, R. (1989). The contingent negative variation in a Go/No Go avoidance task: Relationships with personality and subjective state. International Journal of Psychophysiology, 7, 35–45. Brunia, C. H. M., & Van Boxtel, G. J. M. (2001). Wait and see. International Journal of Psychophysiology, 43, 59–75. Burton, E. C. (1971). State and trait anxiety, achievement motivation and skill attainment in college women. Research Quarterly, 42, 139–144.

66

H. Duan et al. / Brain and Cognition 94 (2015) 60–67

Calapoglu, M., Sahin-Calapoglu, N., Karacop, A., Soyoz, M., Elyildirim, U. Y., & Avsaroglu, S. (2011). Serotonin transporter Bi- and triallelic genotypes and their relationship with anxiety and academic performance: A preliminary study. Neuropsychobiology, 63, 103–111. Cohen, S., & Williamson, G. (1988). Perceived stress in a probability sample of the United States. In S. Spacapan & S. Oskamp (Eds.), The social psychology of health: Claremont symposium on applied social psychology (pp. 31–67). Newbury Park, CA: Sage Publications Inc. Conley, K. M., & Lehman, B. J. (2012). Test anxiety and cardiovascular responses to daily academic stressors. Stress and Health, 28, 41–50. Crasson, M., Lembreghts, M., el Ahmadi, A., Legros, J. J., & Timsit-Berthier, M. (2001). Interindividual variability of contingent negative variation. Clinical Neurophysiology, 31, 300–320. Cui, R. Q., Egkher, A., Huter, D., Lang, W., Lindinger, G., & Deecke, L. (2000). High resolution spatiotemporal analysis of the contingent negative variation in simple or complex motor tasks and a non-motor task. Clinical Neurophysiology, 111, 1847–1859. Donnellan, M. B., Oswald, F. L., Baird, B. M., & Lucas, R. E. (2006). The mini-IPIP scales: Tiny-yet-effective measures of the big five factors of personality. Psychological Assessment, 18, 192–203. Duan, H., Yuan, Y., Zhang, L., Qin, S., Zhang, K., Buchanan, T. W., et al. (2013). Chronic stress exposure decreases the cortisol awakening response in healthy young men. Stress, 16, 630–637. Fischer, T., Langner, R., Diers, K., Brocke, B., & Birbaumer, N. (2010). Temporo-spatial dynamics of event-related EEG beta activity during the initial contingent negative variation. PLoS One, 5, e12514. Funderud, I., Lovstad, M., Lindgren, M., Endestad, T., Due-Tonnessen, P., Meling, T. R., et al. (2013). Preparatory attention after lesions to the lateral or orbital prefrontal cortex – An event-related potentials study. Brain Research, 1527, 174–188. Gaillard, A. W. (1976). Effects of warning-signal modality on the contingent negative variation (CNV). Biological Psychology, 4, 139–154. Glanzmann, P., & Froehlich, W. D. (1984). Anxiety, stress, and contingent negative variation reconsidered. Annals of the New York Academy of Sciences, 425, 578–584. Gotlib, I. H., & Cane, D. B. (1989). Self-report assessment of depression and anxiety. In P. C. Kendall & D. Watson (Eds.), Anxiety and depression: Distinctive and overlapping features. Personality, psychopathology, and psychotherapy (pp. 131–169). San Diego, CA, US: Academic Press. Haier, R. J., Neuchterlein, K. H., Haxlett, E., Paek, J., Browning, H. L., & Buchsbaum, M. S. (1988). Cortical glucose metabolic rate correlates of abstract reasoning and attention studied with positron emission tomography. Intelligence, 12, 199–217. Haier, R. J., Siegel, B., Tang, C., Abel, L., & Buchsbaum, M. S. (1992). Intelligence and changes in regional cerebral glucose metabolic rate following learning. Intelligence, 16, 415–426. Halamandaris, K. F., & Power, K. G. (1999). Individual differences, social support and coping with the examination stress: A study of the psychosocial and academic adjustment of first year home students. Personality and Individual Differences, 26, 665–685. Hamano, T., Luders, H. O., Ikeda, A., Collura, T. F., Comair, Y. G., & Shibasaki, H. (1997). The cortical generators of the contingent negative variation in humans: A study with subdural electrodes. Electroencephalography and clinical neurophysiology, 104, 257–268. Herbert, J., Moore, G. F., de la Riva, C., & Watts, F. N. (1986). Endocrine responses and examination anxiety. Biological Psychology, 22, 215–226. Kamezaki, Y., Katsuura, S., Kuwano, Y., Tanahashi, T., & Rokutan, K. (2012). Circulating cytokine signatures in healthy medical students exposed to academic examination stress. Psychophysiology, 49, 991–997. Kimble, M., Ruddy, K., Deldin, P., & Kaufman, M. (2004). A CNV-distraction paradigm in combat veterans with posttraumatic stress disorder. The Journal of Neuropsychiatry and Clinical Neurosciences, 16, 102–108. Knott, J. R., & Irwin, D. A. (1967). Anxiety, stress and the contingent negative variation. Electroencephalography and clinical neurophysiology, 22, 188. Knott, J. R., & Irwin, D. A. (1968). Anxiety, stress and contingent negative variation (CNV). Electroencephalography and clinical neurophysiology, 24, 286–287. Knott, J. R., & Irwin, D. A. (1973). Anxiety, stress, and the contingent negative variation. Archives of General Psychiatry, 29, 538–541. Knott, J. R., & Peters, J. F. (1974). Changes in CNV amplitude with progressive induction of stress as a function of sex. Electroencephalography and Clinical Neurophysiology, 36, 47–51. Kofman, O., Meiran, N., Greenberg, E., Balas, M., & Cohen, H. (2006). Enhanced performance on executive functions associated with examination stress: Evidence from task-switching and Stroop paradigms. Cognition & Emotion, 20, 577–595. Koh, K. B., Lee, Y. J., Beyn, K. M., Chu, S. H., Kim, D. M., & Seo, W. Y. (2012). Effects of high and low stress on proinflammatory and antiinflammatory cytokines. Psychophysiology, 49, 1290–1297. Kropp, P., Kiewitt, A., Göbel, H., Vetter, P., & Gerber, W. D. (2000). Reliability and stability of contingent negative variation. Applied Psychophysiology and Biofeedback, 25, 33–41. Kurokawa, K., Tanahashi, T., Murata, A., Akaike, Y., Katsuura, S., Nishida, K., et al. (2011). Effects of chronic academic stress on mental state and expression of glucocorticoid receptor a and b isoforms in healthy Japanese medical students. Stress, 14, 431–438. Landgraf, R., & Wigger, L. (2002). High vs low anxiety-related behavior rats: An animal model of extremes in trait anxiety. Behavior Genetics, 31, 301–314.

Leynes, P. A., Allen, J. D., & Marsh, R. L. (1998). Topographic differences in CNV amplitude reflect different preparatory processes. International Journal of Psychophysiology, 31, 33–44. Liston, C., McEwen, B. S., & Casey, B. J. (2009). Psychosocial stress reversibly disrupts prefrontal processing and attentional control. Proceedings of the National Academy of Sciences of the United States of America, 106, 912–917. Liu, Y., & Lu, Z. (2012). Chinese high school students’ academic stress and depressive symptoms: Gender and school climate as moderators. Stress and Health, 28, 340–346. Loveless, N. E. (1973). The contingent negative variation related to preparatory set in a reaction time situation with variable foreperiod. Electroencephalography and Clinical Neurophysiology, 35, 369–374. McCallum, W. C., & Papakostopoulos, D. (1973). The CNV and reaction time in situations of increasing complexity. Electroencephalography and Clinical Neurophysiology, Suppl. 33, 179–185. McCallum, W. C. (1988). Potentials related to expectancy, preparation and motor activity. In T. W. Picton (Ed.), Handbook of electroencephalography and clinical neurophysiology. Human event-related potentials (pp. 427–534). Elsevier Science Publishers. Mogg, K., Bradley, B. P., & Hallowell, N. (1994). Attentional bias to threat: Roles of trait anxiety, stressful events, and awareness. The Quarterly Journal of Experimental Psychology Section A: Human Experimental Psychology, 47, 841–864. Nagai, Y., Critchley, H. D., Featherstone, E., Fenwick, P. B., Trimble, M. R., & Dolan, R. J. (2004). Brain activity relating to the contingent negative variation: An fMRI investigation. Neuroimage, 21, 1232–1241. Oishi, M., Mochizuki, Y., Du, C., & Takasu, T. (1995). Contingent negative variation and movement-related cortical potentials in Parkinsonism. Electroencephalography and Clinical Neurophysiology, 95, 346–349. Papousek, I., Nauschnegg, K., Paechter, M., Lackner, H. K., Goswami, N., & Schulter, G. (2010). Trait and state positive affect and cardiovascular recovery from experimental academic stress. Biological Psychology, 83, 108–115. Roberti, J. W., Harrington, L. N., & Storch, E. A. (2006). Further psychometric support for the 10-item version of the perceived stress scale. Journal of College Counseling, 9, 135–147. Ruchkin, D. S., Sutton, S., Mahaffey, D., & Glaser, J. (1986). Terminal CNV in the absence of motor response. Electroencephalography and Clinical Neurophysiology, 63, 445–463. Sanquist, T. F., Beatty, J. T., & Lindsley, D. B. (1981). Slow potential shifts of human brain during forewarned reaction. Electroencephalography and Clinical Neurophysiology, 51, 639–649. Schmidt, M. V., Sterlemann, V., & Muller, M. B. (2008). Chronic stress and individual vulnerability. Stress, Neurotransmitters, and Hormones: Neuroendocrine and Genetic Mechanisms, 1148, 174–183. Simons, R. F., Macmillan Iii, F. W., & Ireland, F. B. (1982). Anticipatory pleasure deficit in subjects reporting physical anhedonia: Slow cortical evidence. Biological Psychology, 14, 297–310. Siniatchkin, M., Averkina, N., & Gerber, W. D. (2006). Relationship between precipitating agents and neurophysiological abnormalities in migraine. Cephalalgia, 26, 457–465. Siniatchkin, M., Gerber-von, Mül G., Darabaneanu, S., Petermann, F., Stephani, U., et al. (2011). Behavioural treatment programme contributes to normalization of contingent negative variation in children with migraine. Cephalalgia, 31, 562–572. Siniatchkin, M., Sandor, P., Schoenen, J., & Gerber, W. D. (2003). Contingent negative variation (CNV) as a functional genetic marker in migraine without aura. Cephalalgia, 23. 636–636. Spangler, G. (1997). Psychological and physiological responses during an exam and their relation to personality characteristics. Psychoneuroendocrinology, 22, 423–441. Spielberger, C. D. (1983). Manual for the state-trait anxiety inventory. Palo Alto, CA: Consulting Psychologists Press. Tecce, J. J. (1971). Contingent negative variation and individual differences: A new approach in brain research. Archives of General Psychiatry, 24, 1–6. Tecce, J. J. (1972). Contingent negative variation (CNV) and psychological processes in man. Psychological Bulletin, 77, 73–108. Tennant, C., & Andrews, G. (1976). A scale to measure the stress of life events. Australasian Psychiatry, 10, 27–32. Tomiyama, A. J., Dallman, M. F., & Epel, E. S. (2011). Comfort food is comforting to those most stressed: Evidence of the chronic stress response network in high stress women. Psychoneuroendocrinology, 36, 1513–1519. van Boxtel, G. J., & Böcker, K. B. (2004). Cortical measures of anticipation. Journal of Psychophysiology, 18, 61–76. van den Bosch, R. J. (1983). Contingent negative variation and psychopathology: Frontal–central distribution, and association with performance measures. Biological Psychiatry, 18, 615–634. Van Veen, W. J., Peters, J. F., Knott, J. R., Miller, L. H., & Cohen, S. I. (1973). Temporal characteristics of the contingent negative variation: Relationships with anxiety, perceptual mode, sex, and stress. Biological Psychiatry, 7, 101–111. Vedhara, K., Hyde, J., Gilchrist, I. D., Tytherleigh, M., & Plummer, S. (2000). Acute stress, memory, attention and cortisol. Psychoneuroendocrinology, 25, 535–549. Verschoor, E., & Markus, C. R. (2011). Affective and neuroendocrine stress reactivity to an academic examination: Influence of the 5-HTTLPR genotype and trait neuroticism. Biological Psychology, 87, 439–449. Walter, W. (1967). Electrical signs of association, expectancy and decision in the human brain. Electroencephalography and Clinical Neurophysiology, Suppl. 25, 258+.

H. Duan et al. / Brain and Cognition 94 (2015) 60–67 Wang, Z., Chen, J., Boyd, J. E., Zhang, H., Jia, X., Qiu, J., et al. (2011). Psychometric properties of the Chinese version of the perceived stress scale in policewomen. PLoS One, 6, e28610. Watson, D., & Clark, L. A. (1984). Negative affectivity: The disposition to experience aversive emotional states. Psychological Bulletin, 96, 465–490. Werre, P. F., Mattie, H., & Berretty, E. W. (2001). Contingent negative variation, extraversion, reaction time and drug effects. Personality and Individual Differences, 30, 1083–1094.

67

Wynn, J. K., Horan, W. P., Kring, A. M., Simons, R. F., & Green, M. F. (2010). Impaired anticipatory event-related potentials in schizophrenia. International Journal of Psychophysiology, 77, 141–149. Yu, R., & Ho, S. C. (2010). Psychometric evaluation of the perceived stress scale in early postmenopausal Chinese women. Psychology, 1, 1–8. Zhang, M. Y., Fan, B., Cai, G. J., Chi, Y. F., Wu, W. Y., & Jin, H. (1987). Life event scale: Norm result. Chinese Journal of Nervous Mental Diseases, 13, 70–73.