Paranormal believers show reduced resting EEG beta band oscillations and inhibitory control than skeptics

Paranormal believers’ thinking is frequently biased by intuitive beliefs. Lack of inhibition of these tempting beliefs is considered a key element in paranormal believers’ thinking. However, the brain activity related to inhibitory control in paranormal believers is poorly understood. We examined EEG activities at resting state in alpha, beta, and gamma bands with inhibitory control in paranormal believers and skeptics. The present study shows that paranormal belief is related to the reduced power of the alpha, beta, and gamma frequency bands, and reduced inhibitory control. This study may contribute to understanding the differences between believers and skeptics in brain activity related to inhibitory control in paranormal believers.


Participants.
A total of 20 (10 female, 10 male; mean age = 22.50 years, SD = 4.07, age range 19-34) healthy right-handed (Edinburgh Handedness Inventory) students were selected for the study. We were looking for students who believed in or were skeptical about paranormal beliefs. The participants received gifts for participating in the present study.

Materials.
The current study uses the Revised Paranormal Beliefs Scale (RPBS), which is used widely to assess belief in the paranormal. The RPBS consists of 26 items grouped in seven subscales, i.e., Traditional Religious Beliefs, Psi, Witchcraft, Superstition, Spiritualism, Extraordinary Life Forms, and Precognition. The items are rated on a seven-point Likert scale (from 1 = Strongly Disagree to 7 = Strongly Agree) 76 . Items are presented as phrases (e.g., "If you break a mirror, you will have bad luck"), and respondents evaluate the answers on a seven-point Likert scale according to how strongly they agree with them. The RPBS can give scores that corresponds to the subscales, and the aggregate yields a general estimate of the degree of paranormal belief. The RPBS subscales and the measurement show sufficient validity and reliability 76,79 . Internal consistency in previous study was excellent, with Chronbach's α = 0.92 80 . The Cognitive Failures Questionnaire (CFQ) is a self-report questionnaire measuring failures in memory, perception, and motion in daily life. The CFQ consists of 25 questions, and the participants answer the items on a five-point Likert scale (from 0 = never to 4 = always). "Do you neglect to listen to people's names while you are meeting them?" is an example of a question. A high score shows a greater propensity for cognitive failure 81 . It has been demonstrated that CFQ has adequate validity and reliability, including high internal consistency and test-retest stability, and multidimensional factor structure [81][82][83][84] .
The Go/No-Go Task is a behavioral measure of inhibitory control 85 . On a computer monitor, this task was presented. A fixation stimulus in the monitor's center always appeared before the visual Go and No-Go stimuli. The participants were instructed to press a button as rapidly as possible in response to the green rectangle that appeared in the center, which served as the go stimulus. The participants were asked to withhold their response to the No-Go stimulus (blue rectangle), which appeared in the center. The mean of errors (errors in Go trials, errors in No-Go trials, and error rate) were calculated based on the task performance. The Go and No-Go stimuli were presented in random (for details, see Fig. 1).
Procedure. There were two stages to this experiment. First, participants completed the questionnaires including the demographic questions, RPBS, and CFQ. Second, participants (N = 20) were organized for an in laboratory component. The experiment was carried out by participants at a computer station within an electrically isolated EEG-cabin. They were fitted with a 32-channel EEG (Mitsar Medical, Petersburg, Russia) headset after initially providing written informed permission. Although there seems to be a difference between the eyes-open and eyes-close conditions 86 , there are still inconsistent results 87 . We used an eyes-open resting-state The procedure for go and (B) no-go trials was identical, apart from the instruction to press spacebar after the presentation of go stimuli and to withhold a response after no-go stimuli. After 500 ms of stimuli, a blank screen appeared for 500 ms. A fixation point was shown for 800 ms. (C) Illustrations of go-and no go stimuli. An inter-trial interval (ITI) of 700 ms was used. The Go/No-Go task was administered with Inquisit 5 (Millisecond Software, LLC). www.nature.com/scientificreports/ instead of an eye-closed resting-state, because the EEG recording was conducted in the early morning for all the participants. Participants were instructed to keep their eyes open while the EEG was being recorded to avoid drowsiness artifacts in line with previous studies [88][89][90][91][92][93] . Additionally, the eyes-open resting-state represents a more appropriate baseline for comparing two groups together 90,92 . EEG data were collected while participants stared at a blank screen for five minutes. Then, participants completed Go/No-Go task. Eventually, participants completed the manipulation and compliance checks. After receiving a debriefing, the headset was taken off, participants had their hair cleaned, and they were thanked for their participation.
EEG recording, preprocessing, and analyses. The  Statistical analysis. Comparisons of demographic (age, gender, and education) and behavioral variables (PRBS scores, CFQ scores, error rate, error in Go and error in No-Go trials in Go-No/Go task) between the groups (i.e., paranormal believers and skeptics) were conducted using the independent two-sample t-test and chi-square tests. EEG absolute power data were analyzed separately for all the bands, the areas, the hemispheres, and the hemispheres/regions using an independent two-sample t-test between two groups (paranormal believers vs. skeptics) with adjusted p-values using the false discovery rate (FDR) method to limit type I error 98 . FDR correction was not performed in correlation results due to the exploratory nature of this stage. The effect size for the t-test was calculated using Cohen's d. In addition, we used Pearson's correlations to explore the relationships between resting-state EEG activities and demographic/behavioral variables in paranormal believers and skeptics. A multiple regression analysis was performed to show the effect of frequency bands (delta, theta, alpha1, alpha2, beta1, beta2, and gamma) as predictors on the measure of the paranormal belief as the criterion.
In addition, we used model 4 of the PROCESS macro in SPSS 99 to examine whether the effect of paranormal beliefs on inhibitory control (error in No-Go trials in Go/No-Go task) was mediated by frequency bands (delta, theta, alpha, alpha1, alpha2, beta, beta1, beta2, and gamma) in the whole brain (the first mediation model) and frontal lobe (the second mediation model) separately. The statistical significance of the indirect mediation effect on inhibitory control was assessed by bootstrapping (5000 samples) with a 95% confidence interval 99 . Statistical analyses were performed using IBM SPSS Statistics version 24 (IBM Inc., New York, USA), and MATLAB 2021a (MathWorks, Natick, Massachusetts), and p-values less than 0.05 were considered statistically significant.

Results
Behavioral performance in Go/No-Go task and demographic information. As can be seen in Table 1, No significant differences in age, sex, and education were observed between groups. However, paranormal believers had higher scores on the RPBS than skeptics. In addition, the CFQ result showed that paranormal believers reported significantly more general cognitive failures. The paranormal believers showed more errors in No-Go trials (p = 0.04) and errors in Go trials (p = 0.46) in the Go/No-Go task compared with the skeptics (Fig. 2).
EEG oscillatory activity. Figure 3 shows the two groups in terms of the absolute power in each band.
In the whole brain, the independent t-test results showed significant differences between the groups for the absolute power in the beta1 (t 18 = 2.117, p = 0.048, FDR corrected, Cohen's d = 0.997) and beta2 band activities (t 18 = 2.075, p = 0.050, FDR corrected, Cohen's d = 0.978). As shown in Fig. 3, the absolute power in the beta1 and beta2 band activities of the paranormal believers was lower than that of the skeptics in the whole brain. In the frontal area, the independent t-test results showed a significant difference between the groups for the absolute power in the alpha (t 18  www.nature.com/scientificreports/ ers was lower than that of the skeptics in the frontal area. In the right frontal area, the independent t-test results showed a significant difference between the groups for the absolute power in the beta1 (t 18 = 2.548, p = 0.020, FDR corrected, Cohen's d = 1.201) and beta2 band activities (t 18 = 2.680, p = 0.015, FDR corrected, Cohen's d = 1.263). The absolute power in the beta1 and beta2 band activities of the paranormal believers was lower than that of the skeptics in the right frontal area.
In the parietal area, the independent t-test results showed a significant difference between the groups for the absolute power in the beta1 (t 18 = 2.216, p = 0.040, FDR corrected, Cohen's d = 1.044), and beta2 band activities (t 18 = 2.133, p = 0.047, FDR corrected, Cohen's d = 1.005). The absolute power in the beta1 and beta2 band activities of the paranormal believers was lower than that of the skeptics in the parietal area. In the left parietal area, the independent t-test results showed a significant difference between the groups for the absolute power in the beta1 band activity (t 18 = 2.335, p = 0.032, FDR corrected, Cohen's d = 1.100). The absolute power in the beta1 band activity of the paranormal believers was lower than that of the skeptics in the left parietal area. In the occipital area, the independent t-test results showed a significant difference between the groups for the absolute power in the beta1 (t 18 = 2.494, p = 0.023, FDR corrected, Cohen's d = 1.175), and beta2 band activities (t 18 = 2.086, p = 0.050, FDR corrected, Cohen's d = 0.983). The absolute power in the beta 1 and beta2 band activities of the paranormal believers was lower than that of the skeptics in the occipital area.
In the right hemisphere, the independent t-test results showed a significant difference between the groups for the absolute power in the beta1 band activity (t 18 = 2.493, p = 0.023, FDR corrected, Cohen's d = 1.175). The absolute power in the beta1 band activity of the paranormal believers was lower than that of the skeptics in the right hemisphere. However, results showed no significant difference between the groups for the absolute power in other bands, areas, and hemispheres.
In the left temporal area, there were significant negative correlations between the severity of paranormal beliefs and absolute beta2 band activity (r = − 0.471, p = 0.036). In addition, in the occipital area, there were significant negative correlations between the severity of paranormal beliefs and beta1 (r = − 0.507, p = 0.032) and beta2 band activities (r = − 0.471, p = 0.036). Additionally, in the right hemisphere, we found significant negative correlations between the severity of paranormal beliefs and absolute beta1 (r = − 0.516, p = 0.020), and beta2 band  Additionally, the results of the multiple regression analysis show that the regression coefficient value (R) is 0.70 and the determination coefficient is (R2) 0.49. All frequency bands combined could explain 49% of the variance in paranormal beliefs. The amount of relative and effective contribution of the frequency bands (predictors), toward the dependent variable, namely the paranormal beliefs results, is presented in Table 2.

Correlation between inhibitory control and EEG oscillatory activity.
We also observed relationships between EEG frequency bands and the Go/No-Go task performance (Fig. 7, Supplementary Table 13). Error in No-Go trials in the Go/No-Go task was significantly negatively correlated with absolute power in the absolute beta1 power in the whole brain (r = − 0.494, p = 0.027), frontal area (r = − 0.473, p = 0.035), parietal area (r = − 0.490, p = 0.028), occipital area (r = − 0.498, p = 0.025), and right hemisphere (r = − 0. 512, p = 0.021). In addition, there was no significant correlation between Error in Go trials and absolute power in frequency bands.

Model.
Besides, a mediation model with inhibitory control (error in No-Go trials in Go/No-Go task) as a dependent variable, paranormal beliefs as an independent variable, frequency bands (delta, theta, alpha, alpha1,     1. showed that the effect of paranormal beliefs on inhibitory control was not mediated through frequency bands (delta, theta, alpha, alpha1, alpha2, beta, beta1, beta2, and gamma). Additionally, in the second mediation model in frontal lobe, the present study assessed the mediating role of beta2 band oscillation on the relationship between paranormal beliefs and inhibitory control. The results revealed a significant indirect effect of impact of paranormal beliefs on Inhibitory control. Furthermore, the direct effect of paranormal beliefs on Inhibitory control in the presence of the beta2 band oscillation in frontal lobe (mediator) was not significant. Mediation schematic summary is shown in Fig. 8. In addition, the present study assessed the mediating role of other frequency bands (delta, theta, alpha, alpha1, alpha2, beta, beta1, and gamma) in the frontal lobe on the relationship between paranormal beliefs and inhibitory control. In another frequency bands, the results didn't show a significant direct and indirect effect of impact of paranormal beliefs on Inhibitory control.

Discussion
Our main results indicate that paranormal believers show reduced beta power compared to skeptics. Additionally, paranormal believers were more impaired in inhibitory control on the No-Go trials, as compared to the skeptics. As expected, lower absolute beta1 and beta2 power activities during resting-state EEG were shown in paranormal believers. Previous studies have shown that a weaker beta band activity in No-Go trials is associated with poor inhibitory control 42,[45][46][47]58 . Beta band oscillations within the cognitive function occipito-frontal network is modulated by inhibitory control 100 . The top-down control of behavior has been used to illustrate the connection between cognitive processes and frontal beta oscillations 101 . Compared to instructed decisions, a study found that free decisions were accompanied by higher frontoparietal beta-band coherence 102 . Similar to the search and repetition phases of this task, free and instructed decisions demand differing levels of inhibitory control 44 . The "status quo" maintained by this proposed explanation is consistent with an extended theory of beta oscillations 43 . These authors suggest that when a task has a "strong endogenous top-down component," such activity should be connected to the active maintenance of a cognitive set in the cognitive domain. Furthermore, it's important to note that, like the majority of neurophysiological research on inhibitory control, our evidence for the relationship between beta power and inhibitory control is correlational. The causal mechanisms underlying these associations can only be shown by direct experimental changes in brain activity that concentrate on beta oscillations. From the view of neurochemistry, dopamine (DA), which is essential for stabilizing brain networks 103 , is most likely a key player in schizophrenia psychopathology 104 . Furthermore, the DA levels in neural networks could affect the temporal dynamics of beta oscillations 105 . For instance, the subthalamic nucleus's beta-band increased when Parkinson's patients were given the DA precursor Levodopa 106 . However, it is still speculative at this date to attribute the beta-band impairments in schizophrenia patients to the dopaminergic system. Future studies using combined EEG and positron emission tomography may look into the relationship between betaband differences in schizophrenia patients and DA. Additionally, concerning the reduction of the beta power in paranormal believers compared to skeptics in the present study, we can refer to the relationship between paranormal experiences and the DA 107 , as well as lower activity in the rIFG area 63 . Moreover, beta band differences during EEG resting-state in the lack of stimulant may indicate dopamine depletion in the rIFG area 108 , which is related to inhibitory control. Additionally, a study found that less constrained firing rates were caused by enhanced dopaminergic transmission in mesolimbic neurons 109 , and would generate more internal noise and promote, on the phenomenal level, loosened associations, and paranormal beliefs 110 .
Our results showed that in terms of absolute gamma power, paranormal believers showed lower activation in the frontal lobe compared to skeptics. According to the theory of cognitive deficits in paranormal beliefs 3 , it is possible to assume that this cognitive deficit may be associated with differences in brain activity. Several studies showed a relationship between paranormal belief and schizotypy, and other characteristics associated with schizophrenia 1,111,112 . Therefore, studies focusing on paranormal phenomena have reported a reduction in gamma band in individuals with high transliminality 68 , hallucinations, positive schizophrenia symptoms 113 , and dysfunction in attention and cognition 114 . Gamma band activity has been related to a broad range of cognitive processes [52][53][54][55][115][116][117][118] . Therefore, widespread gamma power in cognitive functions, such as inhibitory control 39 , may be linked to the feature binding theory in the brain 119,120 . Given that the gamma band activity has decreased www.nature.com/scientificreports/ in paranormal believers in the frontal lobe compared to skeptics, this can also be attributed to the poor inhibitory control in these individuals. From the cognitive neuroscience perspective, the decrease in gamma band activity may be related to neural dysfunction in the frontal lobe 121 , reflecting impairment in the synchronous inhibition  www.nature.com/scientificreports/ of pyramidal neurons 122 , and the GABAergic neurotransmission 123 . Furthermore, previous studies showed that the frontal lobe is involved in higher-order cognitive functions [124][125][126] . Concerning the frontal lobe dysfunction in paranormal beliefs 74 , the results of the present study also show that in paranormal believers as compared to skeptics, the absolute power of the beta was significantly reduced in the frontal area. These findings indicate that low gamma activity in the frontal lobe and low beta band activity in the frontal lobe may also be associated with the inhibitory control deficit in paranormal believers. We also found that paranormal believers showed a low absolute alpha power in the bilateral frontal lobe compared to skeptics. Alpha power reflected a neural inhibitory mechanism involved in the influence of external sensory information at the resting-state EEG 127 . Through alpha power, this mechanism may be related to the GABAergic inhibitory activity via an inter-neuronal network in the brain 128 , and an unbalance between the GABAergic system and glutamatergic neurotransmission. Alpha power has been associated with the integrity of the cognitive process 129 , while it has been found to play a critical role in inhibitory control 130 . The alpha band differences may be associated with a range of information processing deficits such as inhibitory control 26,131 .
Our results indicate that paranormal believers showed more errors in both the Go and No-Go trials compared to skeptics. Still, there was only a significant group difference in No/Go trials in the Go/No-Go task (Fig. 2). It seems that association of paranormal belief and Go/No-Go task performance is specific for inhibitory control, but that future studies could further investigate general associations of paranormal belief with task performance. Our results also showed that error in No-Go trials in the Go/No-Go task was significantly negatively correlated with beta power in the whole brain. Studies have shown reduced beta power in No-Go trials in different populations 42,[45][46][47][48][49] . Perhaps, deficiency in inhibitory control is associated with beta band reduction. Additionally, our results demonstrated that all frequency bands combined could explain 49% of the variance in paranormal beliefs, although the model was not significant. This may be due to the small sample size in the present study. However, there was no previous study on the relationship between paranormal beliefs and EEG oscillatory activity, these findings seem to prompt further investigations into correlations with EEG oscillatory activity in large sample size. Additionally, in mediation model analysis, beta2 frequency band activation in the frontal lobe (but not in the whole brain) mediated the relationship between paranormal belief and inhibitory control. However, there was no significant direct effect of paranormal beliefs on inhibitory control. The lack of significant correlations between paranormal beliefs and inhibitory control through frequency bands could in part be due to the small sample size. It may be revealed with larger populations and further studies are required with a large sample size.
We had several limitations in this study. Firstly, convenient sampling and selection of the sample from university students constituted one of the limitations of the present study. Based on these results, future studies could address associations of higher order frequency band oscillations with phenomena related to paranormal belief, such as schizophrenia. Secondly, the sample size in the present study was small. Thirdly, cognitive failures were assessed with a self-reported scale. Future studies using objective methods to assess cognitive failures level, including cognitive and behavioral function tasks, are needed. Another limitation of this study is that no resting EEG with closed eyes was measured. In addition, concerning the participation of women in this study, some studies have reported that performance in cognitive tasks related to the prefrontal cortex might be affected by the levels of estrogen and progesterone hormones 132,133 . Therefore, future studies are recommended to control the status of female participants. Finally, according to previous studies based on the greater tendency of women to paranormal beliefs 134,135 , and due to the small sample size of men and women in each group, it was impossible to examine their performance separately. Therefore, we suggest that gender differences be explored in future studies.
In summary, the present study shows that paranormal belief is related to the reduced power of the alpha, beta, and gamma frequency bands, and reduced inhibitory control. This study will provide new insights into the role of differences between believers and skeptics in brain activity in paranormal believers.

Data availability
The datasets used during the current study are available from the corresponding author on reasonable request.
Received: 16 February 2022; Accepted: 23 February 2023 Figure 8. The mediation model. The mediation model tests the relationship between paranormal beliefs as an independent variable, inhibition (error in No-Go trials in Go/No-Go task) as a dependent variable, frequency beta2 band in frontal lobe as mediator. FC frontal cortex, β beta band, / insignificant, *Significant in p < 0.005. www.nature.com/scientificreports/