Higher hypnotic suggestibility is associated with the lower EEG signal variability in theta, alpha, and beta frequency bands

Variation of information in the firing rate of neural population, as reflected in different frequency bands of electroencephalographic (EEG) time series, provides direct evidence for change in neural responses of the brain to hypnotic suggestibility. However, realization of an effective biomarker for spiking behaviour of neural population proves to be an elusive subject matter with its impact evident in highly contrasting results in the literature. In this article, we took an information-theoretic stance on analysis of the EEG time series of the brain activity during hypnotic suggestions, thereby capturing the variability in pattern of brain neural activity in terms of its information content. For this purpose, we utilized differential entropy (DE, i.e., the average information content in a continuous time series) of theta, alpha, and beta frequency bands of fourteen-channel EEG time series recordings that pertain to the brain neural responses of twelve carefully selected high and low hypnotically suggestible individuals. Our results show that the higher hypnotic suggestibility is associated with a significantly lower variability in information content of theta, alpha, and beta frequencies. Moreover, they indicate that such a lower variability is accompanied by a significantly higher functional connectivity (FC, a measure of spatiotemporal synchronization) in the parietal and the parieto-occipital regions in the case of theta and alpha frequency bands and a non-significantly lower FC in the central region’s beta frequency band. Our results contribute to the field in two ways. First, they identify the applicability of DE as a unifying measure to reproduce the similar observations that are separately reported through adaptation of different hypnotic biomarkers in the literature. Second, they extend these previous findings that were based on neutral hypnosis (i.e., a hypnotic procedure that involves no specific suggestions other than those for becoming hypnotized) to the case of hypnotic suggestions, thereby identifying their presence as a potential signature of hypnotic experience.

than right hemispheres, and that low susceptible subjects showed only a weak lateralized asymmetry." Reviewer's Comment: Line 97 Why discuss the methodology/results and conclusions of the experiment in the introduction? Should only be discussing the justification of the study. Some of this material would be better in the abstract.

Authors' Response:
We addressed this issue in two steps.
1. Abstract: We added the following information to the Abstract of the manuscript.
"For this purpose, we utilized differential entropy (DE, i.e., the average information content in a continuous time series) of theta, alpha, and beta frequency bands of fourteen-channel EEG time series recordings that pertain to the brain neural responses of fourteen carefully selected high and low hypnotically suggestible individuals." 2. Introduction: We modified the content of this part (Section Introduction, lines 108-117, in the current version of the manuscript) as follows: "Given these findings, we sought the utility of DE for quantification of the brain neural responses to hypnotic suggestions. Specifically, we utilized DE of the theta, alpha, and beta frequency bands of fourteen-channel EEG recordings of twelve carefully selected high and low hypnotically suggestible individuals. We found that the higher hypnotic suggestibility was associated with a significantly lower variability in information content of theta, alpha, and beta frequencies. We also observed that such a lower variability was accompanied by a significantly higher functional connectivity (FC, a measure of spatiotemporal synchronization) in the parietal and the parietooccipital regions in the case of theta and alpha frequency bands and a non-significantly lower FC in the central region's beta frequency band."

Material and Methods
Reviewer's Comment: Line 127. Would a broader community sample have been more useful than a possibly biased academic one?
Authors' Response: This is in fact an important observation by the reviewer. In the current version of the manuscript, we added a new Section (Limitations and Future Direction, lines 425-495, in the current version of the manuscript) in which we discussed some of the limitations and future direction of this research. With regards to the study sample, we included the following paragraph (lines 478-486) to this Section. "In spite of the fact that the original group of individuals who participated in our hypnosis experiment formed a moderately acceptable sample size (i.e., forty-six subjects), the final validation for their inclusion in LOW and HIGH groups based on Harvard test [63] resulted in a small sample. Furthermore, all of these individuals were university students/staff, some of whom had previous exposure to hypnosis experience. As a result, it is plausible to presume that our participants were able to more readily comprehend and follow our experimental procedure, thereby contributing to an above-average outcome that one might expect from a general population. Therefore, it is crucial to reevaluate these findings while considering a broader general population." Authors' Response: Prior to providing our response to reviewer's comment, the authors would like to clarify that the number of participants that were included in this study were 12 and not 14. The reviewer's observation on 14 participants is correct as we originally identified 14 individuals that were not among the "Mid suggestible group" (Section Hypnosis Test and Suggestibility Score, lines 146-148, in the current version of the manuscript). However, we balanced the number of participants in such a way that LOW and HIGH groups each included 6 participants. Out of these 12 participants, four participants (one female) had previously experienced hypnosis either in form of a stage show or a research experiment. We included this information in Section Statistical Analyses (lines 217-227, in the current version of the manuscript). It reads as follows. Authors' Response: EEG signals were recorded from 14 sites that covered the frontal, central, temporal, parietal and occipital areas. Electrodes were placed on an EEG cap (g.tec, g.GAMMAsys) according to 10-20 international system (F3, Fz, F4, T7, C3, Cz, C4, T8, P3, Pz, P4, O1, Oz, and O2) (Figure 1 (c)) and were selected to cover five main cortical regions (i.e., frontal, central, temporal, parietal and occipital) in both left and right hemispheres (red circles) and midline locations (green circles). We added this information in Section Data Acquisition, lines 171-176, in the current version of the manuscript, as follows.

in both left and right hemispheres (red circles) and midline locations (green circles)."
Reviewer's Comment: Fig 1: were the other electrodes recorded or were just 14 recorded? If so, why were these selected? The electrode map may be misleading if the other electrodes were not used.

Authors' Response:
Only the 14 electrodes that are highlighted in Figure 1 were used in this study. We chose these electrodes due to their relative alignment with the brain regions that were identified by the previous research for their significant involvement in hypnosis: the default mode network (DMN) [64] and fronto-parietal network [1,8,10]. From a broader perspective, these electrodes covered all the major lobes of the brain that are involved in action, emotion, language, cognitive control, and action [65]. We included this information in Section Data Acquisition, lines 176-182, in the current version of the manuscript, as follows.
"We chose these electrodes due to their relative alignment with the brain regions that the previous research identified their significant involvement in hypnosis: the default mode network (DMN) [64] and fronto-parietal network [1,8,10]. From a broader perspective, the channels that were included in our study covered all the major lobes of the brain that are involved in action, emotion, language, cognitive control, and action (see [65], Chapters 9 through 12 for a detailed treatment of the subject)." With regards to the electrodes that were used in our study, we first modified Figure 1 (c) caption to to better clarify the electrodes that we used in our study, as follows.
"Fourteen electrodes placed on the frontal, temporal, central, parietal, and occipital areas in both left and right hemispheres (red circles) and the midline locations (i.e. green circles) recorded EEG signals during the experiment." However, the authors would also like to state that they have no objection to change of this subplot, in case the reviewer finds it necessary.
Reviewer's Comment: Line 176. A faster sampling rate may have been less problematic (especially for measuring higher EEG frequencies).

Authors' Response:
In the present study, we were specifically interested in behavioural responses to hypnotic suggestions that were mainly ideo-motor suggestions, inducing movement and noise in EEG. We clarified this point in Section Data Preprocessing, lines 191-195, in the current version of the manuscript. It reads as follows.
"We excluded gamma band (30-60Hz) from our analysis because the Harvard hypnosis test mainly includes motor items that require movement as a behavioural response and therefore, artefacts from muscle activity during these suggestions could have contaminated high-frequency EEG signals." Therefore, we decided to exclude the gamma band from our analyses, considering its vulnerability to movement related artefacts. Subsequently, we opted for a lower sampling rate of 128 Hz for EEG recordings. This choice was in accord with our overview of the EEG-based hypnosis research that identified 128 Hz and 256 Hz as the most commonly used sampling rates (references [14,31,100] in the current version of the manuscript). However, we agree with the concern of the reviewer on highlighting this limitation along with the use of high density electrodes. Therefore, we added the following discussion to Section Limitations and Future Direction, lines 467-477, in the current version of the manuscript.
"In the present study, we were specifically interested in behavioural responses to hypnotic suggestions that were mainly ideo-motor suggestions, inducing movement and noise in EEG. Therefore, we decided to exclude the gamma band from our analyses, considering its vulnerability to movement related artefacts. Subsequently, we opted for a lower sampling rate of 128 Hz for EEG recordings. This choice was in accord with our overview of the EEG-based hypnosis research that identified 128 Hz and 256 Hz as the most commonly used sampling rates (references [14,31,100] in the current version of the manuscript). However, future research that is empowered with high density electrodes and that includes higher frequency bands can allow for more comprehensive realization of the depth and breadth of brain responses to hypnotic suggestions. Such setting can also provide better testbeds for critical examination of DE and other biomarkers for study of the hypnosis." extending their results during the neutral hypnosis to the case of hypnotic suggestions, thereby identifying such neural activations as potential signatures of hypnotic experience. However, our results fell short in comparative analysis of DE with these previous measures (e.g., COH). Therefore, future comparative analyses of DE and these other measures to clarify their respective dis/advantages will be necessary to thoroughly appreciate their proper domain of application."

Reviewer's Comment:
Requires some clarity about how the 14 states were used to produce the data for subsequent analyses (average of suggestion1-10?). This is mentioned in Discussion line 349. Should be addressed in Results in more detail and why?
Authors' Response: Each individual experienced 14 phases which included a baseline recording, an induction phase (2 initial phases), 10 suggestions (10 separate phases), an awakening from hypnosis, and a post-baseline (2 final phases). The 10 suggestions in the middle where segments of interests in our study. We did not average the 10 segments, but rather computed one DE value for each segment in each frequency band and each EEG location. Given 10 suggestions and that each of HIGH and LOW groups included 6 participants, we had 6 X 10 = 60 DEs, per frequency band and for each of HIGH and LOW groups (e.g., 60 DEs for alpha band at F3). In the case of FC, we used these 60 DE values, per frequency (i.e., 6 participants × 10 suggestions), per channel, to compute the pairwise correlations among the channels. We verified this information in Section Statistical Analysis, lines 228-237, in the current version of the manuscript. It reads as follows.
"Each individual experienced 5 main stages (Figure 1) that included 14 phases: a baseline recording phase and an induction (2 initial phases), 10 suggestions (10 separate phases), an awakening from hypnosis phase, and a post-baseline (2 final phases). The 10 suggestions in the middle where segments of interests in our study. We computed one DE value for each segment in each frequency band and each EEG location. Given 10 suggestions and that each of HIGH and LOW groups included 6 participants, we had 6 × 10 = 60 DEs, per frequency band and for each of HIGH and LOW groups (e.g.,60 DEs for alpha band at F3). In the case of FC, we used these 60 DE values, per frequency (i.e., 6 participants × 10 suggestions), per channel, to compute the pairwise correlations among the channels." We further modified the information associated with the FC analysis (Section Change in Functional Connectivity (FC), lines 250-263, in the current version of the manuscript) to more clearly explain how DEs were used during the FC analysis. The modified Section reads as follow.
"To determine any potential significant change in functional connectivity among EEG channels of HIGH versus LOW suggestible groups, we performed all-pair FC analysis. For this purpose, we combined DEs of all participants for a given channel at a given frequency band and computed the pairwise FC using Pearson correlation (i.e., every pair of channels). This resulted in 14 × 14 FC matrices, per frequency band, where 14 refers to the number of EEG channels. For each channel, we then computed the average Pearson correlations that it had with the remainder of the channels and only considered those channels whose averaged Pearson correlations were ≥ 0.70 (i.e., primarily strong and very strong correlations) in our analysis. For the selected channels, we also counted the number of channels that they were synchronized with (i.e., number of channels that they showed ≥ 0.70 correlation with). For both of these measures (i.e., averaged correlation and number of synchronized channels, per selected channel), we applied Kruskal-Wallis test to determine the effect of suggestibility on FC. We followed this test with post-hoc paired Wilcoxon rank-sum test."

Discussion
Reviewer's Comment: A well written discussion about the effects/implications for the findings, however there should also be a discussion about the potential limitations of the study.

Authors' Response:
We added a new Section (Limitations and Future Direction, lines 425-495) in which we discussed some of the limitations and future direction of our research. It highlights the following topics:

Comparison with other measures (lines 425-432, in the current version of the manuscript):
We discussed how the future can benefit from comparative analysis of DE and other measures that are used in the study of the neural correlates of hypnosis. Please refer to the authors' response to reviewer's comment "It would have been interesting to see a comparison with standard coherence..." for the content of the discussion that has been included in this paragraph.

The use of resting period EEG (lines 433-441, in the current version of the manuscript):
In this paragraph, we stated the reason why we excluded the resting state EEG recordings of the participants in our study. We further pinpointed how the inclusion of this recordings in analysis of the effect of hypnosis on the brain activity can benefit the future research. for the content of the discussion that has been included in this paragraph. Please refer to the authors' response to reviewer's comment "Line 192. Why were the rest times excluded? These could have been associated with a control baseline." for the content of the discussion that has been included in this paragraph.

The use of Mid-hypnotisable group in the future research (lines 442-449, in the current version of the manuscript):
In this paragraph, we underlined the exclusion of the larger portion of our sample that corresponded to the mid-hypnotisable group, given the main objective our study. We further underlined how the inclusion of this group can help determine the potential relationship between high and low suggestible individuals. Please refer to the authors' response to reviewer's comment "Line 146: Would it also have been on some research interest to also examine..." for the content of the discussion that has been included in this paragraph.

Use of DE, minimum requirements, limitations, and potential solutions (lines 450-466, in the current version of the manuscript):
In this paragraph, we summarized the our motivation for considering the entropy (in general) and highlighted the major and previous studies that brought the DE for EEG analyses to the spot light. We then briefly discussed the main assumption for the use of DE and finally highlighted how an alternative solution might be considered to lighten it. This paragraph reads as follows.
"Considering the crucial role of the cortical self-organized criticality [82][83][84] in maximizing its information capacity [85][86][87], entropy has been proven as a powerful tool for quantification of the variability in brain functioning [88] and cortical activity [89] in such broad area of research as information processing capacity of working memory (WM) [47] and the state of consciousness [91]. In this regards, although the use of DE in neuroimaging (e.g.,Tononi et al. [90] and ) and EEG studies (e.g.,Duan et al. [56], Zheng and Lu [57],and Shi et al. [58], Zheng et al. [94]), its application for modeling of the brain functioning requires further investigation. Specifically, parametric adaptation of DE for the analysis of EEG time series [56][57][58][59] assumes that the time series data under investigation is normally distributed. Although the applicability of such an assumption in neuroimaging studies has been investigated [60,94,95], similar theoretical studies to better position the use of DE in EEG-based brain research is currently (to the best of our knowledge) lacking. Such analyses can help determine the domain of applications in which DE may not be an adequate measure for modeling the EEG time series of the brain activity. Along the same direction, it is also interesting to further examine the utility of the non-parametric formulation of the differential entropy [97,98] for modeling of EEG time series of the brain signal variability [99]."

5.Sampling rate and number of electrodes in the present study (lines 467-477, in the current version of the manuscript):
In this paragraph, we briefly discussed why we chose the sampling rate adapted in this study. Furthermore, we underlined the necessity for research using dense electrode EEGs and higher sampling rate for collecting the brain responses to hypnosis suggestions in higher frequencies to more comprehensively and critically examine the utility of DE and other biomarkers in hypnosis studies. Please refer to the authors' response to reviewer's comment "Line 176. A faster sampling rate may have been ..." for the content of the discussion that has been included in this paragraph.

Sample size and demographic limitations (lines 478-486, in the current version of the manuscript):
This paragraph discussed the shortcomings imposed by the sample of participants that were included in our study. Please refer to the authors' response to reviewer's comment "Line 127. Would a broader community sample have been more useful than a possibly biased academic one?" for the content of the discussion that has been included in this paragraph.

Prospect of future utilization (lines 487-495, in the current version of the manuscript):
We closed this Section by pointing at one potential real-world application of the findings such the results that we presented in this manuscript.
"The present findings are not only of interest to the psychology and neuroscience community but also to the researchers in the field of AI and brain-computer interfaces [101,102]. For instance, the use of DE as a biomarker of hypnosis can be utilized in development of real-time EEG classifiers that detect their users' responses to hypnotic suggestions. This, in turn, can expedite the deployment of the automated hypnotherapeutic systems of the future for clinical treatment of mental and behavioural disorders at brain functional level [23][24][25][26]. Such adaptations, in turn, can take the field a step closer to personalized hypnosis interventions that are tailored around the individuals' suggestibility level." Reviewer's Comment: There should also be some discussion about how the study could be expanded in future.