1 Introduction

Viewing Earth from space yields a profound feeling of awe, with an accompanying strong sense of interconnectedness and environmental awareness, similar to self-transcendence. This complex event is called the ‘Overview Effect’, described as a perceptually and conceptually overwhelming experience (Yaden and Newberg 2022) that evokes powerful emotional reactions in the observer, triggering an admiration of the planet’s beauty, self-transcendence, and feelings of belonging (Stepanova et al. 2019a; White 2014). It can lead to changes in the mental models related to our planet and its inhabitants, towards an increase in environmental awareness and connectedness to humankind (Kanas 2020; Stepanova et al. 2019b; Voski 2020).

The vastness and danger of space, as well as the overwhelming beauty of Earth, are profound elicitors of the experience of awe, which is a vital aspect of the Overview Effect (Keltner and Haidt 2003; Krenzer et al. 2018; Yaden et al. 2019). The Overview Effect is considered to be long lasting, and more powerful than ‘regular’ instances of awe (Yaden et al. 2016), triggered by for example the sight of a valley viewed from a mountain top. Research focusing on awe therefore often uses the Overview Effect, specifically having participants view our planet from space, as an induction procedure (Gordon et al. 2017; Nelson-Coffey et al. 2019; Reinerman-Jones et al. 2013). The emotion of awe is characterized by a complex set of emotions that include wonder, reverence, and fear, and can thus be both positive and negative in valence (Guan et al. 2019b; Keltner and Haidt 2003). Experiencers might notice time slowing down, feel small and connected to other beings, perceive the vastness of the witnessed phenomenon, and physical sensations like the chills. Their eyes might be widening, and they feel a need to mentally accommodate the information to existing mental models of the world (Yaden et al. 2019). In general, triggers of awe include visual or conceptual vastness, for example, a view from great heights such as in the Yosemite National Park (Bai et al. 2017), natural disasters like tornados, or other phenomena so grand that the spectator feels small in comparison (Piff et al. 2015). This can induce fear, threat, vulnerability, and fragility (Bethelmy and Corraliza 2019) in the observer, but may also increases respect for nature, improves well-being, promotes social behavior, and a feeling of connectedness to others (Stellar et al. 2017).

From a long-term perspective, awe can elicit a fundamental change in existing mental models of the world. It requires an adaptation to the new perceptual or identity related information, also called the ‘need for accommodation.’ In effect, awe is a driver of a fundamental belief change, which is why it has been characterized as an epistemic emotion that is inherently linked to scientific inquiry and learning (Anderson et al. 2020; Gail Jones et al. 2022; McPhetres 2019; Urban 2022; van Limpt-Broers et al., 2020). According to Yaden et al. (2016), the lasting effects of the Overview Effect as reported by astronauts in the past, can best be understood in terms of awe accompanied by the process of self-transcendence.

The short-term and long-term effects caused by the Overview Effect have been investigated using interviews (Gallagher et al. 2014; Lauren Reinerman-Jones et al. 2013; White 2014) or the Positive Effects of Being in Space questionnaire (PEBS) (Ihle et al. 2006). In addition, implicit association tests were used to reveal an underlying connection to nature with the Overview Effect leading to environmental awareness (Stepanova et al. 2019b). These instruments are sometimes enriched with measures of awe, where most studies rely on longer questionnaires like the AWE-S scale (Yaden et al. 2019) and the Situational Awe Scale (Krenzer et al. 2018), or single-item Likert questions to measure awe after an experience (Chirico et al. 2020; Guan et al. 2019a; Hu et al. 2017; Piff et al. 2015), and questionnaires about disposition for awe (Guan et al. 2018; Nakayama et al. 2020; Shiota et al. 2006). The reduction in conceptual self-size, the ‘small-self’, has been investigated with the ego-dissolution questionnaire (Nour et al. 2016), and pictorial methods to measure self-size. For example, when participants are instructed to draw themselves after experiencing awe, the human-figure drawings tend to be smaller (Bai et al. 2017; Colantonio and Bonawitz 2018; Sawada and Nomura 2020; Sturm et al. 2022; Elk et al. 2016; van Limpt-Broers et al. 2024).

The drawback of self-reports is that they are based on the assumption that participants are able to identify and convey their own emotions, emotional states, and feelings. However, retrospective emotion reporting can be biased by outside factors such as the time that has passed since feeling the emotion, and the participant’s personality and beliefs (d'Mello and Graesser 2014; Feldman Barrett 1997; Rasinski et al. 2005). Participants may also find it difficult to verbalize their experience (Stepanova et al. 2019a). A combination of standardized questionnaires and neurophysiological measures through brain imaging would likely give a more reliable result in the study of the Overview Effect and awe.

Past studies demonstrate that awe can be measured through brain imaging such as functional magnetic resonance imaging (fMRI) (Takano and Nomura 2022; Elk et al. 2019), voxel-based morphometry (VBM) (Guan et al. 2018; 2019b) and electroencephalography (EEG) (Chirico et al. 2020; Gallagher et al. 2014; Hu et al. 2017; Reinerman-Jones et al. 2011; 2013). Based on these outcomes, it can be expected that the Overview Effect, of which awe appears to be an important emotional element, would be associated with changes in brain signals. This assumption has been partially confirmed by several studies conducted by Reinerman-Jones et al. (2011; 2013), and Gallagher et al. (2014). However, these were based on participants viewing still images rather than being engaged in a fully immersive, animated experience, raising questions about the comparative value of the measurement.

Compared to other brain imaging techniques, EEG has the advantage that it can easily be combined with a VR headset and provides optimal temporal resolution (Gevins et al. 1999). EEG is a method where electrodes are placed in a standard grid upon the scalp, and differences in electric potential caused by activated synapses in the brain can be detected (Michel and Brunet 2019; Yürdem et al. 2019). Raw EEG data must be pre-processed to filter out noise and artifacts such as muscle movements and eye-blinks (Delorme and Makeig 2004). Finally, the cleaned recorded electric potential can be divided into frequency bands per sensor. These frequency bands include delta (0.5–4 Hz), theta (4–8 Hz), alpha (8–3 Hz), beta (13–30 Hz) and gamma (30–100 Hz) frequencies (Tian et al. 2021; Valenzi et al. 2014; Yürdem et al. 2019). Variations in potential within frequency bands have previously been reported in relation to different kinds of cognitive processes, some of which are opposites within the same frequency bands. Activity has been defined by bursts, changes, and oscillations, which can be positive, or negative, and can explain the seemingly contradictory processes which have been observed. The lowest frequency band concerns delta waves, which are related to deep sleep, unconsciousness, and deep meditation (Liu and Sourina 2012; Ramírez-Moreno et al. 2021; Stinson and Arthur 2013; Yürdem et al. 2019). Changes in the theta signal have been observed during sleep or deep relaxation, but also during periods of (emotional) arousal and stress, as well as during concentration and sustained mental effort (Liu and Sourina 2012; Ramírez-Moreno et al. 2021; Stinson and Arthur 2013; Yürdem et al. 2019). Alpha rhythms have been associated with stages of relaxation where the potential is higher. Alpha power, however, decreases during perceptual and emotion processing and memory engagement, as well as during creative processes and conscious thoughts (Liu and Sourina 2012; Ramírez-Moreno et al. 2021; Stinson and Arthur 2013; Yürdem et al. 2019). The beta band has been linked to focus, concentration, arousal in emotions, active thinking, and awareness (Liu and Sourina 2012; Ramírez-Moreno et al. 2021; Stinson and Arthur 2013; Yürdem et al. 2019). Finally, activity within the gamma frequency band is associated with sensing, feature binding, attention, memory, bursts of insight, high-level cognitive processing, and matching perceived stimuli to remembered ones (Herrmann et al. 2004; Liu and Sourina 2012; Penolazzi et al. 2009; Ramírez-Moreno et al. 2021; Stinson and Arthur 2013; Yürdem et al. 2019).

Table 1 provides an overview of EEG studies that investigated the Overview Effect using still images (Gallagher et al. 2014; Reinerman-Jones et al. 2011; 2013) and shows that it appears to be associated with changes in theta, and beta band activity. When experiencing the Overview Effect, theta band activity decreases, potentially indicating engagement and alertness (Gallagher et al. 2014; Reinerman-Jones et al. 2011). Oscillatory activity changes in the beta band have also been related to engagement levels (Gallagher et al. 2014; Reinerman-Jones et al. 2011). In addition to decreased theta and variation in beta spectral power, awe has been reported to lead to increased functional connectivity in all areas of the brain for these two frequency bands, similar to what happens in meditative states (Chirico et al. 2020). Furthermore, when EEG is recorded while different emotions are induced, awe appears to be strongly correlated with other positive emotions over the beta and alpha frequency bands, which indicates that in this case the brain interprets awe as being positive in valence (Hu et al. 2017).

Table 1 Overview of EEG research on awe
Full size table

Despite the low spatial resolution of EEG, the location of electrical impulses on the scalp has provided some insights into the brain areas associated with the Overview Effect. These include activation in the right parietal region, associated with general behavioral arousal (Schmidt and Trainor 2001), and frontal regions, involved in emotion processing, with activity in the left region representing positive emotions, and in the right region representing negative emotions (Martin and Altarriba 2017; Schmidt and Trainor 2001). Awe has been linked to EEG-signals in the right and left hemispheres (Gallagher et al. 2014; Reinerman-Jones et al. 2011; 2013) and also correlated with other positive emotions in the central brain area (Hu et al. 2017).

Past EEG studies of the Overview Effect, and awe used images (Gallagher et al. 2014; Reinerman-Jones et al. 2011, 2013), or non-immersive 30-s video clips (Hu et al. 2017). While the use of images is a validated method in emotions research, it should be taken into account that one of the most prominent aspects of the Overview Effect, as well as many other awe-triggering experiences described in the literature, involves the perception of vastness which may be difficult to convey in still images. Still images presented through a VR headset are not comparable to real life (Higuera-Trujillo et al. 2017) because of their low immersion (Amin et al. 2016; Kober et al. 2012). VR simulations, on the other hand, provide a great alternative to real experiences for the purposes of emotion research (Chirico and Gaggioli 2019a; Clemente et al. 2014). Immersiveness in VR is achieved through the richness of sensory information mediated by technology (Amin et al. 2016; Baños et al. 2004). High immersion might lead to a feeling of presence, or the ‘sense of being there’. When a person is fully engaged and forgets about the real environment in favor of the virtual environment, they will experience a high sense of virtual presence. As a consequence, they are likely to respond to the virtual environment in the same way as they would respond to a real environment (Baños et al. 2004; Chirico and Gaggioli 2019a; Witmer and Singer 1998). What is therefore needed is a fully immersive simulation that provides a realistic Overview Effect experience (Kalantari et al. 2021).

In the current study, we utilized an immersive Overview Effect VR simulation and assessed its effectiveness with standard questionnaires. The events in the simulation that were predicted to elicit the Overview Effect were subsequently analyzed in relation to the EEG signal. The neurophysiological measurements were further enriched with self-reports collected after the simulated experience. Based on the findings of the studies listed in Table 1, we predicted that the spectral power within the theta frequency band would decrease, and oscillatory activity would change for the beta frequency band. For delta, alpha, and gamma frequencies, we expected no change, or could not make a prediction based on the literature. However, all frequencies were included in this research to obtain a complete picture of EEG changes related to the Overview Effect.

2 Method

2.1 Participants

Forty-two students from Tilburg University, the Netherlands, participated in this study. Inclusion criteria for participants involved normal or corrected to normal vision. The participants were healthy (i.e. no pacemaker and no known epilepsy conditions). Furthermore, all participants were Dutch-speaking because of the Dutch narrated simulation. One participant was removed from the analysis because of computer failure, leaving a sample size of 41 (Mage = 20.07 years; SDage = 1.94 years; 14 male, 27 female). Students participated in this study for course credit. The study was approved by the Ethical Review Board of Tilburg University (REDC # 2019/04b). A post-hoc power analysis was conducted using G*Power (Faul et al. 2007), which indicated that for 41 participants in Wilcoxon signed-rank tests for matched pairs with a 0.05 one-tailed Type I error probability, the power of the detected medium sized effect was 92%.

2.2 Materials

A fully immersive European Space Agency (ESA) astronaut endorsed VR simulation was used, which was originally developed for the educational program of the non-profit organization SpaceBuzz. This VR journey simulated a rocket launch into orbit around Earth in order to experience the Overview Effect. The launch included movement of a hydraulic chair in the beginning and end of the simulation and immersive visuals and audio to learn about Earth (Fig. 1).

Fig. 1
figure 1

Schematic of a participant within the rocket segment

Full size image

The SpaceBuzz VR simulation was created in Unity with a high resolution and contrast using an HTC Vive Pro VR headset (Resolution: 1440 × 1600 pixels per eye, 615 PPI, 3D Spatial Audio, refresh rate of 90 Hz). Weightlessness in space was simulated using a hydraulic chair where participants’ feet did not touch the floor and by the presence of visual stimuli such as floating objects in the captain’s cabin. An avatar of ESA astronaut André Kuipers narrated the experience in Dutch.

The VR simulation was segmented in 34 meaningful events (Table 2) for experimental purposes. These were given a sequence number (T1, T2, T3, etc.). The simulation started with the rocket launching into space during which a hydraulic chair vibrated to create an immersive experience of the launch, and orbiting Earth (T1, still image presented in Fig. 2a). Once in orbit, the user’s chair moved (moment of movement excluded from the data), and one of the doors opened to enlarge the viewing area. The participant was then ‘suspended in the vastness of space’ and Earth came into view for the first time (T5). This is the moment the Overview Effect is experienced: ‘a state of awe with self-transcendent qualities, precipitated by a particularly striking visual stimulus’ (Yaden et al. 2016), both perceptual and conceptual vastness. As Table 2 shows, this event matches all Overview Effect criteria as described by Stepanova et al. (2019a): 1) showing vastness, 2) high resolution and contrast, vividness, dynamic and changing visuals of Earth, 3) a strong focus on Earthgazing, right after the initial excitement from a launch to space, weightlessness is simulated, 4) with a personal connection that was created by the narrative. Following the experience of the Overview Effect, topics such as deforestation (example in Fig. 2c), excessive fishing, and pollution were presented. After a short detour to the moon to witness Earthrise (Fig. 2d), the rocket returned back to Earth, ending the VR journey. The total time of the VR simulation was 14 min and 25 s. The simulation was accompanied by music associated with awe (Neidlinger et al. 2017; Silvia et al. 2015), which should have no negative effect on the experienced emotions (Chirico and Gaggioli 2019b).

Table 2 Meaningful segments of the SpaceBuzz Simulation in chronological order
Full size table
Fig. 2
figure 2

Screenshots of the SpaceBuzz VR simulation. still images include a Cockpit Time (T1), b Overview Effect (T5), c Pop-up rainforest (T21), d Earthrise (T32) (cf. Table 2)

Full size image

2.3 Equipment

EEG-signals were collected using a wireless B-Alert X10 system (ABM) transmitterFootnote 1 and sensor strips for nine electrodes (Fz, F3, F4, Cz, C3, C4, POz, P3, and P4). The Acqknowledge 5.0 software package (BIOPAC Systems, Inc.) was used to record and save the data at 256 Hz.

2.4 Instrumentation

2.4.1 Disposition for awe.

The awe items from the Dispositional Positive Emotion Scales (DPES) were used to measure disposition for awe (Shiota et al. 2006), with a Cronbach’s α of 0.57. Considering that this measure only consists of 6 items, a low reliability is not unexpected. The DPES-Awe questions survey dispositional positive affect, where a higher score indicates that awe is experienced more frequently, see Appendix D.

2.4.2 Overview effect.

The Overview Effect was measured using the Positive Effects of Being in Space questionnaire (PEBS), including questions divided in sub-scales a) perception of Earth, b) space, and c) the effects the experience had on daily life (Ihle et al. 2006). As is shown in Appendix 1, the PEBS allow for questions being answered not having experienced being in space, and therefore will allow for being included in a pre-test (i.e. prior to the VR space flight). It had a Cronbach’s α of 0.62 for the pre-test (baseline) questionnaire, and of 0.69 for the post-questionnaire. Removing sub-scale c would significantly increase the α to 0.75 for the pre-, and 0.79 for the post-questionnaires. However, for completeness, we have decided to keep all three subscales in the analysis. The current reliability levels can be considered acceptable (Taber 2018; Tavakol and Dennick 2011).

2.4.3 Awe.

Awe was measured using a human-figure drawing method where participants were asked to draw themselves on a piece of paper, see Appendix 2. A smaller figure indicates smaller self-size relating to a greater experience of awe (Bai et al. 2017). To measure perceived self-size, we recorded the height of the depicted person from the longest leg to the top of the head in centimeters with one decimal. To account for size variation between drawings, we normalized the data using the following formula (1), based on the pre- and post-human-figure drawing height (Craig et al. 2004; Hake 2014; Meltzer 2002). Drawing gains reflects the relative normalized change from pre- to post-drawing, thereby allowing for a between-participants comparison. Subsequently, gains of 0 mean that the pre- and post- drawings are the same. Negative gains indicate a decrease in size, and positive gains indicate an increase. Hence, gains should be compared to “0”, during analysis.

$$Drawing\;gains = \left( {\frac{{\left( {\frac{{Postheight}}{{Preheight + Postheight}}} \right) - \left( {\frac{{Preheight}}{{Preheight + Postheight}}} \right)}}{{1 - \left( {\frac{{Preheight}}{{Preheight + Postheight}}} \right)}}} \right)$$
(1)

In addition to the pictorial method, the AWE-S scale was used to measure experienced awe (Yaden et al. 2019). This 30-item seven-point scale questionnaire can be divided into factors indicating what is felt during the experience of awe, namely a) ‘time slowing down’, b) ‘ego-dissolution’, c) ‘increased connectedness’, d) ‘perceived vastness’, e) ‘physiological changes’, and f) a ‘need for accommodation’, seeAppendix 3. It had a Cronbach’s α of 0.86 for the entire questionnaire.

2.4.4 Simulator sickness.

To ensure that none of the participants had experienced nausea (which can result in unreliable data), the Virtual Reality Simulator Sickness Questionnaire (Kim et al. 2018) was administered and analyzed. It is a nine-item four-point scale questionnaire that lists possible symptoms regarding simulator sickness, see Appendix 5. Participants scored low on this questionnaire (M = 1.76, SD = 0.43), indicating only slight discomfort. It had a Cronbach’s α of 0.71.

2.5 Procedure

Upon arrival in the lab, participants were welcomed, and signed an informed consent form. While EEG sensors were applied, they filled out a demographics questionnaireFootnote 2 the DPES items for dispositional awe, and the pre- or baseline drawing to measure awe, and the pre- or baseline PEBS questionnaire to measure the Overview Effect. The experimenter measured the participant’s head to fit the correct size sensor strip for EEG. The strip was placed on the participant’s head, and sensor-foams were filled with electrode skin preparation gel (ELPREP). After correct application of the headset, and impedance check of the sensors, a wig-cap was placed over the participant’s head to avoid movement of the sensor strip when the VR headset was applied. Following the completion of the questionnaires, and the calibration of the sensing equipment, the participants sat down in the hydraulic chair and the VR headset was placed on their head. The EEG recordings were time aligned with the VR simulation. The experiment concluded with post-experimental questionnaires, namely a repetition of the drawing to measure perceived self-size or awe, the post- PEBS questionnaire to measure Overview Effect, the Awe-S Scale, and a VR simulator sickness questionnaire. The sensing equipment was removed, and the participant was debriefed and thanked for participation. See Fig. 3 for the flow of the experiment.

Fig. 3
figure 3

Flow of the experiment (The knowledge questions about space, as well as a questionnaire about science motivation are included in this overview for completeness and transparency but are not further discussed in this paper.)

Full size image

2.6 EEG pre-processing

Pre-processing of the raw EEG data was performed using MATLAB R2021b (Mathworks Inc., 2021) and the EEGLAB toolbox (Delorme and Makeig 2004). First, a band-pass filter between 1 and 50 Hz was applied to the raw EEG data. Then, the EEG data was re-referenced to average reference. Lastly, independent component analysis (ICA) was conducted to find and manually reject ICA components that represented noise and artifacts (e.g., eye-blinks).

Pre-processed EEG data was segmented into epochs based on the length of phases (i.e., events) of the VR journey to ensure equal segments across participants (Table 2). This resulted in 34 epochs of the EEG data per participant with a length between 3 and 38 s (M = 17.94, SD = 7.35), excluding moments in the video where the chair was vibrating or moving because of possible interference with the EEG recordings. Spectral power values of delta (0 – 4 Hz), theta (4 – 8 Hz), alpha (8 – 13 Hz), beta (13 – 30 Hz) and gamma (30 – 100 Hz) frequency bands per electrode were extracted for each epoch. Baseline EEG-signals were extracted from the 3-s time windows before the onset of each epoch overlapping with previous epoch (Hu et al. 2017; Koelstra et al. 2012). Subsequently, spectral power values were baseline-corrected by subtracting the spectral power of corresponding baseline EEG. This a common method for baseline-correction and is used when no neutral baseline period has been recorded (Ali et al., 2015; Lai et al. 2018). It provides a form of normalization and noise reduction; when the data from the 3-s window prior to the epoch is removed, theoretically, only true neural activation remains. This solution allowed for comparison between participants and epochs. To reduce the number of statistical tests, mean spectral power values were then calculated for the following cortical areas: Frontal (F3, Fz, F4), Parietal (P3, Poz, P4), Central (Fz, Cz, Poz), Right (F4, C4, P4), and Left (F3, C3, P3) regions of the brain (Gallagher et al. 2014; Reinerman-Jones et al. 2011). This resulted into 850 spectral power features in total (5 frequency bands × 5 areas × 34 events). See Online Resource 1 for an overview of the data. These values were used for statistical analyses.

To explore whether there was a potential bias because of differences in epoch lengths, given that the durations of topics varied, the first part of the simulation was segmented into 4-s epochs. These 4-s epochs were then baseline corrected as well. When all epochs are set the same length, such a length bias can be avoided. See Fig. 4 for the selection of 4 s epochs on the simulation timeline.

Fig. 4
figure 4

Division of 4 s epochs for the analysis on the simulation timeline

Full size image

2.7 Statistical analysis

As this study has a pre-post design, the outcome variables include pre- to post- differences in the PEBS questionnaire, showing the extent to which participants experienced the Overview Effect, and the pre- to post- drawings, revealing awe through the small-self principle. Disposition for awe was measured prior to the experiment with the DPES-awe questionnaire, and situational awe was measured post-VR using the AWE-S Scale. EEG was measured only for the duration of the VR simulation and analyzed based on VR content. See Table 3 for an overview of the statistical analyses.

Table 3 Overview of statistical analyses
Full size table

Wilcoxon Signed rank tests were used to analyze a difference between pre- and post- questionnaires and drawing gains, because of non-parametric data distribution tested using Shapiro–Wilk tests. EEG data were analyzed using a Wilcoxon Signed rank test as well. For non-parametric data distributions, and tests, reporting the median and interquartile range is more informative than means and standard deviations, and will therefore be reported in the current study (Candia-Rivera and Valenza 2022; Habibzadeh 2017). Bonferroni correction on repeated tests for five regions resulted in an α of 0.01. Spectral power values from the Overview Effect (T5) were compared to the general median of all other events to establish whether it stands out compared to the other events in the simulation.

The problem with a comparison between the Overview Effect (T5) and all other events (T1-T34), is that a rather brief event is compared with all other events of different and longer durations, which might distort the results by possibly yielding a Type I error. Therefore, the Overview Effect event (T5) was also compared with three other individual events of interest that are more similar in length: 1) the first event after launch (T1, Cockpit time) where participants just arrived in space and had not yet been exposed to a view of space; 2) the second selected event concerns the moment just prior to the Overview Effect event (T4, Space view oxygen), where the participant has been exposed to a view of space but not of Earth; 3) the third selected event is the event following the Overview Effect (T6, Beautiful view), in which the visuals are the same as that of the Overview Effect, but a spoken voice makes the event less transcendent.

Other events that were included in the analysis that are visually comparable to T5 (Overview Effect) but do not match the Overview Effect criteria, are T7 (Earth in view) and T25 (Silent over Earth).

Finally, two events were included that are considered similarly transcending to the first time that Earth is viewed from space are compared to the Overview Effect (T5), the moment of Earthrise (T31 (Earthrise start), and T33 (Earthrise continued)). Contrary to the other events (T4, T6, T7, T25) presumed to yield activation different than that following the Overview Effect (T5), we expected event T31 and T33 not to differ with findings from the Overview Effect (T5).

In addition to the main analyses that involve epochs based on topic (i.e., events), an analysis was performed comparing spectral power values from the Overview Effect to 4 s epochs prior to this moment.

3 Results

3.1 Questionnaires

Participants reported a disposition for awe above the mid-point (DPES Awe Mdn = 5.00), a score higher than that reported in other DPES awe questionnaire studies (Anderson et al. 2020; Guan et al. 2019a; Nakayama et al. 2020). The PEBS Overview Effect questionnaire increased from baseline (Mdn = 5.15) to post-questionnaire (Mdn = 5.54), using a Wilcoxon Signed rank test, Z = -5.31, p < 0.001, demonstrating participants experienced the Overview Effect. Awe was confirmed by both the questionnaire, and pictorial method. The AWE-S score (Mdn = 4.73) was similar to awe-experiences in other studies, and higher than neutral conditions (Graziosi and Yaden 2019; Yaden et al. 2018). Drawing gains were calculated from the pre-drawing (Mdn = 5.50) to the post-drawing (Mdn = 5.00). Gains were significantly different from zero, using a Wilcoxon Signed rank test, Z =  −22, p = 0.03, in the expected direction, where a smaller figure indicates a smaller self. These results show that participants experienced awe by drawing smaller figures without being asked directly about their feeling of small-self after viewing the simulation, it is thus an implicit measure of awe. The Overview Effect post-score was positively correlated with the DPES Awe, r(41) = 0.33, p = 0.04, and the AWE-S score, r(41) = 0.49, p < 0.001. In short, the self-report data (i.e., questionnaires and drawings) showed support for the experience of awe and the Overview Effect in participants. For an overview of the questionnaire results, see Table 4.

Table 4 Overview of experimental results, Z from wilcoxon signed rank test comparing pre to post
Full size table

3.2 EEG

The spectral power values from the Overview Effect (T5) were compared to the general median across all events barring the Overview Effect. Results show a significant difference between this event and the general median for the frontal, central, and right hemisphere beta power, and for the frontal, parietal, central and left hemisphere gamma power (Table 5).

Table 5 Related samples wilcoxon signed rank Z-values for comparisons to the overview effect
Full size table

For the events T1 and T4 prior to the Overview Effect (T5), decreased spectral power for the Overview Effect did not reach significance (Table 5). For event T6 following the Overview Effect (T5), significant differences in beta and gamma frequency bands were found, similar to the comparison of this event with the general median, see Table 5.

Confirming our expectations regarding moments T7 and T25, events visually comparable to the Overview Effect but not matching the Overview Effect criteria, a significantly lower relative spectral power was found for the Overview Effect event in most brain areas for the beta and gamma frequency bands in comparison to T7 and T25 (Table 5). Finally, for events T31 and T33, that are considered equally self-transcending as the Overview Effect (T5), no significant differences between these two experiences were found, see Table 5. Baseline-corrected median spectral power was plotted over time per brain region with all comparisons highlighted for beta and gamma bands, see Fig. 5.

Fig. 5
figure 5

Median values (N = 41) for beta and gamma frequency bands over the duration of the VR simulation

Full size image

Results of the 4-s epoch analysis revealed a significant difference between the moment prior to the Overview Effect and the Overview Effect for right hemisphere theta, Z = -2.72, p = 0.01, and frontal beta, Z = -2.64, p = 0.01, with lower spectral power values for the Overview Effect than the moment prior to it. This takes away the possibility that segment length may have biased the results, as differences in the beta frequency band are still present.

Taken together, the results of the analyses are in line with previous EEG research on awe where a stronger awe experience is reflected in lower spectral power over the beta frequency band. Contrary to the findings in the literature, however, no difference was found for the theta frequency band. Instead, the gamma frequency band showed a reduction in spectral power in all areas of the brain.

4 Discussion

The current study investigated EEG evidence for the Overview Effect. It used an immersive VR space journey and tested whether it induced the Overview Effect using self-reports. Prior studies either focused on self-report measures to obtain evidence for the Overview Effect or used neurological measures. The current study combined a variety of self-report measures – questionnaires and drawings – with spectral power values from EEG-signals.

The results are in line with self-reports from previous work where the Overview Effect led to a greater awareness of Earth and space (Ihle et al. 2006; Kanas 2020). The Overview Effect questionnaire answers correlated with the DPES-awe and Awe Experience questionnaires (Stepanova et al. 2019a; Yaden et al. 2018). The small-self was confirmed implicitly using pre- and post-test drawings of the self (Bai et al. 2017; Shiota et al. 2007).

In addition to the self-reported measures, the findings obtained from this study are in line with prior work on the Overview Effect concerning variation in spectral power values in left and right hemisphere beta bands (Gallagher et al. 2014; Reinerman-Jones et al. 2011; 2013), but extend to other brain areas. The beta band is associated with focused attention, decreasing in power during high-load cognitive tasks and when working memory load is higher (Baceviciute et al. 2021). This may be related to the need for accommodation in this context, which is the motivation to process newly acquired information and adjust internal models of the world (Gottlieb et al. 2018; Yaden et al. 2016). This kind of cognitive processing of altering mental schemas may thus be reflected by a lower beta frequency. That other brain areas (Gallagher et al. 2014; Reinerman-Jones et al. 2011; 2013) showed activation in our results may not come as a surprise because of emotion processing being known to occur in the frontal regions (Martin and Altarriba 2017; Schmidt and Trainor 2001), behavioral arousal occurring in the right parietal region (Schmidt and Trainor 2001), and correlations to awe having been found in the central brain area (Hu et al. 2017). While the current study did not yield any more conclusive evidence on the location of the Overview Effect in the brain, it showed the expected decrease in beta spectral power during the moment that Earth first comes into view.

Past EEG studies on both awe and the Overview Effect found activity change in the theta band (Chirico et al. 2020; Gallagher et al. 2014; Reinerman-Jones et al. 2011; 2013), which the present study did not observe. Previous research has indicated that a decrease in theta activity meant that participants were engaged and alert (Gallagher et al. 2014; Reinerman-Jones et al. 2011). In addition, theta activity has also been associated with encoding new information, increased memory load (Nigbur et al. 2011), and memory formation (Greenberg et al. 2015). It is possible that because we did not compare to a baseline where participants were not in virtual space, the null effect on theta denotes no difference in engagement, alertness, and memory formation throughout the entire simulation. Future research where a comparison with a neutral baseline could be conducted would provide more clarity on this phenomenon. A different possibility is that the theta activity change could have been due to spatial disorientation or motion sickness, as several of the referenced literature of Overview Effect-related research used VR-based applications (Li et al. 2015; Naqvi et al. 2015). Since the current study used a moving chair, the discrepancy between visual and physical movements was likely reduced. More research is needed to confirm this assumption, especially because past findings on motion sickness are not entirely conclusive (Lim et al. 2021).

Finally, we found lower spectral power in the gamma frequency band, which previous studies did not always include (e.g., Gallagher et al. 2014; Reinerman-Jones et al. 2011; 2013). Gamma power increases are associated with successful matches between memory and encountered stimuli, whereas decreases denote incongruency between the incoming visual or semantic information and internal models (Willems et al. 2008). As gamma rhythms are associated with bursts of insight, and high-level processing (Stinson and Arthur 2013), a drop in spectral power can indicate a violation of known mental structures (Penolazzi et al. 2009). Similarly, an awe-experience can lead to deactivation in the middle temporal gyrus (MTG), related to a discrepancy between existing knowledge and internal models of the world (Guan et al 2019b). Since Overview Effect eliciting experiences can lead to the need for accommodation as a result of expectancy violations (Yaden et al. 2016), decreased gamma activity can be expected. In principle, these experiences can be both positive and negative in valence (Keltner and Haidt 2003). A decrease in both beta, and gamma frequencies have been observed following stressful stimuli, accompanied by an increase in theta and alpha bands (Luijcks et al. 2015). These patterns were not observed in the current study. Interestingly, higher levels of beta and gamma have been associated with fear of heights in VR (Apicella et al. 2023), which could form a possible confound in our study, as subjects were placed at a far distance from the (virtual) Earth.

One finding within the study was that not all comparisons showed the expected results. For example, comparing pre-Overview Effect to Overview effect did not reach significance. A possible reason for the less salient difference is that awe can also be felt from seeing space (Reinerman-Jones et al. 2013). For the purposes of this study, the analysis focused on time frames associated with the Overview Effect (and awe) experience, as defined in prior research. A full comparison of all time frames with the Overview Effect can be found in Online Resource 2. It is worth noting that moments that do not show significant differences with the Overview Effect (T5) mostly involve views of space, or other awe-inspiring views such as the Terminator or the Northern Lights. Moments for which we observed significant differences in the EEG signal when compared to the Overview Effect are mainly parts of the simulation where factual information is provided (for example, about pollution, or the International Space Station).

The findings in this study should be seen in light of its limitations. The system that was used for recording was a 9-channel EEG, which is the same as has been used in prior research (Gallagher et al. 2014; Reinerman-Jones et al. 2011; 2013). We also used the same definition of regions for comparison purposes. This setup, however, inevitably caused overlap between regions (Gallagher et al. 2014; Reinerman-Jones et al. 2011; 2013). While a nine-channel EEG has clear advantages with respect to the speed of application, a higher number of channels can also provide a better spatial resolution. Even though we were able to describe EEG indicators of Overview Effect-inspired awe in the frequency oscillations, we were not able to identify the specific brain areas potentially involved in the process. We do not think this had a major effect on the findings, however, further research is needed to ensure the replicability of the results.

Also, as with most experimental psychology studies, the participant group consisted of a homogenous sample of Dutch-speaking students, within a relatively narrow age-range. This choice helped us minimize potential age-related effects on the EEG output (Duffy et al. 1993; Stacey et al. 2021) but it remains to be seen to what extent the results generalize to a broader population, across ages and across cultures to determine the generalizability of the Overview Effect.

In conclusion, the findings of the current study confirm that the Overview Effect can be induced using a VR journey into space. We presented evidence of a reduction in spectral power in the beta and gamma EEG frequency bands when awe and the Overview Effect are experienced. A reduction in spectral power in the beta frequency corresponds to increased cognitive load, and a reduction in spectral power in the gamma frequency is associated with failure to match memory with sensory information. Both can be linked to the disruption of mental structures, likely due to the need for accommodation arising from awe experiences. Based on the outcomes of the current study, we suggest that future research into awe-phenomena should explore the process of disruption in mental models, typically accompanying awe experiences, in more detail.