Introduction

Posttraumatic stress disorder (PTSD) is a severe and complex mental disorder precipitated by exposure to a catastrophic event involving actual or threatened death or injury, or a threat to the physical integrity of him/herself or others (DSM-5)1. Impairments in cognitive functions, such as emotion, attention, memory and executive control functions are frequently observed in PTSD population2,3,4,5,6.

The capacity for continually monitoring and updating our actions is critical for effective performance in daily life. Response inhibition, the ability to inhibit prepared or proponent behaviors, is the key determinant of successful cognitive and motor control7. Neuroimaging studies have shown that fronto-basal ganglia networks, especially pre-supplementary motor area (preSMA), inferior frontal gyrus (IFG) and anterior cingulate cortex (ACC) are critical for inhibiting motor system and response tendencies8,9,10. Patients with PTSD generally exhibit abnormal top-down inhibitory control in the motor system, i.e., less activation in frontal brain areas and less deactivation of the motor cortex, suggesting there may be a deficiency of frontal inhibitory control to an enhanced motor readiness or an increased prepotency to respond11,12,13.

Studies with event-related brain potentials (ERPs), which have high time resolution, have also demonstrated abnormal inhibitory function at different processing stages in patients with PTSD. Specifically, during a Go/NoGo task, the frontocentral N2 elicited 200–400 ms following NoGo stimuli is considered an earlier step of response inhibition, i.e., detection of the conflict between the internal representation of the Go response and the NoGo stimulus14,15,16. The N2 is followed by a frontocentral P3 which is a positive component typically seen 300–700 ms post-stimulus onset, and this frontocentral P3 might represent a later stage of response inhibition, i.e., response evaluation/decision or response inhibition success16,17. Wu et al.18 found that adolescent earthquake survivors with PTSD exhibited a shorter NoGo-N2 latency than survivors without PTSD when performing a Go/NoGo task. Using a continuous performance task, Shucard et al.19 demonstrated that PTSD veterans had longer NoGo-P3 latency than civilian controls.

Incorporating emotionally salient stimuli into the Go/NoGo paradigm might help us elucidate the inhibition mechanism from the perspective of affective-modulated executive function. Previous studies showed that both implicit and explicit emotion stimulus processing modulates response inhibition20,21,22,23 and that behavioral dysregulation from psychiatric disorders became more prominent within emotional contexts24. From the perspective of PTSD, an exaggerated fear response and an inability to inhibit fear responses after trauma exposure are proposed to be a risk factor for the PTSD patients or an acquired trait of the illness25. The affective Go/NoGo task was proposed as a useful tool to assess psychiatric pathologies characterized by abnormal emotion processing26. Several studies with affective Go/NoGo paradigm in healthy participants have shown that emotional stimuli can influence the performance of response inhibition, as higher false alarm (incorrect NoGo response) rate for positive stimuli than negative and neutral stimuli20,21,23,27. With the method of ERPs, Zhang and Lu28 found smaller amplitudes and shorter latencies of Go-N2 following positive and negative faces than neutral faces, with larger P3 amplitudes and shorter P3 latencies for positive and negative faces than neutral faces in both Go and NoGo trials. On the other hand, Albert et al.20 found that NoGo-P3 amplitude is greater to positive pictures than to negative and neutral pictures. However, to our knowledge, there were no ERPs studies exploring the emotional response inhibition in individuals with posttraumatic stress symptoms to date.

PTSD is a highly heterogeneous clinical syndrome which is composed of distinct symptom clusters. Examining the relationships between cognitive functions, such as response inhibition, and severity of PTSD dimensional symptoms on a continuum rating may both yield higher reliabilities29 and help to further elucidate the psychopathology and core elements that underlie PTSD, in turn improving clinical assessment and intervention. A few studies to date have examined this relationship. For example, with three-factor DSM-IV model (i.e., re-experiencing, avoidance and hyperarousal)30, Swick et al.31 reported that commission errors were most strongly associated with re-experiencing symptoms in a behavior Go/NoGo task. Shucard et al.19 showed that NoGo-P3 latency was positively related to hyperarousal symptoms in Vietnam veterans. On the other hand, Wu et al.32 found that NoGo-P3 latency was positively associated with avoidance symptoms in a five-factor dysphoric arousal PTSD model (i.e., re-experiencing, avoidance, negative alterations in mood and cognitions, dysphoric arousal and anxious arousal) based on the DSM-51.

In the most recent version of the DSM (DSM-5), negative affect and anhedonia are grouped into one cluster as negative alterations in mood and cognitions1. However, the cluster of negative alterations in mood and cognitions is a diverse construct within PTSD, which comprises symptoms involving enhanced negative affect/general distress and symptoms of reduced positive affect/anhedonia. Previous theoretical and empirical studies also suggested that negative affect and positive affect are different constructs33,34,35. More importantly, the Research Domain Criteria (RDoC) project initiated by the National Institutes of Mental Health (NIMH) specified that negative valence and positive valence are two distinct domains in psychopathology field36,37. Based on this evidence, Liu et al.38 proposed a new six-factor model in which the single symptom cluster of “negative alterations in mood and cognitions” from DSM-5 model was further divided into a negative affect symptom cluster (negative affect potentiation) and an anhedonia symptom cluster (positive affect deterioration) in an epidemiological sample of Chinese earthquake survivors. This six-factor model emerged as the best-fitting as compared to the current DSM-5 models38. Moreover, previous work from our own group showed that cortisol activity during cognitive task was only related with negative affect cluster, but not anhedonia cluster in participants who suffered from posttraumatic stress symptoms, providing preliminary evidence for this differentiation of negative affect and anhedonia39.

Thus, the aim of the present study was to explore the relationship between indices of emotional response inhibition (commission error and NoGo related ERP components) and each of the six symptom clusters from newly proposed PTSD six-factor model in trauma-exposed victims by using an emotional Go/NoGo paradigm adapted from Albert et al.20. To increase statistical power and decrease parameter estimation bias40, all participants, instead of only probable PTSD cases, were included in the regression analysis to capture the full range of symptom severity. This approach is also consistent with the RDoC proposal that the biological and clinical variables can be measured in a dimensional way, i.e., on a continuum ranging from normal to pathological41,42 as well as Lobo et al.43 suggestion that traumatized populations without PTSD diagnoses should also be included in dimensional analysis to identify putative EEG biomarkers of posttraumatic stress symptom severity. Based on the hallmark of this six-factor model and result from our previous work (distinction of negative affect and anhedonia), we predicted that the negative affect cluster and anhedonia cluster will have distinct associations with the behavioral performance or amplitude/latency of NoGo-N2/P3.

Results

Descriptive results

Table 1 showed the demographic and clinical variables of the study group (n = 58). The mean age of the participants was 50.16 ± 5.67 yrs. Among the 58 participants, 26 (44.83%) were male. The mean PCL-5 total score was 30.76 ± 16.06. The mean trauma exposure score was 5.12 ± 1.40, and the mean depression score was 41.16 ± 10.30.

Table 1 Demographic and clinical variables of the study group.
Full size table

Behavioral performance

Table 2 showed the behavioral performance in the Go/NoGo task (means and standard deviations (SDs)).

Table 2 Descriptive statistics for the behavioral performance (n = 58).
Full size table

For the commission error (CE), there was a marginally significant valence effect (F(2,114) = 2.717, p = 0.070). Post-hoc analysis showed that participants have significantly lower commission error for negative than neutral pictures (p = 0.008). There was no significant difference between positive and negative pictures (p = 0.154) or between positive and neutral pictures (p = 0.464).

The grand mean of omission error (OE) rate was 1.6 ± 3.0%, and there was no valence effect on OE (F(2,114) = 0.133, p = 0.876).

For the RT of correct go trials, there was a significant valence effect (F(2,114) = 6.4053, p = 0.002). Post-hoc analysis showed that participants had a faster RT for negative/neutral than positive pictures (p = 0.002 and p = 0.008, respectively), but there was no significant difference between negative and neutral pictures (p = 0.781).

ERPs

As illustrated in Fig. 1, the N2 amplitude was larger for NoGo stimuli than for Go stimuli (F(1,57) = 8.843, p = 0.004). The valence main effect was also significant (F(2,114) = 6.233, p = 0.003), and post-hoc analysis showed that negative picture elicited a decreased N2 as compared to both neutral and positive pictures (p = 0.008 and p = 0.003, respectively), while there was no difference between neutral and positive pictures (p = 0.706). The interaction effect between valence and GoNoGo was not significant (F(2,114) = 1.42, p = 0.246).

Figure 1
figure 1

Grand average ERP for the NoGo and Go conditions at FCz under three different valence conditions.

Full size image

For P3 amplitude, NoGo stimuli elicited significantly larger amplitude than Go stimuli (F(1,57) = 123.762, p = 0.000). The valence main effect was also significant (F(2,114) = 5.161, p = 0.007), and post-hoc analysis showed that negative picture elicited significantly larger P3 as compared to both neutral and positive pictures (p = 0.000 and p = 0.039, respectively), with no difference between neutral and positive pictures (p = 0.347). The interaction effect between valence and GoNoGo was not significant (F(2,114) = 0.292, p = 0.747).

There was no significant correlation between Go/NoGo-N2 amplitude and CE/RT in negative valence condition (ps > 0.10). There was a significantly negative correlation between Go-P3 amplitude and RT (r = −0.351, p = 0.007) as well as between NoGo-P3 amplitude and RT (r = −0.381, p = 0.003) in negative valence condition, i.e., the larger of P3 amplitude, the faster participants responded to Go trials under negative context. There was no significant correlation between Go/NoGo-P3 amplitude and CE (ps > 0.10).

Regression analysis

The relationship between total PCL-5 and emotional response inhibition

The multivariate regression analyses showed that total PCL-5 score model predicted 27.3% for NoGo-N2 amplitude under negative valence condition (R2 = 0.273, F(6,57) = 3.197, p = 0.01): PCL-5 score had a positive association with NoGo-N2 amplitude (β = 0.301, t = 2.034, p = 0.047, see Fig. 2). Because N2 is a negative component, thus the positive correlation coefficient, in fact, denotes a negative correlation, i.e., the higher of total PCL-5 score the smaller of NoGo-N2 amplitude in negative context. There was no significant correlation between PCL-5 total score and NoGo-N2 amplitude in neutral or positive condition (ps > 0.10).

Figure 2
figure 2

The partial regression scatter plot of NoGo-N2 amplitude under negative valence condition at FCz site with PCL-5 total score. Note: Because N2 is a negative component, thus the positive correlation coefficient, in fact, denotes a negative correlation, i.e., the higher of PCL-5 total score, the smaller of NoGo-N2 amplitude in the negative context.

Full size image

No significant correlations were found between PCL-5 total score and commission error, NoGo-N2 latency, or NoGo-P3 amplitude/latency in each of the three valence conditions (ps > 0.10).

The relationship between six symptom clusters and emotional response inhibition

Multivariate regression analyses showed that the six-cluster model predicted 41.5% for commission error under positive valence condition (R2 = 0.415, F(11,57) = 2.972, p = 0.005): only anhedonia cluster had a negative association with it (β = −0.476, t = −2.418, p = 0.02, see Table 3 and Fig. 3). None of these six clusters were significantly associated with commission error under neutral or negative valence condition (ps > 0.10).

Table 3 Results of regression analysis for whole participants with commission error under positive valence condition as a dependent variable and six clusters as independent variables (with age, gender, education, trauma exposure, and depression as covariates).
Full size table
Figure 3
figure 3

The partial regression scatter plot of commission error rate under positive valence condition with anhedonia score.

Full size image

None of these six clusters models were significantly associated with NoGo-N2 amplitude/latency in each of three valence conditions (ps > 0.10).

None of these six clusters models were significantly associated with NoGo-P3 amplitude/latency in each of three valence conditions (ps > 0.10).

Discussion

The current study investigated the relationship between posttraumatic stress symptom clusters based on the newly proposed six-factor PTSD model and response inhibition process under the emotional context at the behavioral and neural levels. First of all, the NoGo trials elicited larger N2 and P3 amplitudes than Go trials, which suggested the classical response inhibition effects on ERPs in this emotional Go/NoGo paradigm. Secondly, the results from the present study showed that emotional valence can modulate the response inhibition process: compared to in the neutral or positive context, participants committed fewer errors to NoGo and responded faster to Go in the negative context; there was a smaller N2 amplitude and larger P3 amplitude in the negative context compared to in the neutral or positive context. Furthermore, by utilizing a dimensional approach, we were able to identify distinct relationships between symptom clusters and distinct stages of response inhibition under emotional context within trauma-exposed individuals. We found that PCL-5 total score was negatively correlated with NoGo-N2 amplitude in the negative context, suggesting that PTSD severity is associated with worse conflict detection in the negative context. Furthermore, regression analysis showed that anhedonia, but not negative affect symptom cluster was associated with commission errors, which confirmed our hypothesis that anhedonia and negative affect symptoms should be differentiated.

Picture valence influenced the response inhibition performance. Behaviorally, participants had a lower commission error rate to negative pictures compared to neutral pictures. This result was partially consistent with several previous studies with healthy participants, as indicated by the fewer commission error for negative/neutral than positive stimuli (e.g., words, facial expressions, and pictures)20,21,23,26,27. These results suggested that “approach pleasant and reward-related events might make stopping responses to positive stimuli–more difficult than to other types of emotional stimuli”20, and avoidance of negative events makes stopping responses to negative stimuli easier in traumatized individuals.

On the neural level, ERP results showed that negative pictures elicited smaller N2 amplitude than neutral and positive pictures in traumatized individuals. The frontocentral N2 component has been proposed as an index of executive function, including attention deployment and conflict detection arising from competition between execution and inhibition of a single response in Go/NoGo task10,53,54, and the N2 amplitude increases with the amount of conflict and the extent to which cognitive control resources are recruited to detect current and future conflict53,55. In both normative and clinical populations, negative or threat-related stimuli can automatically capture attention and also interfere with the top-down cognitive control56,57. These negative interference effects have been explained by changes in processing priority that emotional processing is prioritized over cognitive processing when there is conflict or competition in the task58,59. Therefore, in the negative context, the facilitation of irrelevant negative characteristic in frequent Go trials (enhanced emotional engagement) may reduce the cognitive resources to detect the conflict between Go and NoGo trails. This study is among the first to show that N2 is sensitive to the emotional context in traumatized individuals. The reduced N2 amplitude to negative stimuli suggest poor conflict detection and less discrimination between Go and NoGo trials in the negative context.

The negative pictures elicited a larger P3 amplitude compared to neutral or positive stimuli, which was consistent with other studies in healthy population60,61. Furthermore, the correlation analysis showed that the larger P3 amplitude was associated with better behavior performance (RT to Go trials) in the negative context, which echoes the previous finding that larger P3 amplitudes in young adults were associated with more efficient working memory performance (RT and accuracy)62. P3 component reflects higher order cognitive processes such as evaluation of stimulus salience, categorization processing, and the subsequent memory encoding63. P3 amplitude increases with task relevance, motivational significance, arousal level, and top-down mental resource allocation63. Recent reviews proposed that P3 is modulated by dopaminergic activity which is exerted by the locus coeruleus-norepinephrine arousal system64. As stimuli with negative valence would be evaluated as more evolutionarily important, negative stimuli will elicit higher arousal level56 and recruit more top-down attentional resources65, thus leading to a larger P3 amplitude in the negative context. Therefore, the positive association between higher P3 amplitude and better performance under negative context suggested that individuals achieve better performance partially by recruiting more top-down attentional resources to the task.

Previous taxometric investigations have suggested that PTSD might be better classified as the upper extreme of the posttraumatic stress continuum instead of a categorical mental disorder in diagnosis and treatment66,67,68. In the present study, regression analysis within the whole group showed that PCL-5 total scores were negatively associated with NoGo-N2 amplitude only in the negative context. This result was consistent with the previous research that PTSD symptom severity is associated with worse inhibitory performance11,69. Our results further indicated that the overall PTSD symptoms severity was associated with the early stage of inhibition processing, i.e., poorer conflict detection, especially in the negative context.

Interestingly, we found that the severity of anhedonia symptoms was inversely related to commission error to NoGo trials only in the positive context, i.e., individuals with more anhedonia symptoms could hold their response better when response inhibition is needed in the positive context. Previous studies found that individuals with high level of anhedonia reported diminished positive experience to stimuli in a lab70 as well as in daily life situation71. Frewen and colleagues72,73,74 have found that participants with severe trauma exposure reported decreased positive emotion to positive stimuli, and this decreased positive response was related with anhedonia severity and increased amygdala activation. These findings indicated that less processing of positive stimuli or under-engagement of appetitive motivation system in individuals with high anhedonia level could generate less interference when inhibition response is needed, which might explain the association between anhedonia and the lower commission error to positive stimuli in the present study.

In the DSM-5, negative affect and anhedonia are grouped into one cluster as negative alterations in mood and cognitions1. However, within PTSD, negative alterations in mood and cognitions is a diverse construct involving symptoms of enhanced negative affect/general distress and symptoms of reduced positive affect/anhedonia. Furthermore, the RDoC project by the NIMH specified that negative valence and positive valence are two distinct domains in psychopathology field36,37. Previous theoretical and empirical studies also suggested that negative affect and positive affect are two different constructs33,34,35. This study showed that only anhedonia cluster, but not the negative affect cluster, was associated with commission error, which provides electrophysiological support for the current distinction of anhedonia and negative affect symptoms in the 6-factor DSM-5 model.

The present study had several limitations. First, we use a self-report measure, the PCL, to evaluate PTSD symptoms. The clinically administered instruments will be needed in the future studies. Secondly, all the participants were exposed to a deadly earthquake, thus the generalizability of the current findings was limited. Last but not the least, we only measured the response inhibition process among the survivors about five years after the trauma, thus it is hard to discriminate whether our results reflect the vulnerability factor before the trauma or consequences after the trauma. Nevertheless, the emotional response inhibition paradigm we took in this study is a promising first step towards the development of an electrophysiological index to explore the emotional inhibitory control in individuals with posttraumatic stress symptoms.

In conclusion, results from the present study showed that emotional stimuli can influence response inhibition. The negative stimuli impair conflict detection (smaller N2 amplitude) by prioritizing irrelevant emotional processing over cognitive processing on the one hand. On the other hand, negative stimuli facilitated behavioral performance (faster RT) partially by recruiting more top-down attentional resources to the task at hand (larger P3 amplitude). Consistent with prior work32,66,75, our study further emphasized the importance of assessing symptom severity dimensionally to improve our understanding of psychiatric disorders. The clinical–electrophysiological regression results showed that the severity of posttraumatic stress symptoms is associated with impaired inhibition processing at an early stage, i.e., poorer conflict detection in the negative context. Furthermore, the severity of anhedonia rather than negative affect symptom cluster was related with less commission error only in the positive context, which might be associated with less interference from positive emotion by under-engagement of motivational system. This result also provided electrophysiological evidence for the six-factor model, i.e., a distinction should be made between negative affect symptom cluster and anhedonia symptom cluster.

Methods

Participants

Participants were recruited through advertisements posted at local resident communities in Hanwang county. The participants selected were those who had been directly exposed to the devastating earthquake in Wenchuan County, Sichuan Province, China, on May 12, 2008, and with the current age range from 41 to 60 years. We excluded participants by self-report with (1) significant substance abuse, including drug, alcohol, and nicotine; (2) a past or current head injury; (3) self-reported neurological diseases or other serious medical condition. Sixty-one qualified volunteers meeting the inclusion and exclusion standards were selected to participate. The data from three participants were discarded because of too few accepted ERP trials due to poor behavioral performance and/or excessive movement artifacts.

All the participants didn’t take any psychiatric medication for at least four weeks before the experiment, which was conducted from 13–31 December 2013, approximately five and a half years after the earthquake. All the participants reported normal hearing and normal or corrected-to-normal vision. There were four participants reporting left-hand dominance.

Questionnaires

The PTSD Checklist for DSM-5 (PCL-5) used in the current study is a self-report instrument to assess the PTSD symptoms44,45. The PCL-5 is adapted from the original PCL to map onto PTSD symptoms of the DSM-545, and it includes six symptom clusters according to Liu et al.38: re-experience (B1–B5), avoidance (C1–C2), negative affect (D1–D4), anhedonia (D5–D7), dysphoric arousal (E1–E2, E5–E6), and anxious arousal (E3–E4). In the PCL-5, respondents rate all the 20 items from 0 (not at all) to 4 (extremely), to measure the severity of a particular symptom that has bothered them during the past month. The original PCL has been proven, reliable and valid46,47, and the Chinese version has been applied in previous research on earthquake-related PTSD symptomatology38,40.

The Center for Epidemiological Studies-Depression Scale (CESD) was used to assess depressive symptoms in the past week48. The version used here was translated into Chinese and its reliability and validity have been confirmed49.

The severity of trauma exposure was assessed by Trauma Exposure Scale38.

Stimuli

Thirty trauma-unrelated pictures were selected from the International Affective Picture System (IAPS)50, including ten positive pictures, ten negative and ten neutral pictures. These pictures were re-rated by 48 Chinese volunteers, and analysis showed that positive > neutral > negative pictures at the valance value (7.29 ± 0.21 vs. 5.11 ± 0.25 vs. 2.88 ± 0.45, ps < 0.001). Positive pictures have a similar arousal value as negative pictures (5.69 ± 0.65 vs. 5.63 ± 0.75, p > 0.05), and both positive and negative pictures have significantly higher arousal scores than neutral pictures (4.64 ± 1.14) (ps < 0.05). Pictures were edited into three color frames, i.e., red, blue and green, resulting in 90 pictures total. All pictures were presented on a black background in the center of a computer screen at a visual angle of approximately 10° horizontally and 8° vertically.

Procedure

After completing the questionnaires, the participants were seated comfortably in a normally lit room and completed an emotional Go/NoGo task which was adapted from Albert et al.20. After a practice block of 18 trials, two experimental blocks were completed with a short break between the blocks. Each block consisted of 45 pictures and each of them repeated four times, thus 180 trials for each block (1/3 NoGo and 2/3 Go). Each trial started with the presentation of a color framed picture for 300 ms, followed by a 1000–1400 ms interval with a white central fixation-cross presented on a black screen. The participants were required to either press a button (Go) as accurately and quickly as possible when pictures with one of the two color frames (e.g., red and blue: Go stimulus) were presented, or withhold their response (NoGo) when pictures with the other color frame (e.g., green: NoGo stimulus) were presented. Go/Nogo and picture valance conditions were presented in random order. The frame color indicating NoGo was counterbalanced across participants.

EEG Recording and Preprocessing

An electroencephalogram (EEG) was recorded from 64 scalp sites using Ag/AgCl electrodes mounted in an elastic cap (Compumedics Neuroscan, Charlotte, NC). The EEG had an online reference to the left mastoid and an offline algebraic reference to the average of the left and right mastoids. The vertical and horizontal electrooculograms were recorded from two pairs of electrodes. One pair was placed above and below the left eye, and the other, 10 mm from the outer canthi of each eye. Interelectrode impedance was maintained at <5 kΩ. The signals were amplified with a 0.05–100 Hz bandpass filter and digitized at 500 Hz.

The EEG data were digitally filtered using a 30-Hz low-pass filter and were epoched into periods of 1000 ms (including a 200 ms prestimulus baseline) time-locked to the onset of the picture. Ocular artifacts were removed from the EEG signal using a regression procedure available through Neuroscan software. Trials with various artifacts were rejected if they exceeded the criterion of ±70 μV. The ERPs from both the Go and NoGo conditions were individually averaged. Behaviorally incorrect trials were not included in the ERP analysis.

The peak amplitudes and latencies of the frontocentral N2 and frontocentral P3 were measured at the Fz (frontal region), FCz (frontocentral region), and Cz (central region) sites. The N2 was defined as the minimum voltage in the time-window between 200 and 400 ms after stimulus onset, and P3 was defined as the maximum voltage in the time-window between 300 and 500 ms after stimulus onset. These sites and time windows were chosen in agreement with the previous literature32,51,52.

Data Analysis

All statistical analyses were conducted using SPSS software (version 19.0). Descriptive statistics were gathered from the questionnaire scores and behavioral data, including the reaction time (RT) in correct trials, the rate of omission errors for the Go trials, and the commission error (CE) for the NoGo trials in each of all three valence picture conditions. Repeated measures analysis of variance (ANOVA) was carried out on behavioral performance and latency/amplitude at the three measurement sites (FZ, FCz, and Cz) under each of the three valence conditions (negative, neutral and positive). The Greenhouse–Geisser correction for degrees of freedom was applied when the sphericity assumption was violated. For cases where repeated measures ANOVA procedures revealed a significant main effect, post hoc analyses of least square difference were used to examine the specific effects and significant levels. To explore if there were relationships between behavior performance and electrophysiological index of response inhibition, the Person correlation coefficiency was calculated between N2/P3 amplitude and CE/RT.

Multivariate regression analyses were conducted to examine the associations between the six symptom clusters and indices of emotional response inhibition (behavioral performance and ERP components to NoGo condition). Amplitudes/latencies for the NoGo-N2 and NoGo-P3 components extracted at FCz site under each of the three valences (positive, neutral, and negative) were treated as dependent variables in regression analysis, as the amplitudes for both N2 and P3 were largest at this site. The symptom scores were treated as predictors, and demographic variables (age, gender, education level) and clinical variables (trauma exposure and depression) were entered as covariates in the regression analyses. The same method was also used to test the relationship between total PTSD severity and emotional response inhibition. All p values below 0.05 were considered statistically significant, and the tests were two-tailed.

Ethics

All the participants gave written informed consent and were paid for their participation. The experiment was approved by the Ethics Committee of Human Experimentation at the Institute of Psychology, Chinese Academy of Sciences. The experiment was conducted in accordance with relevant guidelines and regulations.