DOI: https://doi.org/10.21203/rs.3.rs-2138858/v1
Consciousness is classically described as alertness and awareness. Damasio claims that this change is not enough for consciousness to emerge, that a basic element is missing, which is the "self". The purpose of this study was to investigate the role of electrical vestibular stimulation (VS) in the formation of bodily self-consciousness in healthy individuals by utilizing brain hemodynamic activations with the functional Near Infrared Spectroscopy (fNIRS) system.
The study protocol included three sessions: a session without VS, a VS session with application on the right temporo-parietal junction (rTPJ) and a VS session with application on the mastoid level with transcranial direct current stimulation (tDCS).
In VS provided at the rTPJ level, unlike the other sessions, increased hemodynamic activation was observed at primary somatosensory cortex-supramarginal gyrus in the left lobe and in the supramarginal gyrus in the right lobe. In VS provided on the mastoid level, increased hemodynamic activation was observed in secondary somatosensory cortex-ventral posterior cingulate cortex connection on left lobe, which we did not observe in other sessions.
After artificial VS, the perception of bodily ownership in individuals requires coding of personal space information as well as vestibular information processing and extrapersonal space integration information.
The concept of self-consciousness has received less attention and considered inapproachable by many scientists in neuroscientific studies examining the neurobiological mechanisms of consciousness. Scientists and philosophers such as William James and James Gibson have suggested that investigating the psychological, physiological, and neural mechanisms involved in bodily perception and bodily experience may be crucial for understanding self-consciousness (1–3). Bodily self-consciousness, which enables us to distinguish between parts of our own physical self and parts of the outside world; is essential for survival and constitutes a fundamental aspect of human self-consciousness. This sense arises from the integration of visual, tactile, vestibular, proprioceptive, and other bodily signals into a coherent multisensory information about one's own body. Importantly, this integration not only contributes to the awareness of body ownership, but also plays an important role in the perception of one's own localization.
While body ownership is defined as the immediate and continuous experience of our body and its parts belonging to us, self-localization is defined as the experience where the self is localized at a particular position of the body in space. The issue of one's own localization and own body ownership is closely related to how people encode spatial information. Under normal circumstances, self-localization and body ownership are closely linked to the body-centered frame of reference, but this relationship can lead to abnormal self-conscious states specifically in patients with neurological disorders and experimentally in healthy subjects. It has been argued that research on body part illusions fails to provide insight into the mechanisms of global aspects of the bodily self, such as self-identification with the body as a whole, self-positioning in space, and perspective. Thus, two main versions of multidimensional illusions targeting more holistic aspects of the self were used previously (4). First, the full-body illusion, where the person sees their own body or a fake object from behind in front of them from a third-person perspective (5), and second, the body-swap illusion, where a fake body is seen from a first-person perspective (6). In both versions of these illusions, simultaneous visio-tactile touch to the fake and real body increases body ownership of a virtual body (i.e. full-body ownership) compared to asynchronous touch (6, 7). Importantly, it has been suggested that only the third point of view produces a change in self-localization (5, 8, 9). While self-identification with a false body seen from a first-person perspective is associated with activity in premotor regions, changes in self-localization and visio-spatial perspective have been suggested to be associated with activity at the temporo-parietal junction (10, 11). Similar to body illusions, a participant immersed in a virtual environment receives conflicting multisensory information about own position (12). Here, the visual information indicates that the person is in a virtual world, while proprioceptive information exists in the real world, implying that there is a different body location between the physical body and the virtual body.
Autoscopic phenomena are illusory dissociations of one's own body associated with striking abnormalities in body ownership. Although various Autoscopic phenomena have been described, the main ones are “Autoscopic hallucination”, “heautoscopy” and “out-of-body experience” (13–17). These phenomena have been reported to be observed after damage to the temporo-parietal cortex or temporo-occipital cortex. They arise from different multisensory dissociations of knowledge from the bodily and surrounding area, leading to phenomenological differences in the individuals, including self-attribution, self, or self-position (13, 15, 16, 18, 19). The room tilt illusion is a sudden and temporary tilt of one's own body in the visual environment without misplacement (13, 20, 21). Typically, subjects report a sudden upturn (180◦ reversal of the visual field) with respect to their body, or a 90◦ tilt in the outer field. Although this illusion is associated with brain stem and vestibulocerebellar system lesions (21–26), parieto-occipital and frontal cortex lesions (20), peripheral vestibular disorders, it has also been described in healthy individuals without such disorders (27, 28). Since visual, extra-personal, and otolithic vestibular cues cannot be integrated during the tilted room illusion, it has been assumed that this illusion is related to the interaction with the parieto-insular vestibular cortex (PIVC), where the gravitational pathways are terminated (29).
The temporo-parietal junction is a region located close to the PIVC, suggesting that the vestibular cortex may play a role in self-positioning and visuospatial perspective. Studies in neurological patients showing that there is a frequent relationship between vestibular illusions and out-of-body experiences have stated that this is due to the regulation of vestibular signals with signals from the personal and impersonal fields (30–33), and different studies have suggested that this dissociation is mostly the result of abnormal neural activity at the temporo-parietal junction (10, 19, 24, 32). Changes in self-localization and perspective have been associated with perceived aspects of the individual visual field. In other words, it is suggested that various senses have different contributions to the construction of the bodily self in a visual task, depending on which of them is weighted when compared to visual information and vestibular information (20).
The prepose of this study was to investigate the healthy participants' brain hemodynamic responses to autoscopic phenomena and room tilt illusion enabled by a feasible virtual reality (VR) environment. Unlike previous bodily self-consciousness studies with the combination of VR and functional Near Infrared Spectroscopy (fNIRS), we aimed to compare responses to illusions using artificial vestibular stimulation (VS) on the right temporo-parietal junction (rTPJ) level and mastoid levels in our study. Therefore, the purpose of this study was to investigate the role of VS in the formation of bodily self-consciousness in healthy individuals by utilizing brain hemodynamic activations with the fNIRS system.
This study was carried out in the Istanbul Medipol University. Ethical approval was obtained from non-invasive research ethics committee of Istanbul Medipol University (No. E-10840098-772.02-2489). Oral and written informed consent were obtained from all the participants and details of the study protocol was explained to all the participants before the study. Consent form was prepared for Figure 3. Only right-handed healthy participants between the ages 18-30 were included in the study. Individuals with a diagnosis of epilepsy, having a pacemaker or metal implants in the head, history of brain surgery, participants with a sensitive scalp, and those taking neuroepileptic/antiepileptic medication and antidepressant were excluded from the study due to the counterindications in tDCS application. Additionally, individuals with vertigo, head trauma, head surgery, neurological/psychiatric disease, physical disability, and vision problems were excluded from the study due to the contraindications in VR application.
Study sample size was calculated using the "G*power sample size calculator" (34). In the power analysis we conducted to determine the sample size, the number of samples was obtained as 13. However, considering the possibility of missing data that may occur as a result of fNIRS analysis and erroneous data that may occur due to noise, a total of 20 people, 10 women and 10 men, were included in our study.
All participants were assessed once with the Cybersickness Susceptibility Questionnaire before starting the experiment (35). In order to eliminate the effects of the electric current on the study results, the experimental design was panned as 3 sessions for three days, with at least 24 hours between each session. No VS was applied in the first session, with fNIRS montage participants were immersed in the VR environment by wearing VR headset. Inside VR environment, participants were asked to perform 4 specific tasks. VS was applied only in the second and third sessions. VS was applied on the right temporo-parietal junction in the second session and on the bilateral mastoid region in the third session, and the participant was asked to perform the 4 tasks in the immersive VR environment. fNIRS measurement was conducted in all sessions. In order to eliminate the effect of participants learning the tasks in the VR environment, all three session sequences were randomly applied for each participant (Fig. 1).
Complaints including symptoms similar to motion sickness have been reported in participants being immersed in VR environments, and emerged motion sickness was seen at a rate of approximately 30% (36). It has been reported that motion sickness and virtual reality sickness are diseases with largely overlapping symptoms, although their stimuli and trigger mechanisms are different. For this reason, the "Cybersickness Susceptibility Questionnaire" was applied to the participants to be included in the experiment, once before participating in the experiment, in order to determine the sensitivities that may occur in the VR environment. Participants who are very rare and rare risk of cybersickness symptoms were included the study.
HTC Vive Cosmos Elite Virtual Reality headset was used to provide a VR environment. There were four different tasks that the participants were asked to complete, and at the end of each task, there was a target that the participant were asked to reach (Fig. 2). Task 1 was designed to create an autoscopic hallucination state by creating a virtual body 1 meter in front of the participant, in which the participant were asked to attribute to their own body from a third-person perspective. Task 2, was aimed to create the heautoscopic hallucination with the target in the middle of the virtual body and the real body. In the tasks 3 and 4, in addition to the first and second tasks respectively, room tilt illusion was created inside the VR environment where the room was virtually tilted upside down 180 degrees. In all tasks, participants were asked to stop moving when they reached the target. The target was in a 45° angle on the right side of the virtual body. Female and male avatars were used for female and male participants.
The hemodynamic activities of the participants were recorded before and during all four tasks. In the initial of each task, the brain resting hemodynamic activities of the participants were recorded for 20 seconds starting from the moment they attributed the virtual body to study the brain hemodynamic response on bodily ownership. Next, participants were asked to complete the tasks and at the end of each task, where participants stopped at their target, 20-second brain resting hemodynamic activites were recorded from the moment they stopped.
Hemodynamic activity responses were recorded using NIRSport (NIRScout 8–16, NIRx Medizintechnik GmbH, Germany) together with VR headset (Fig. 3). To determine the anatomical locations of the NIRS channels, the “spatial registration of NIRS channel locations” function of the NIRSite 2020.7 software was used (Fig. 4). NIRS channels were created between the beam source and detector and placed on the scalp. Thus, the beam sent from the source optodes was expected to show the hemodynamic activity under these channels with the NIRS signals coming from the detector's optodes. To record extracerebral signals, 2 detector optodes were placed 3 cm from the source optodes (37, 38). The changes in concentration of oxyhemoglobin (ΔHbO2) was calculated by using the modified Beer-Lambert law (39). Rays at 2 different wavelengths (760 and 850 nm) were sent from the beam source optodes and the beams were detected by the beam detector optodes. The top position of the head was placed in the Cz region according to the 10–20 EEG system. Recordings were taken from 18 channels created with 8 source and 7 receiver optodes positioned on the temporal and parietal cortex according to the EEG 10–20 system.
Neurostim device was used to provide vestibular stimulation (NeurostimTherapeutic, Neurosoft/Neurostim, Russia). Stimulation on the right temporoparietal junction was applied with carbon electrodes at 0.08 mA per 1 cm2, and with 25 cm2 (5 cm x 5 cm) electrodes at 2 mA for 20 minutes. According to the 10–20 EEG system, the anode electrode was in the CP6 region and the cathode electrode was on the left wrist (40). We conducted the current with a battery-operated constant current stimulator and two surface sponge electrode pads moistened with saline solution (41). VS on the mastoid region was applied with 1 cm radius gel-filled electrodes at 0.64 mA for 20 minutes, with the anode electrode on the right mastoid and the cathode electrode on the left mastoid (42). The parameters used were adjusted to be within the limits of safety established from previous studies in humans (43–45). Because of the tDCS has prolonged effects of up to 90 minutes in the human cortex, depending on the stimulation duration and current density (46), each session of the study was planned to be completed within 40 minutes from the end of the electrical stimulation.
Data processing was performed using Homer2 software. The collected signals were first converted into the optical density using function “hmrIntensity2OD”. For the preprocessing purpose, “hmrMotionArtifact, hmrMotionCorrectWavelet, enStimRejection” functions were applied to remove motion artifact. In addition, the optical density was bandpass filtered from 0.00 Hz to 0.5 Hz by “hmrBandpassFilt” function. Next, the hemodynamic responses including concentration of HbO was derived from the optical density by “hmrOD2Conc” function with a differential path length factor of 6.0. The detrended concentration between the baseline value of 2 seconds before the avatar appears and the time period between 20 seconds after the avatar appears was evaluated using the hemodynamic response function (HRF). Also, the same values were valid when the participants reached the target.
“SPSS 24.0 (Statistical Package for Social Science) for Windows” program was used for statistical analysis. In descriptive statistics, mean, standard deviation and percentage values were presented. The normal distribution of data was measured with the Kolmogrov-Smirnov Test. Multivariate Analysis of Variance (MANOVA) was used for statistical analysis within sessions, Repeated Measures ANOVA was used for statistical analysis between sessions, and One-way analysis of variance (ANOVA) was used for statistical analysis of inter-session coordinate and time data. A probability value of p < 0.05 was considered significant.
Twenty participants completed the study protocol and final analysis was performed with 20 participants. Ages of participates were between 19 and 27, mean age was 23.60 ± 2.28 (M ± SD) years. As a result of the Cybersickness Susceptibility Questionnaire, 3 participants marked "very rare" risk and 17 participants marked "rare" risk.
n | % | ||
---|---|---|---|
Gender | Female | 10 | 50.00 |
Male | 10 | 50.00 | |
Total | 20 | 100.00 | |
Cybersickness Susceptibility Questionnaire Results | Very Rare | 3 | 15.00 |
Rare | 17 | 85.00 | |
Total | 20 | 100.00 |
Hemodynamic activation in the session without vestibular stimulation
Significant difference was observed in channels 5, 7, 11, 12, 13, 14 and 15 compared to other fNIRS channels (p < 0.05, Table 2). Significantly higher ΔHbO2 was obtained in channels 7, 12, 13, 14 and 15 in autoscopic hallucination (task 1), and in channels 5 and 11 in heautoscopic hallucination & room tilt illusion (task 4) (Fig. 5).
Hemodynamic activation in vestibular stimulation at rTPJ
Significant difference was observed in channels 3, 4, 5, 6, 7, 8, 9, 11, 12, 13, 14, 15 and 16 compared to other fNIRS channels (p < 0.05, Table 2). Significantly higher ΔHbO2 was obtained in channels 9 and 15 in autoscopic hallucination & room tilt illusion (task 3), and in channels 3, 4, 5, 6, 7, 8, 11, 12, 14 and 16 in heautoscopic hallucination & room tilt illusion (task 4) (Fig. 5).
Hemodynamic activation in vestibular stimulation at mastoid
Significant difference was observed from channels 4 and 15 compared to other fNIRS channels (p < 0.05, Table 2). Significantly higher ΔHbO2 was obtained in channels 4, 9, 10, 11 and 15 in autoscopic hallucination & room tilt illusion (task 3), and in channels 5, 6, 7, 8, 12 and 14 in heautoscopic hallucination & room tilt illusion (task 4) (Fig. 5).
Channels | Vestibular Stimulation Sessions | ||||||||
---|---|---|---|---|---|---|---|---|---|
Without Vestibular Stimulation | rTPJ Level | Mastoid Level | |||||||
F | Effect Size | p | F | Effect Size | p | F | Effect Size | p | |
CH3 | 1.367 | 0.084 | 0.258 | 4.731 | 0.240 | 0.002* | 2.339 | 0.115 | 0.075 |
CH4 | 2.284 | 0.132 | 0.060 | 5.303 | 0.261 | 0.001* | 3.532 | 0.164 | 0.017* |
CH5 | 3.851 | 0.204 | 0.006* | 8.779 | 0.369 | 0.000* | 7.079 | 0.282 | 0.000* |
CH6 | 1.910 | 0.113 | 0,141 | 8.009 | 0.348 | 0.000* | 5.315 | 0.228 | 0.001* |
CH7 | 4.338 | 0.224 | 0.005* | 9.288 | 0.382 | 0.000* | 9.428 | 0.344 | 0.000* |
CH8 | 1.345 | 0.082 | 0.237 | 6.118 | 0.290 | 0.001* | 3.702 | 0.171 | 0.015* |
CH9 | 0.965 | 0.060 | 0.419 | 3.806 | 0.202 | 0.011* | 3.592 | 0.166 | 0.016* |
CH10 | 1.873 | 0.111 | 0.174 | 2.398 | 0.138 | 0.105 | 3.276 | 0.154 | 0.035* |
CH11 | 3.564 | 0.192 | 0.024* | 9.096 | 0.377 | 0.000* | 4.719 | 0.208 | 0.003* |
CH12 | 5.170 | 0.256 | 0.001* | 11.053 | 0.424 | 0.000* | 8.495 | 0.321 | 0.000* |
CH13 | 4.904 | 0.246 | 0.001* | 6.883 | 0.315 | 0.000* | 5.828 | 0.245 | 0.001* |
CH14 | 2.827 | 0.159 | 0.035* | 8.365 | 0.358 | 0.000* | 5.744 | 0.242 | 0.001* |
CH15 | 2.700 | 0.153 | 0.038* | 7.324 | 0.328 | 0.000* | 3.255 | 0.153 | 0.023* |
CH16 | 2.486 | 0.142 | 0.059 | 4.644 | 0.236 | 0.002* | 2.007 | 0.100 | 0.130 |
(* : p < 0.05)
Considering the links between vestibular cortical processing and body ownership and bodily consciousness, it was thought that interference with peripheral vestibular signals might lead to consequences related to the neurobiology of bodily self-consciousness. Neuroimaging studies have reached conclusions showing that PIVC-associated brain networks are critical in the presence of VS. The aim of this study was to examine the contributions of artificial VSs on bodily self -consciousness along with their hemodynamic activities. For this purpose, after electrical stimulation of the participants' vestibular cortex at the level of the right temporoparietal junction and the level of the mastoid region, their performance in the VR environment was compared with fNIRS. According to the results of the study, we could verify our hypothesis that after artificial VS at the rTPJ and mastoid levels, the participants contributed to the formation of bodily self-consciousness in the VR environment.
Our results on brain hemodynamic activity showed cortical HbO activation in various cerebral regions during four different tasks within the VR environment. All of the tasks performed in the presence of without VS, VS on rTPJ level and VS on mastoid level sessions showed the strongest oxygenated brain hemodynamic activation in supramarginal gyrus-angular gyrus region connections on right and left lobes and secondary association sensorimotor cortex on left lobe. Here; it can be thought that the increase in the hemodynamic activity of these regions in the session without VS may be due to the encoding of spatial focus, visual-spatial processing and visual movement response information on the basis of tasks. As a result of VS on both rTPJ and the mastoid region, we observed hemodynamic activation in primary somatosensory cortex on right lobe and secondary sensorimotor cortex-ventral posterior cingulate cortex (PMC) junction. Electrophysiological studies in the field of body ownership have mainly focused on somatosensory processing and activity of different cortical areas using somatosensory evoked potentials, rTMS and EEG (47–49). However, few studies have investigated the effect of body ownership on sensorimotor integration and motor cortex excitability. Sensory conflict provided in a VR environment can alter sensorimotor integration, so that its corticomotor output can provide more insight into the neural basis of bodily ownership. Our results regarding strong activation in the supramarginal and angular gyrus in the inferior parietal lobe are in line with the findings of other studies that vestibular processing becomes stronger with performance of tasks (50–52).
After the VS provided at the rTPJ level, unlike the other sessions, increased hemodynamic activation was observed at primary somatosensory cortex-supramarginal gyrus in the left lobe and in the supramarginal gyrus in the right lobe. The neurocognitive body ownership model presented by Tsakiris proposed a neural network responsible for body ownership perception (53). According to this model, the conflict between somatosensory and visual inputs during bodily illusions interrupts multisensory integration and triggers the process of body ownership. The rTPJ, somatosensory cortex, posterior parietal cortex, and right posterior insula have been suggested as strategic areas of this process. It has also been suggested that the ventral premotor cortex contributes to a sense of body ownership and provides a link to the motor system (53–55). Consistent with these studies, we observed hemodynamic activation in the left part of the primary somatosensory cortex, which we did not encounter in other sessions, in addition to the supramarginal gyrus region with stimulation from the rTPJ level.
In the VS provided at the mastoid level, we observed an increase in hemodynamic activity in secondary somatosensory cortex - ventral posterior cingulate cortex (PMC) connection on left lobe, which we did not observe in other sessions. It has been suggested that the posterior medial cortex plays a role in generating subjectivity (56). In the literature, in a case of a patient with seizures originating from the posteromedial cortex and characterized by a sense of self, it has been reported that the patient has “depersonalisation”, that is, unreality, disconnection, or being an outside observer regarding the person's thoughts, feelings, sensations or body (10, 23). This report demonstrated the body's self-dissociation during seizure auras and after stimulation of the PMC. In our study, we achieved an increased hemodynamic response in the PMC region with VS on the mastoid region. Also, the fMRI study findings suggesting that the posterior insula region is activated only during galvanic stimulation applied at the mastoid level supports its vestibular function. This location is best associated with the PIVC region, where neurons in this region respond to the electrical stimulation of the peripheral vestibular nerve and vestibular, somatosensory, and optokinetic stimulation.
When we evaluate the hemodynamic activity results created by the sessions on different tasks of VR environment; unlike sessions with VS, in the session without VS, increased hemodynamic activation was observed in task of autoscopic hallucionation. Also, in both sessions with VS, significantly higher activation was observed in fNIRS channels in the case of autoscopic hallucination&room tilt illusion. It can be thought that the reason for this is that the autoscopic hallucination is observed only with a disintegration in the personal area and is not affected by a pathology in the processing of vestibular information. When there is no VS, because there is a conflict in the personal area, it can be thought that the stimuli from visual information in the VR environment may increase the hemodynamic activation in the relevant regions. However, due to modulated vestibular processing after both applied VSs, it requires the ability process vestibular information and encode extrapersonal space integration information available with the room tilt illusion association, in addition to the ability to encode personal space information in autoscopic hallucination. Therefore, more oxyhemoglobin concentration changes were observed in this case compared to the session without VS. Moreover, high hemodynamic activation was observed in all sessions in the heautoscopic illusion&room tilt illusion task. As this task requires more vestibular processing and personal and extrapersonal integration skills than any other task, we can claim that increased hemodynamic activity was observed in certain fNIRS channels in this task. It was hypothesized that the effect of the parieto insular vestibular cortex on the room tilt illusion differs from that in out-of-body experiences and heutoscopy because during the room tilt illusion, there is only abnormal processing of body position in the person's extrapersonal space, without pathologies of body and body ownership (22).
Although most studies have identified the contribution of visual signals in determining self-position, other studies have highlighted the role of vestibular signals in integrating bodily experience by combining signals from bodily, retinal, and geocentric spaces in multimodal spaces (22). Because of TPJ activity receives visual, tactile, proprioceptive, and vestibular signals about body orientation depending on the environment, it was considered essential for calculating self-positioning and visuospatial perspective (27, 57). The multisensory integration that takes place in this region seems necessary for self-positioning and consistent perspective experience of the form. Therefore, the experience of self-positioning appears to be a necessary condition for self-consciousness and also for awareness of external events (58).
With the sensory conflict of bodily self-consciousness created in the VR environment in healthy individuals, it is ensured that people encode information about spatial focus, visual-spatial processing and reaction to visual movement. While artificial VS applied at the mastoid region level together with the VR environment cause sensorimotor integration, applied at the rTPJ level also cause changes on somatosensory integration, the cortical output of these stimulations can provide more information about the neural basis of bodily ownership. After vestibular processing modulated with artificial VS, the perception of bodily ownership in individuals requires coding of personal space information as well as vestibular information processing and extrapersonal space integration information. It has been observed that the fNIRS neuroimaging technique, which we used to reach these conclusions, is a method that can be used to measure the activation of bodily self-consciousness in cortical areas in people with artificial VS.
Electroencephalography
Functional Near Infra-Red Spectroscopy
Oxy-Hemoglobin
Milliampere
Parieto-Insular Vestibular Cortex
Right Temporo-Parietal Junction
Virtual Reality
Vestibular Stimulation
Transcranial Direct Current Stimulation
Conflicts of Interest:
The authors declare that they have no conflict of interest.
Ethics approval and consent to participate
Ethical approval was obtained from non-invasive research ethics committee of Istanbul Medipol University (No. E-10840098-772.02-2489). I declare that all methods are carried out in accordance with the relevant guidelines and regulations.
Consent for publication
Consent form was prepared for Figure 3. The form will be sent to you during the publication phase based on the information specified in the "Consent for Publication".
Availability of data and materials
All data generated or analysed during this study are included in this published article.
Competing interests
The authors declare that they have no competing interests
Funding
This study was funded by The Scientific and Technological Research Institution of Turkey (TUBITAK) under the 1002 program (Project No: 221S984).
Authors' contributions
H.Y. contributed significantly to the design of the study, the interpretation of the data, the creation of the new software used in the study, the preparation of the draft of the study and the writing of the manuscript. F.H. contributed the interpretation of the data in the creation of new software used in the study and was a major contributor in writing the manuscript. L.H. contributed the design of the study, the interpretation of the data, and was a major contributor in writing the manuscript. All authors read and approved the final manuscript.
Acknowledgements
We would like to thank The Scientific and Technological Research Institution of Turkey (TUBITAK) for supporting this study.