Since the appearance of the term Nature Deficit Disorder in 2005 by Richard Louv, many studies have tried to make people aware of the importance of visiting nature for their personal health (Martínez and Caballo, 2020). Scientific literature has for years shown the importance of exposure to natural environments as opposed to urban environments, being associated with better psychologycal health (De Vries et al., 2003), better psychological health (Ferguson and Evans, 2018) and even better physical-sports health (Eigenschenk et al., 2019).

Scientific evidence shows that when a person observes a scene, photo or video of a natural environment compared to an urban environment, positive psychological effects occur, such as a reduction in anxiety (Ulrich, 1981), in physiological and psychological stress (Hartig et al., 2003; Tyrväinen et al, 2014 ), a reduction, in work-related stress (Bjørnstad et al., 2015; Korpela et al., 2015; Sianoja et al., 2018), and even an improvement in mood (Hartig et al., 2003; Stigsdotter and Grahn, 2004). The effect on observation of these images is such that even positive effects on the satisfaction and well-being of coronary and pulmonary patients have been found simply by observing environments in nature or with houseplants from a window (Raanaas et al., 2010).

Currently, only seven studies have compared the effects of human exposure to a natural, green environment with an unnatural (urban or indoor) environment (Aspinall et al., 2013; Bailey et al., 2018; Bratman et al., 2015; Joung et al., 2015; Neale et al., 2017; Park et al, 2007; Tilley et al., 2017), with little research now using image visualisation tasks with Neuroimaging (Chang and Chen, 2005; Kim et al., 2010; Kim and Jeong, 2014; Lee, 2017; Martínez-Soto et al., 2013; Roe et al., 2013; Tang et al., 2017; Ulrich, 1981).

Regarding the brain impact of urban images compared to nature images, some EEG results (Aspinall et al., 2015; Essawy et al., 2014; Caballero et al., 2018) have shown that in a natural environment brain activity was associated with a more meditative state, represented by a greater presence of Alpha waves, while in an urban environment (Choi et al., 2015), brain activity was associated with higher levels of Beta waves and greater psychological stress. Closer to the methodology of this study, the limited work carried out with fMRI found that visualisation of images of urban environments compared to natural environments was associated with increased activity in prefrontal and occipital areas, as well as temporal areas, (Kim et al, 2010; Lee, 2017), parts of the limbic system (Kim et al., 2010; Kim and Jeong, 2014 ), and default mode network (DMN) areas such as the cuneus and posterior cincture (Martinez-Soto et al., 2013; Tang et al., 2017). Meanwhile, nature scenes, compared to urban images, produced greater activation in subcortical areas such as the insula, associated with higher-order emotional and cognitive processes, and the basal ganglia, involved in voluntary movements (Kim et al., 2010). Furthermore, the activation of areas of the frontal, parietal and upper temporal cortices and the occipital cortex were also associated with the visualisation of natural scenes (Martínez-Soto et al., 2013). Interestingly, those studies that in addition to analysing brain processing with neuroimaging used validated subjective tests showed that the visualization of urban environments was associated with responses of fear (Kim et al., 2010), stress (Choi et al., 2015) and sadness (Ulrich, 1981), in addition to being associated with over-exertion in cognitive processing tasks (Martínez-Soto et al., 2013; Tang et al., 2017). On the other hand, activation when seeing natural environments was associated with happiness (Kim et al., 2010), well-being (Chang and Chen, 2005),, cognitive restoration (Martínez-Soto et al., 2013; Tang et al., 2017) and improved attention (Ulrich, 1981), even looking for differences according to the type of landscape stimulation (with/without water, with/without trees etc.) (Tang et al., 2017).

In addition to visualizing images, Aspinall et al. (2015) and Chen et al. (2016), carried out a recording of brain activity with EEG in an outdoor activity, obtaining results that corroborated those obtained in the laboratory about stress and full attention shown by the aforementioned research. In the case of fMRI and the visualisation of nature images, work has been carried out mostly with young adults (Kim et al., 2010; Kim and Jeong, 2014; Lee et al., 2015; among others), and in primary school children (Eigenschenk et al., 2019), but never comparing adult sportsmen and women specialised in the practice of their favourite sport in a particular type of environment (outdoor or indoor). As stated by Grassini et al. (2019), few studies have yet used the fMRI technique to test the effect of natural environments and urban environments at the brain level, so this research will help to fill this theoretical gap. Therefore, the first objective of this research is to study the differences between athletes who practice outdoor sport (OG) and those who practice indoor sport (IG) in their levels of stress and awareness. On a second objective, we aimed to explore the brain reactivity, in these two groups (OG and IG), to the visualization of images of natural environments in comparison with images of urban environments, where they practice their sports. The last objective would be to study the associations between the psycho-emotional measures and the brain reactivity related to nature and urban environments in each group separately. To do so, we study the differential relationship of OG and IG in the previously mentioned brain activation with the levels of perceived stress and full attention/consciousness. As a result, two hypotheses were established. First, we expect OG to present better levels of awareness and lower levels of perceived stress (Berto, 2014). We also expect to find a differential brain activation in the OG, compared to the IG, in frontal areas when visualising images of nature compared to urban images (Elsadek et al., 2019; Yu et al., 2017), and that these resulting areas are positively related to levels of awareness and negatively related to levels of stress in the OG, more than in the GI (Berto, 2014; Brown et al., 2013).

A convenience sample of 49 participants were recruited throughout the province of Granada (Andalucía, Spain) and divided in two groups depending on their preferred areas for physical exercise: outdoor group (OG; n = 21) and indoor group (IG; n = 25). Participants were informed about the study’s aims, signed an informed consent form and completed one session including an fMRI scan session and the assessment of several self-report questionnaires. Besides, a sample of hair was collected from the participants to measure hair cortisol accumulation (see Measures for more details). Inclusion criteria included (i) ages between 18 and 65 and (ii) being a habitual sportsperson (at least exercising 3 times a week during last five years). In this case, for the OG group, they had to exercise outdoors and conversely, the IG had to exclusively exercise indoors. In both cases, water sports were excluded to avoid confusion with natural environment (Grassini et al., 2019). On the other side, exclusion criteria were: (i) having any psychological disorder or any medical or nervous system disease diagnosed, (ii) take any medication regularly, (iii) having suffered an epileptic attack, (iv) taking any drugs and drinking regularly (3 times for week) any contraindication to undergo the fMRI session (i.e. metal prosthesis or claustrophobia). All participants had normal or corrected-to normal vision.

The data from each participant were collected in a same-day session of approximately 1 hour and 20 minutes of total duration. In the first place, all the participants completed an fMRI task within the scanner. After that, self-report questionnaires were filled out and a sample of hair was collected in order to assess cortisol concentrations. Further neuropsychological tests beyond the study’s aims were completed by the participants.

The Human Research Ethics Committee of the University of Granada approved the study with number 621/CEIH/2018, and all the participants were informed about the aim of the study and signed an informed consent.

All participants performed an fMRI scan in a 3T Tim Trio Siemens equipment located at the Mind, Brain and Behavior Research Center, University of Granada, with a total duration of 7 minutes and 16 seconds equipped with a 32-channel phased-array head coil. A T2*-weighted echo-planar imaging (EPI) was obtained (TR = 2000ms, TE = 25ms, FOV = 240 x 240mm, 68x68 pixel matrix; flip angle = -0.4º, number of slices = 35, 160 whole-brain volumes). Additionally, a high resolution T1-weighted anatomical image was also acquired for each participant. This image, of 176 slices (TR = 1900ms; TE = 2.52ms; flip angle = 9°; FOV = 256 x 256 mm; slice thickness = 1 mm) was used to discard structural alterations and for the co-registration step in the preprocessing stage. In addition, 3D field maps were also collected by acquiring a gradient image with two echoes. Thus, two magnitude images were generated (TR = 468ms; TE1 = 4.92ms; TE2 = 7.38ms, FOV = 240x240mm; 68x68 pixel matrix; number of slices = 35) together with a phase map, showing the difference between the two echoes.

This study shows an experimental design. An image-viewing task including images of nature and urban areas was used. All pictures were shot ad hoc for the study using standardized presentation and lighting conditions. Furthermore, all the photographs (urban and nature) were taken at the morning, namely, with the same ambient light. Three main types of images were utilized: “green nature”(defined as those images captured in field areas with vegetation where no urban area is depicted), “water nature” (defined as field areas without urban features, as previous images, but this time including water extensions and leafier fields) and urban (defined as those images depicting city center, buildings, traffic and urban-related characteristics). In each trial, each of the pictures was presented and participants had to rate these pictures. Participants were instructed to choose between two rating options (pleasant or unpleasant) according to their own preferences. Each trial began with a fixation cross, which appeared for approximately 1s. Then, images appeared until the participant chose if they seem pleasant or unpleasant. An inter-trial interval was presented after the rating to adjust each trial duration in approximately 5200ms.Then, the fixation cross was presented again. The order in which the images were presented was counterbalanced among the participants. There was a total of 80 image-viewing trials. After each trial, participants were instructed to press a button to rate the images as pleasant or unpleasant. This button pressing did not form part of the objectives of this study but, instead, allowed to assure that the participants remain alert to the images during the whole task.

We used the gold standard method to assess cortisol in hair, which has been validated and provides a test-retest reliability as a biomarker of chronic cortisol exposure in adults (Russell et al., 2012). Hair samples consisting of approximately 150 strands of hair were collected from the posterior vertex with a length no greater than 3 cm (assuming an average growth rate of 1 cm/month; Russell et al., 2012). A 3 cm segment contains the cortisol that has been deposited over approximately the last 3 months. The hair was then wrapped in a piece of aluminum foil to protect it from light and humidity, and was stored in an envelope in room temperature. Later, the hair samples were sent for analysis to the Faculty of Pharmacy of the University of Granada. The samples were weighed and ground to a fine powder to break up the hair’s protein matrix and increase the surface area for extraction using a ball mill (Bullet Blender Storm, Swedesboro NJ, USA). Cortisol from the interior of the hair shaft was extracted into HPLC-grade methanol by incubation of the sample for 72 hours at room temperature in the dark with constant inversion using a rotator. After incubation, the supernatant was evaporated until completely dry using a vacuum evaporator (Centrivac, Heraeus, Hanau, Germany) and the extract was reconstituted in 150 ul of phosphate buffered saline (PBS) at pH 8.0. The reconstituted sample was immediately frozen at -20ºC for later analysis (Sauve et al., 2007). The cortisol in the hair sample was measured using the Salivary ELISA Cortisol kit© (Alpco Diagnostics®, Windham, NH) as per the manufacturer’s directions with the reagent provides. The cross-reactivity, as reported by the manufacturer, is as follows:

Prednisolone 13,6%, Corticosterone 7,6%, Deoxycorticosterone 7,2%, Progesterone 7,2%, Cortisone 6,2%, Deoxycortisol 5,6%, Prednisone 5,6%, and Dexamethasone 1,6%. No cross reaction was detected with DHEAS and Tetrahydrocortisone.

The Spanish version (2.0) of the Perceived Stress Scale (PSS; Cohen et al., 1983 and adapted by Remor, 2006) was used. This scale is a self-report questionnaire assessing the level of perceives stress during the last month. It comprises 14 items with a Likert-like scale ranging from 0 (never) to 4 (very often). It is noteworthy that the total score of the PSS is obtained by reversing the scores of the following items: 4, 5, 6, 7, 9, 10 y 13 and adding the other items’ scores. Higher scores in this scale represents higher levels of perceived stress. Importantly, this scale provides an index of perceived chronic stress and has previously shown stability through repeated measures tests. The Cronbach’s Alpha value was .87.

The Spanish version of the Mindful Attention Awareness Scale (MAAS; Brown and Ryan, 2003; Johnson et al., 2014; Soler et al. 2012), was used to assess the general ability to be aware and to focus the attention to what is taking place in the present moment for each situation (i.e., focus on the sensory properties of food when eating). Thus, this test reports a measure of trait mindfulness. This instrument consists of 15 items, 6-point Likert-type responses ranging from 1 (almost always) to 6 (almost never). All the items form a unique total score measuring attention-awareness with higher scores showing higher awareness. The Cronbach’s Alpha value was .92.

Functional data was preprocessed using the CONNv17 functional connectivity toolbox (Whitefield-Gabrielli and Nieto-Castañón, 2012), implemented in MatlabR2017a (The MathWorks Inc., Natick, Massachusetts, USA). The preprocessing pipeline included realignment, unwrapping and distortion correction (with fieldmaps), denoising of motion artifacts and head motion (following aCompCor; Power et al. (2012)), segmentation, co-registration to each participant’s anatomical scan, normalization, re-sliced to a 2mm isotropic resolution in MNI space and smoothing using a Gaussian kernel of FWHM 6mm. Additional steps after denoising included band-pass filtering of the BOLD timeseries (between 0.008-0.09 Hz) and linear detrending.

The conditions of interest were modelled from the onset time at which the image was presented to the time at which participants responded. Task regressors were convolved with the hemodynamic response function. Our contrasts of interest were “green nature vs urban” “water nature vs urban” and “green nature vs water nature”, as defined in the first-level analyses (single subject) and in the second level between-group analyses.

The minimum threshold extent for the activation analyses were estimated by 1000 Monte-Carlo simulations using the cluster-extent based AlphaSim thresholding approach, implemented in the SPM RESTplusV1.2 toolbox. For the between-group effects in the activation under the image-viewing task, the required cluster extent was calculated including as input parameters an individual voxel threshold probability of 0.001, a cluster connection radius of 5mm and the actual smoothness of imaging data after model estimation, incorporating a whole-brain image mask (184283 voxels). The minimum cluster size was 109 voxels (327 mm³).

To explore the relationship between brain activation during the exposure to nature/urban images and the data resulting from the questionnaires, we extracted the peak beta eigenvalues from each significant result of between-group brain differences. Using partial correlations, these brain values were correlated with physiological (cortisol) and psychological (perceived) stress and awareness in each group separately. Also, we correlated both measures of stress (perceived and physiological). Sex, age, body mass index, sport level and education were used as covariables of no interest (as in the neuroimaging analysis). Further Fisher’s tests and Zou’s intervals were carried out to explore the potential between-group differences in these correlations. Specifically, to explore whether the correlation between a resulting brain network and the psychological measures in outdoor group differed from the correlation between those two variables in the indoor group.

No between-group differences were shown in the main demographic variables of the study. However, we found between-group differences in perceived stress and the total score of MAAS with the outdoor group (OG) showing higher MAAS scores and lower perceived stress than the indoor group (IG). These results are shown in Table 1. Besides that, physiological cortisol and perceived stress did not significantly correlate in any of the groups.

Between-group differences during the exposure to green nature versus urban images were shown. The OG showed higher activation in the middle occipital cortex and in a cluster comprising the supplementary motor area (SMA), the premotor cortex and the pre-SMA compared to the IG. No other result surpassed the threshold of significance. (Table 2).

In OG, the higher activation of the left SMA cluster negatively correlated with perceived stress (R = -0.767; p = 0.001), while in the IG, and curiously not in the OG, the higher premotor cortex activation was positively related to the total score on MAAS (R = 0.628; p = 0.003). No significant association was found with the hair cortisol levels. Fisher’s Z and Zou’s interval (p = 0.045 and -0.8307-0.0062, respectively for perceived stress, and p=0.001 and 0.3715 1.3224 for the MAAS score) confirmed that these correlations showed a significant between-groups differences. (Figure 1 and Figure 2).

This research aimed to find out the differences in psycho-emotional processes and brain activation when faced with the task of visualising images in a group of sportsmen and women practising in a natural environment (outdoor) as opposed to urban sportsmen and women (indoor). As a third objective, this study sought to explore the relationships between this differential activation and its relationship with stress and full attention/awareness in the two groups separately.

Concerning our first objective, the results show significant differences in the perception of stress, with participants who belonged to the OG perceiving less of it. On the other hand, the OG obtained higher values in the MAAS test, which would imply greater attention and awareness. However, no differences were found in accumulated cortisol. These results would be partially in line with other research that suggests an improvement in stress levels and attention/awareness, not only when exposed to nature ((Dijkstra et al., 2008; Johnsen, 2013; Mutz and Müller, 2016)), but also when practising sports in natural environments (34-35). Therefore, this study shows that the practice of exercise, specifically in natural environments, would add more benefits to the already known advantages of practising sport on stress reduction and full attention/awareness (35), thus establishing the relevance of the type of environment in which we choose to perform physical activity to improve our processes of emotion regulation. Contrary to the results in psychological stress, we found no differences in physiological stress. This would go against other research that has measured stress using saliva (23), although we have to take into account that these studies have methodological differences from ours since previous studies have taken measures of acute stress, while our study has collected a sample of cortisol in hair, that is to say, it reports an index of accumulated stress during the last three months. Also, a recent systematic review has shown that the results of self-reported measures give the most convincing evidence of the beneficial effect of outdoor sports on stress (Kondo et al., 2018). Future studies will reveal whether sport in natural environments is useful in reducing physiological stress levels in the long term and not only in the short term, as has been studied so far.

In the case of the second objective, the OG showed greater activation in the mid-occipital cortex, premotor cortex and SMA, areas related to movement control and management, mental and motor imagery, but also to high-level cognitive processes such as decision-making, memory and attention, among others (Nachev et al., 2007; Tanaka et al., 2005). These areas found while viewing green natural environments differ from those already presented by other authors on urban environments (Kim et al., 2010; Kim and Jeong, 2014; Martínez-Soto et al., 2013; Tang et al., 2017). Furthermore, we did not find any significant difference in the “urban versus water nature” contrast, unlike other previous studies (Tang et al., 2017). These authors also found that two of the studied natural stimuli (mountain and river) produced a similar perceived restorative potential. Thus, the difference in results might be due to the characteristic of our sample. Our sample was comprised of sportsmen and sportswomen who did not perform any physical activity related to water sports (due to exclusion criteria). This particularity, although speculatively, might lead to a higher restorative perception of the green nature, mainly by the outdoor group as green nature is the environment they are used to. Despite this, our data support the results of Zhang et al., (2019), in which they find the involvement of SMA and the occipital cortex in the processing of images of green environments. The difference in areas found between the different studies and ours can be explained by one main reason and that is the sample used by our study, which for the first time used a sample with two groups of athletes who have been doing sports in nature or in an urban indoor environment at least three times a week for at least five years. Therefore, we are dealing with a population with long exposure to a particular type of environment. The most similar results were found in the study by Wood et al. (2014) in which they found brain activation in motor areas, but in this case when listening to sounds related to their favourite sport. This would suggest the involvement of these areas of “motor preparation” in the recognition of and response to known stimuli. In fact, Aspinall et al. (2013) have already proposed that the better known the sensory inputs, the greater the activation of these areas. This would also explain why the OG group shows more activation in these areas than the IG group. Another similar study is that of Davand et al. (2018) who carried out research with school children who had grown up in green areas and found that exposure to these natural areas was positively associated with more grey matter in the left premotor cortex and with a greater volume of white matter in the left premotor area cluster, among others. All these results together could suggest that people accustomed to natural environments have a better functioning of these motor areas. Furthermore, these motor areas, together with visual-occipital areas, are also involved in motor imaging (Stevens, 2005), and walking (La Fougere et al., 2010).

Taking into account all the above, we could think that prolonged exposure to green space can affect the increased reactivity of the brain to certain stimuli, presenting activations of brain areas different to those shown by subjects with prolonged exposure to another type of environment, such as the urban one. For this reason, the visualisation of natural stimuli, where the OG group carries out its physical activity, could have triggered greater imagery than in the IG group. This contribution should be contrasted with later studies.

Regarding our last objective, we found that the OG, rather than the GI, had a negative correlation between SMA cluster activation and perceived stress. Conversely, and unexpectedly, the IG, rather than OG, obtained a positive correlation between activation of the premotor cortex and MAAS, which measures awareness and full attention. Both MAAS and the premotor cortex have been related in sports subjects to motor imagery (Lorey et al., 2010; Mizuguchi et al., 2012), but also to mental imagery of various kinds (Frank et al., 2014). The negative SMA-perceived stress relationship in OG is a novel finding, as the SMA is not conceived as an area responsible for stress regulation. However, some previous studies have involved the SMA in other cognitive processes (Tanaka et al., 2005), and more importantly for our purposes, in emotional regulation processes, specifically in “up-regulation” processes (Frank et al., 2014; Kohn et al., 2014; New et al., 2009). This result would suggest that in view of the exposure to natural environments in the OG group the activation of the SMA would help to increase the positive affect of these athletes, something that is related to their predisposition to lower perceived stress.

On the other hand, we found a positive relationship between activation of the premotor cortex and greater awareness in IG than in OG. The association between the premotor cortex and awareness seems to be supported by previous studies. According to an influential meta-analysis, the premotor cortex is associated with voluntary regulation of thought and meditation (Fox et al., 2016) and processes such as mind-wandering would involve this region (Hasenkamp et al.,). Despite this, this finding is contrary to the hypothesis of this study, where we expected a greater relationship in the OG than in the GI. One possible explanation is that natural settings (like those shown in the study images) provoke higher states of meditation than any other environment in most individuals (Aspinall et al., 2013), but athletes who practice sports in natural environments have a greater amount of external stimuli during their practice that provides less opportunity to focus on their body signals, thoughts and emotions than those who exercise in indoor environments with few distracting elements (Bigliassi et al., 2020).

This study also has a number of limitations that need to be addressed. For example, one would be that the brain-emotion associations found in our results do not allow us to know if brain functioning is what predicts or is the consequence of the processing/regulation of emotions (stress and full attention), so we cannot determine cause and effect relationships between variables. Evaluation of longitudinal measures could solve this limitation in future studies because, as Norwood et al. (2019) state, there is still a need for further research to understand the effects of various environments on the brain (Coburn et al., 2017; Frumkin et al., 2017). Another limitation is the difficulty in controlling the prior exposure of each subject to their sporting environment. Besides, this paper compares subjects who, on some occasions, practice similar sports (track racing with mountain racing), but in different environments. The type of activity chosen should also be taken into account in future studies. Therefore, the outcome of this study emphasises the need to carry out more research on this subject in order to understand the effects of different environments on the brains of sportsmen and women. Finally, our study lack of qualitative data to support the individual’s perception expressed in the psychological questionnaires. As regards strengths, we can highlight the novelty of using a measurement technique that is unusual in these studies, fMRI, which allows us to broaden existing theoretical knowledge and bring up the brain as a possible mediator between the type of environment chosen for physical activity and better processing/regulation of emotions. In addition, for this study we have used various assessment techniques that take into account physiological factors, such as the measurement of cortisol accumulation in the hair, cerebral, with the fMRI and psychological factors. Few if any studies so far have taken into account so many different measurement techniques in a single work. Therefore, the data obtained in this research would provide a new line of work to the already existing literature that would need further investigation.

Altogether, our findings show that the activation of motor brain areas while viewing images of green nature associates with a lower perceived stress in those participants who exercised outdoor, but not in those who exercised indoor. These results highlight the relevance of the type of environment chosen for exercising over psychological wellbeing, as supported by the lower perceived stress in the outdoor group. In addition, they also pinpoint a specific pattern of brain activation that could act as a mediator in the association between the aforementioned variables, which is to say, type of environment while exercising and psychological wellbeing.