Occupational functional plasticity revealed by brain entropy: A resting‐state fMRI study of seafarers

Abstract

Recently, functional magnetic resonance imaging (fMRI) has been increasingly used to assess brain function. Brain entropy is an effective model for evaluating the alteration of brain complexity. Specifically, the sample entropy (SampEn) provides a feasible solution for revealing the brain's complexity. Occupation is one key factor affecting the brain's activity, but the neuropsychological mechanisms are still unclear. Thus, in this article, based on fMRI and a brain entropy model, we explored the functional complexity changes engendered by occupation factors, taking the seafarer as an example. The whole‐brain entropy values of two groups (i.e., the seafarers and the nonseafarers) were first calculated by SampEn and followed by a two‐sample t test with AlphaSim correction (p < .05). We found that the entropy of the orbital‐frontal gyrus (OFG) and superior temporal gyrus (STG) in the seafarers was significantly higher than that of the nonseafarers. In addition, the entropy of the cerebellum in the seafarers was lower than that of the nonseafarers. We conclude that (1) the lower entropy in the cerebellum implies that the seafarers’ cerebellum activity had strong regularity and consistency, suggesting that the seafarer's cerebellum was possibly more specialized by the long‐term career training; (2) the higher entropy in the OFG and STG possibly demonstrated that the seafarers had a relatively decreased capability for emotion control and auditory information processing. The above results imply that the seafarer occupation indeed impacted the brain's complexity, and also provided new neuropsychological evidence of functional plasticity related to one's career.

1. INTRODUCTION

1.1. fMRI and brain entropy

After the development of magnetic resonance technology, in the 1990s, Ogawa, Lee, Kay, and Tank (1990) first proposed blood‐oxygen‐level‐dependent functional magnetic resonance imaging (BOLD‐fMRI), such that brain structure imaging is extended to brain functional imaging. Estimating the complexity of BOLD‐fMRI signals can help detect the different characteristics of brain activity associated with age, development, diseases, and so on, as well as provide more detailed insights into clinical and biomedical applications (Ferenets, Vanluchene, Lipping, Heyse, & Struys, 2007; Richman and Moorman, 2000; Sokunbi et al., 2011; Sokunbi, Cameron, Ahearn, Murray, & Staff, 2015; Xie, Guo & Zheng, 2010; Yang et al., 2013). This complexity can be estimated by measuring the entropy values of the signal. Several studies have shown that the characterization of the brain based on its complexity may provide a better understanding of the health and robustness of an individual (Goldberger, Peng, & Lipsitz, 2002), the ability to adapt to the aging brain (Sokunbi et al., 2011), the cardiovascular disease (Richman and Moorman 2000), and the neural effects of drugs (Ferenets et al., 2007). The human brain, as we know, is one of the most complex biological systems. To maintain its normal operation, it is extremely necessary to maintain the brain entropy in a certain range (Bergström, 1969; Singer, 2009; Wang, Li, Childress, & Detre, 2014). Many factors contribute to the changes in the complexity of the system, such as age, sex, diseases, and occupations (Drachman, 2006; Gao, Zeng, Wang, Shi, & Chen, 2016; Hervais‐Adelman et al., 2015; Li, Fang, Hager, Rao, & Wang, 2016; Shen et al., 2016; Sokunbi et al., 2013, 2011, 2015; Sokunbi, 2014; Wang, 2016; Wang, Zeng, Shi, & Yan, 2017; Yang, Hou, Lu, Luo, & Yao, 2016; Zeng, Wang, Shi, & Yan, 2016). Recently, the sample entropy (SampEn) as a representative has been widely used to characterize the alteration of brain function with single‐scale and multi‐scale biological signals using EEG (Abásolo, Hornero, Espino, Alvarez, & Poza, 2006), MEG (Gómez and Hornero, 2010), and fMRI (Wang et al., 2014; Sokunbi, 2014; Sokunbi et al., 2015).

1.2. Occupational functional plasticity

The brain is a complex dynamical system with structural and functional plasticity; the brain's morphological structure and functional activity patterns can be altered by internal or external factors such as learning, development, aging, and lesions (Sporns, 2011). Recently, many studies have shown that there is a significant correlation between the brain's structural/functional plasticity and people's occupations. Many occupations were studied including musicians, expert athletes, acupuncturists, simultaneous interpreters, taxi drivers, professional sailors, and so on (Dong et al., 2013, 2015; Hervais‐Adelman, Moser‐Mercer, Michel, & Golestani, 2015; Park et al., 2009, 2011; Schlaug, Jäncke, Huang, Staiger, & Steinmetz, 1995; Shen et al., 2016; Wang et al., 2017; Wei, Zhang, Jiang, & Luo, 2011; Yang et al., 2016).

For example, musicians who began their musical training at an early age owns larger anterior midsagittal corpora callosa compared with nonmusician controls or musician who started training later (Schlaug et al., 1995). Additionally, Yang et al. (2016) found that for professional composers, the functional connectivity strength among certain paired resting‐state networks had the significant changes between two resting‐state conditions, one before the composition task, and the other after the composition task. For the athletes, some studies further suggest that continued practice and repetitive performance of basketball‐related motor skills may induce plastic structural changes in the cerebellum vermis and striatum (Park et al., 2009, 2011); In addition, a significant positive correlation between the diver's status and cortical thickness in the right parahippocampal gyrus was reported; this correlation might be due to the effect of extensive training on the diver's brain structure (Wei et al., 2011).

Besides, Dong et al. (2015) found that the acupuncturists had higher amplitude of low‐frequency fluctuations (ALFF) in the left ventral medial prefrontal cortex (VMPFC) and in the contralateral hand representation of the primary somatosensory area compared to the matched nonacupuncturists. Hervais‐Adelman et al. (2015) used the BOLD‐fMRI scanning to reveal that brain functional plasticity is associated with the emergence of expertise in extreme language control by exploring the functional response of participants with simultaneous interpretation training. Surprisingly, they found that during simultaneous interpretation, the selection and control functions of the lexico‐semantic system was closely related with the caudate nucleus and that the continued control of language output depended on the putamen. Shen et al. (2016) focused on the functional connectivity differences between professional taxi drivers and nontaxi drivers and found that a significant difference in the connection pattern existed in the early warning functional network. This difference was used to distinguish between professional taxi drivers and nontaxi drivers with an accuracy of 90%. Wang et al. (2017) showed that compared to nonsailors, sailors had one distinct atomic functional connectome that was made up of four specific subnetworks: the auditory network, the visual network, the executive control network and the vestibular function‐related network. This connectome was most likely linked to sailing experience, that is, continuously being subjected to auditory noise, having to maintain balance, locating one's position in three‐dimensional space while at sea, and obeying orders (Gao et al., 2016; Wang et al., 2017; Zeng et al., 2016). All the above studies have showed that a person's occupation has a significant effect on brain functional plasticity.

1.3. Study purpose

From the previously outlined studies, it is clear that the functional plasticity of the brain allows it to adapt to occupational factors and to form specific modes of brain activity, which could be potentially characterized by brain complexity (Tononi, Edelman, & Sporns, 1998). Many studies have also demonstrated that brain entropy is an effective tool for capturing the changing in the sophisticated brain system. However, knowledge regarding the relationship between the occupational influences and the brain complexity is still lacking. In this study, using the professional seafarer as an example and taking our measure of complexity to be brain entropy, we explore about how long‐term occupational factors influence the functional complexity of the human brain (Tononi et al., 1998). We expect that this work will increase our understanding of the relationship between brain functional plasticity and occupation.

Specifically, as we have stated in our previous study, seafarers are highly suited for use in exploring the relationship between brain functional plasticity and career experience due to their occupational stability (i.e., the seafarers are often engaged in their jobs for long periods of time) and their professional particularities. These include being exposed to the marine working environment (small working space with high levels of machine noise), single‐sex colleagues (all male sailors), and long periods of isolation from their families. Additionally, this job requires professional maritime training and skills, good psychological health, strong environmental adaptability, ability to follow a chain of command strongly, and so on (Wang et al., 2017). Due to the stability and particularity of a sailor's occupation, we speculate that the long‐term sailing experience and career training of seafarers alters the brain complexity in some brain regions; such an outcome could help to characterize the intrinsic neural substrates of career experience. We test this hypothesis as described in the following sections.

The remainder of this article is organized as follows: the section Materials (Section 1.1) and Methods is first presented, followed by the Results and Analysis (Section 1.2), where we present our results for the sailor and nonsailor groups. Finally, in the Discussion (Section 1.3) and Conclusion (Section 2), we discuss how brain functional plasticity is related to the sailors’ career training and experience.

2. MATERIALS AND METHODS

2.1. Data acquisition

Twenty male professional seafarers (ages: 42–57 years, mean age = 49 years old, right handed) from various positions, for example, mate, helmsman, and seaman, were recruited from a shipping company in Shanghai, China. All seafarers had about 10–20 years of experience in navigation. For the nonseafarers, 20 Chinese male participants (ages: 48–55 years, mean age = 51 years old, right handed), were recruited from land‐based jobs (i.e., campus landscaping and office support) at the university or secondary school campus. All the subjects in the nonseafarer group were without the navigational skills, maritime professional training, or long‐term experience on the sea. The ages of the participants in the two groups were matched. The education levels for all the participants were junior college or equivalent. All participants were informed about the purpose of the study and signed a written consent form according to procedures approved by the IRB of East China Normal University (ECNU). The participants had no reported history of neurological or psychiatric disorders and were further evaluated by the symptom checklist 90 (SCL‐90) (Derogatis, Rickels, & Rock, 1976) with no abnormalities. During the data acquisition phase, all participants wore earplugs and were instructed to keep the body alive with their eyes closed and to stay relaxed. The resting‐state BOLD fMRI data for each participant was collected at Shanghai Key Laboratory of Magnetic Resonance at ECNU. The specific scan parameters were as follows: GE 3.0 T using a gradient echo EPI, a total of 36 slices covering the whole‐brain area, 160 time points, TR (time of repetition) = 2 s, matrix size = 64 × 64, in‐plane resolution = 3.75 mm × 3.75 mm, and slice thickness = 4 mm. This dataset was also reported in our previous research (Wang et al., 2017).

2.2. Formulation of brain entropy

The sample entropy (SampEn) is a complex statistical measure in biological data processing and may solve the shortcomings of nonlinear time series analysis techniques such as the correlation dimension (Pritchard et al., 1994) and Lyapunov exponent (Wolf, Swift, Swinney, & Vastano, 1985), which require large data sets and assume that the time series is stationary. This assumption is problematic for most biological data sets (Sokunbi, 2014). Specifically, the sample entropy is only slightly affected by low‐level noise and is robust to the occasional very large or small artifacts; it also gives meaningful information with reasonable data lengths (Richman and Moorman, 2000; Sokunbi, 2014). Furthermore, Wang et al. (2014) have shown that the SampEn is a robust and sensitive entropy measure for differentiating generic and fMRI signals with different regularities and that SampEn‐based fMRI‐derived brain entropy is a reliable brain activity measure, giving distinct values from a nonliving object and background noise, and is sensitive to task‐induced regional brain activity alterations. The relatively low brain entropy value in the cortex may reflect the “higher” psychological function of the cortex (Wang et al., 2014).

Next, we will give the mathematical definition of the SampEn. For a time series with $N$ time points $\{u(j)|1\le j\le N\}$, we can form the vector $X_m(i)=\{u(i+k):0\le k\le m-1\}$ with a length given by $N-m+1$ where $\left\{i\right|1\le i\le N-m+1\}$ and $m$ is a predefined dimension. By defining the maximum distance between $X_m(i)$ and $X_m(j)$ as $d[X_m(i),X_m(j)]=\max \left\{\right|u(i+k)-u(j+k)|:0\le k\le m-1\}$, the distance between $X_{m+1}(i)$ and $X_{m+1}(\mathrm{j})$ can be expressed as $d[X_{m+1}(i),X_{m+1}(j)]$. We further define $B_i^m(r)=B_i{(\mathrm{N}-\mathrm{m}-1)}^{-1}$, where $B_i=\mathrm{num\{}d\mathrm{[}{X}_m\left(i\right),X_m(j\mathrm{)]}<r\}$ and $j$ ranges from 1 to $N-m$( $j\ne i$). Similarly, we can obtain $A_i^m(r)=A_i{(\mathrm{N}-\mathrm{m}-1)}^{-1}$ and $A_i=\mathrm{num\{}d\mathrm{[}{X}_{m+1}\left(i\right),X_{m+1}(j\mathrm{)]}<r\}$. Then, the SampEn can be formulated as follows:

$$B_{}^m(r)={(\mathrm{N}-\mathrm{m})}^{-1}{\displaystyle {\sum }_{i=1}^{N-m}{B}_i^m(r)}$$ (1)

$$A_{}^m(r)={(\mathrm{N}-\mathrm{m})}^{-1}{\displaystyle {\sum }_{i=1}^{N-m}{A}_i^m(r)}$$ (2)

$$\text{SampEn}(\mathrm{m},\mathrm{r},\mathrm{N})={-\ln [\mathrm{A}}^{\mathrm{m}}(r)/B^{\mathrm{m}}(r)].$$ (3)

The SampEn is a measure of system complexity, where a large value represents unpredictability and random variation (high complexity), and a low value represents predictability and stable structure (low complexity). If uncertainty and complexity are increased, then the SampEn increases and vice versa.

2.3. Data processing

The data processing procedure included four major steps: data preprocessing, brain entropy calculation, statistical analysis, and the display of results.

In the data preprocessing step, the popular SPM12 software developed by KJ Friston and his colleagues (http://www.fil.ion.ucl.ac.uk/spm/software/) was applied with the discarding of the first 10 volumes, slice timing, realignment, normalization, and smoothing with a full‐width at half‐maximum (FWHM) Gaussian kernel equal to 8 mm.

The brain entropy maps for each subject in each group, that is, the seafarers and non‐seafarers, were calculated based on Equation (3), with the corresponding parameters $m$ and $r$ empirically set to 2 and 0.30, respectively, as other studies have suggested (Chen, Zhuang, Yu, & Wang, 2009; Richman and Moorman, 2000; Sokunbi, 2014).

Two groups of brain entropy maps were first tested using a two‐sample t test analysis with AlphaSim correction (p < .05) using the REST software (Song et al., 2011); then, the brain areas that exhibited significant differences and their corresponding MNI coordinates were obtained by the wfu_pickatlas toolbox (Maldjian, Laurienti, Burdette, & Kraft, 2003; Maldjian, Laurienti, & Burdette, 2004). Finally, the results were shown by MRIcro (http://www.cabi.gatech.edu/mricro/mricro/).

3. RESULTS AND ANALYSIS

The negatively activated brain areas presented in Figure 1 (seafarers’ entropy value < nonseafarers’ entropy value) indicated the regions where the brain entropy values of the seafarers were lower than that of the control group. This activity area was mostly concentrated in parts of the cerebellum, especially the AAL brain regions of Cerebelum_9_L and Cerebelum_8_L (Tzourio‐Mazoyer et al., 2002), and the corresponding MNI coordinates and involved brain regions are shown in Table 1.

Figure 1.

Negatively activated brain areas (i.e., seafarers’ entropy value < nonseafarers’ entropy value) identified by two‐sample t test with AlphaSim correction (p < .05) [Color figure can be viewed at http://wileyonlinelibrary.com]

Table 1.

The MNI coordinates and brain atlas of the negatively activated brain areas (seafarers’ entropy value < nonseafarers’ entropy value)

Area	MNI (x, y, z)	AAL atlas
Cerebellar tonsil	(−16, −54, −55)/(48, −58, −52)	Cerebelum_9_L/cerebelum_Crus2_R
Inferior semi‐lunar lobule	(−21, −67, −55)	Cerebelum_8_L

The positively activated brain regions presented in Figure 2 (seafarers’ entropy value > nonseafarers’ entropy value) indicated the regions where the brain entropy of the seafarers was higher than that of the control group. This activity was mostly concentrated in some portions of the superior temporal gyrus and the orbital frontal gyrus. The corresponding MNI coordinates and the altas of the involved brain regions are shown in Table 2.

Figure 2.

Positively activated brain areas (i.e., seafarers’ entropy value > nonseafarers’ entropy value) identified by two‐sample t test with AlphaSim correction (p < .05) [Color figure can be viewed at http://wileyonlinelibrary.com]

Table 2.

The MNI coordinates and brain atlas of the positively activated brain areas (seafarers’ entropy value > nonseafarers’ entropy value)

Area	MNI (x, y, z)	AAL & Brodmann
Frontal lobe	(−4, 22, −24)/(13, 22, −22)	Rectus_L/Frontal_Sup_Orb_brodmann area 11/brodmann area 47
Superior temporal lobe	(−34, 12, −22)/(38, 13, −22)	Temporal_Pole_Sup_L/Temporal_Pole_Sup_R brodmann area 38

4. DISCUSSION

The aim of this study was to explore the relationship between the occupational factors and the brain complexity using a resting‐state fMRI technique, taking seafarers as an example. The low entropy in the human brain has long been postulated and observed (Bergström, 1969; Singer, 2009; Wang et al., 2014). Furthermore, as stated by Rakic (2009), the relatively low brain entropy in the neocortex may reflect the “higher” mental functions subserved by the cortex (Wang et al., 2014). Thus, in this study, we postulate that the cortical regions with relatively low entropy correspond to the seafarers who could have better psychological or physiological functions compared to that of the nonseafarers and vice versa.

On one hand, according to the presented results in Figure 1 and Table 1, we found that the brain entropy in the cerebellum location of the seafarers was lower than that of the nonseafarers, revealing that the seafarers require more specialized cerebellar functions due to their occupation, compared to the nonseafarers, especially in the Cerebelum_9_L and Cerebelum_8_L regions. In humans, the cerebellum plays an important role in motion control and may be involved in cognitive functions such as attention and language (Desmond and Fiez, 1998; Fox et al., 2005; Leiner, Leiner, & Dow, 1993; De Smet, Paquier, Verhoeven, & Mariën, 2013), and its movement‐related function is most firmly established. As to the seafarers with the occupational training and long‐term offshore operations, the cerebellum is an important part of the vestibular system, and it helps the sailors to maintain their body balance and clear vision (Morton and Bastian, 2004; Stein and Glickstein, 1992); the cerebellum also improves their ability to determine their own position in three‐dimensional spaces within the maritime environment (Angelaki and Cullen, 2008; Highstein, Fay, & Popper, 2004; Wang, 2016; Wang et al., 2017). Meanwhile, according to the guidelines of STCW’95 for seafarers, the professional training courses for sailors touch upon safety training courses, the manipulation of large‐size ships, the training for ship collision prevention, and so on, which also specifically reshape the cerebellum's function to better satisfy the occupational requirements (Horck, 2006). Thus, this complexity of the changes in the cerebellum location of the sailors is possibly related to their maritime skills, compared to controls with land‐based jobs.

On the other hand, according to the results presented in Figure 2 and Table 2, we found that the brain entropy in the superior temporal gyrus (STG) (i.e., Temporal_Pole_Sup_L/Temporal_Pole_Sup_R) and some portions of the frontal lobe (Rectus_L/Frontal_Sup_Orb) of the seafarer were higher than the entropy of the non‐seafarer in those regions. The superior temporal gyrus is an essential structure involved in auditory processing, as well as the function of hearing, speech, language (Bigler et al., 2007; Howard et al., 2000; Tang, Hamilton, & Chang, 2017), and integration of multimodal sensory input (Barraclough, Xiao, Baker, Oram, & Perrett, 2005). Thus, we inferred that the increase in brain entropy in the superior temporal gyrus of the seafarer was possibly due to the marine working environment, which has continuous machine noise, the monotonous loud sound of surf, and the less communicative community of single‐sex colleagues (Wang, 2016; Wang et al., 2017). Furthermore, previous studies have demonstrated that the orbito‐frontal gyrus plays a key role in the regulation of the peripheral physiological response during emotional experiences (Barbas, 2007; Barrett, Mesquita, Ochsner, & Gross, 2007; Bechara, Damasio, & Damasio, 2000; Bishop, Duncan, Brett, & Lawrence, 2004; Kim, Gee, Loucks, Davis, & Whalen, 2011; Meyers, Berman, Scheibel, & Hayman, 1992; Ohira et al., 2006) and in the planning behaviors associated with reward and punishment (Bechara, Damasio, Damasio, & Anderson, 1994; Wei et al., 2011). The sailors were required to work on the sea and to live in relatively small and closed spaces for long periods of time, far from the land and their families. This could cause some adverse effects on the function of the orbito‐frontal gyrus affecting the control of emotion. Thus, we inferred that the increase in the entropy in some portions in superior frontal gyrus of the seafarer possibly reflected their decreasing ability to control their emotions in contrast with the nonseafarer. This may suggest an anxiety‐prone trait that results from the long‐term experience of a boring and lonely life on the sea. Also, according to the hierarchical structure and strict accountability system in sailor management, it might be inferred that the changes of complexity captured by brain entropy in OFG region with respect to seafarers have close relation to the changes of regulative ability with respect to rewards (occupational promotion and honor) and punishment (bearing the responsibility of the accident) (Bechara et al., 1994; Wei et al., 2011). Besides, we also found the increase in the brain entropy in the inferior frontal gyrus (Brodmann Areas 11 and 47) of the seafarers in contrast to the nonseafarers, whose function was closely associated with the processing of speech and sign language (Cooper and Paccia‐Cooper, 1980). It was also reported that the fractional anisotropy values of uncinate fasciculus (including Brodmann Area 11) were decreased significantly in children who experienced socioemotional deprivation (Eluvathingal et al., 2006). Thus, our finding of the increase in entropy of this region with respect to the seafarers might partly contribute to the decreased control capability of socioemotion in sailors who had long periods of sailing experience, due to the less communicative community and the situation far from the their families on the sea.

In this research, we investigated how long‐term career experience reshaped the functional complexity of the human brain. All the sailors had sailing experience durations of greater than or equal to 10 years. However, when they were recruited, they answered whether or not they had the requisite number of years of experience by a verbal “Yes” or “No.” Thus, the actual number of years of sailing experience that each sailor had was not recorded. This is a limitation in the exploration of how the duration of seafaring experience affects brain entropy. Additionally, the number of subjects in this study was relatively small, potentially affecting the reproducibility of the results, although the used SampEn portrays the well discriminatory capability for the small‐size group analysis (Sokunbi, 2014). Thus, the reported complexity of the changes in the superior temporal gyrus, frontal gyrus, and cerebellum whether depended on the duration of the sailor's experience, and the reproducibility of this altered complexity should be further investigated in large‐size samples in a follow‐up study.

5. CONCLUSION

In this study, we have explored the relationship between occupational factors and brain functional plasticity using brain entropy calculated from the resting‐state fMRI scans. By performing statistical analysis on the brain entropy maps from the seafarer and nonseafarer groups, we found that in contrast to the nonseafarers, the seafarers had lower entropy values in the cerebellum, which contributed to their professional maritime skills such as maintaining their body balance and clear vision, and improving their ability to determine their own position in three‐dimensional spaces in the maritime environment. We further found that in comparison to the nonseafarers, the seafarers had higher entropy values in some portions of the superior temporal gyrus, rectus gyrus, and superior frontal gyrus; these values possibly reflected the relatively decreased capability for emotional control and auditory information processing. We posited that this may be the case due to the specific marine working environment with continuous machine noise, the monotonous loud sound of surf, and the less communicative single‐sex communities and the long periods of isolation from their families. In summary, this study provides new insight into how occupational factors change the brain's complexity, and it further verifies the brain's strongly functional plasticity.

Wang N, Wu H, Xu M, et al. Occupational functional plasticity revealed by brain entropy: A resting‐state fMRI study of seafarers. Hum Brain Mapp. 2018;39:2997–3004. 10.1002/hbm.24055