Skip to main content
NCBI home page
Human Brain Mapping logoLink to Human Brain Mapping
. 2018 May 1;39(9):3636–3651. doi: 10.1002/hbm.24200

Estimation of vocational aptitudes using functional brain networks

Yul‐Wan Sung 1,✉, Yousuke Kawachi 1, Uk‐Su Choi 2, Daehun Kang 1, Chihiro Abe 1, Yuki Otomo 1, Seiji Ogawa 1
PMCID: PMC6866321  PMID: 29717529

Abstract

The success of human life in modern society is highly dependent on occupation. Therefore, it is very important for people to identify and develop a career plan that best suits their aptitude. Traditional test batteries for vocational aptitudes are not oriented to measure developmental changes in job suitability because repeated measurements can introduce bias as the content of the test batteries is learned. In this study, we attempted to objectively assess vocational aptitudes by measuring functional brain networks and identified functional brain networks that intrinsically represented vocational aptitudes for 19 job divisions in a General Aptitude Test Battery. In addition, we derived classifiers based on these networks to predict the aptitudes of our test participants for each job division. Our results suggest that the measurement of brain function can indeed yield an objective evaluation of vocational aptitudes; this technique will enable a person to follow changes in one's job suitability with additional training or learning, paving a new way to advise people on career development.

Keywords: brain function networks, vocational aptitudes, classifier, resting‐state fMRI, multiclass, support vector machine

1. INTRODUCTION

In modern society, the success of an individual is highly dependent on their occupation. Therefore, choosing and developing a career that takes advantage of one's aptitudes is of utmost importance.

The suitability for a job is typically assessed on the basis of psychometric properties, such as intellectual, clerical, perceptual, numerical, spatial performance, and so forth, and there is a battery of tests available to measure these abilities (Gottfredson, 1998; Gottfredson, 2003; Frey & Detterman, 2004). For example, the Minnesota Clerical Test (Andrew, Paterson, & Longstaff, 1979), the Differential Aptitude Test (Bennett, Seashore, & Wesman, 1974), the Wonderlic tests (Wonderlic, 1983), the General Aptitude Test Battery (GATB) (United States Department of Labor, 1970), and the Armed Services Vocational Aptitude Battery (United States Department of Defense, 1984) are used to evaluate vocational aptitude. These tests have been widely used for career guidance, selection, counseling, and so on (Hunter, 1986; Barrick, Mount, & Strauss, 1993; Lent, Brown, & Hackett, 1994; Legree, 1995; Kwallek, Lewis, Lin‐Hsiao, & Woodson, 1996; Brown & Lent, 1996; Chan, 1997). However, these vocational aptitude tests are not oriented to evaluate the properties that affect suitability over time (i.e., progress or improvement of abilities), because repeating these tests can lead to biases in the results as participants become familiar with the questions.

The abilities evaluated in these tests are considered to be influenced by genetics, experiences, and environment (McGee, 1979; Plomin, Owen, & McGuffin, 1994; McClearn et al., 1997; Thompson et al., 2001; Turkheimer, Haley, Waldron, D'Onofrio, & Gottesman, 2003; Hackman & Farah, 2009), and these characteristics must be stored and represented in the brain. Therefore, we hypothesized that the vocational aptitudes measured by these tests could also be represented in the brain as intrinsic brain functions associated with job requirements. Conversely, by measuring brain function, we may be able to assess specific vocational aptitudes; hence, it will become possible to take repeated measurements, allowing us to evaluate changes in vocational aptitudes or to perform follow‐up studies.

Only a few functional brain studies related to vocational aptitude have been reported, and these studies do not directly examine how specific vocational aptitudes are represented in the brain (Yurgelun‐Todd, Killgore, & Cintron, 2003; Fan et al., 2013; Burgaleta et al., 2014; Jung et al., 2015). Technical difficulties in measuring brain function that specifically reflect vocational aptitudes may be the reason of this limitation.

Recently, numerous studies have shown that the effects of cognitive and physical training are reflected in morphological and functional brain changes (Golestani, Paus, & Zatorre, 2002; Dong & Greenough, 2004; Bermudez & Zatorre, 2005; Draganski et al., 2006; Maguire, Woollett, & Spiers, 2006; Pereira et al., 2007; Deng, Saxe, Gallina, & Gage, 2009; Voss & Schiff, 2009; Aimone, Wiles, & Gage, 2009; Tremblay, Lowery, & Majewska, 2010; Rhyu et al., 2010; Dosenbach et al., 2010; Teipel et al., 2010; Decety & Michalska, 2010; May, 2011), indicating the role of environment in neuroplasticity. Therefore, it is likely that changes in vocational aptitudes caused by the environment would also be reflected as functional and/or morphological changes in the brain and that measuring these changes would allow us to assess vocational aptitudes.

In this study, we attempted to assess vocational aptitudes by measuring brain function. We measured resting‐state brain function using functional MRI and evaluated vocational aptitudes using the Japanese version of GATB (Employment Security Bureau, Ministry of Health, Labour & Welfare, 2013). To identify functional networks from the resting‐state fMRI signals, we performed a region of interest (ROI)‐based analysis, in which ROIs defined on the basis of previous studies (Desikan et al., 2006; Dosenbach et al., 2010) were used. Based on these functional networks, classifiers for vocational aptitude for 19 job divisions in GATB (Table 1) were derived using a multi‐class (three levels) support vector machine (SVM).

Table 1.

Vocational aptitudes: 19 divisions

Job 1 (M) Plant culture/animal breeding
Job 2 (GV) Animal training and service/aquaculture/gardening
Job 3 (KM) Physical works in mechanical/plants/animals
Job 4 (PM) Hand work
Job 5 (KFM) Industrial machine manipulation
Job 6 (PKM) Manufacture/assembly
Job 7 (SPM) Cutting/molding/construction/installation
Job 8 (SPF) Barber/beauty services
Job 9 (NSP) Drawing and architecture design/laboratory technology
Job 10 (GQM) Nursing/therapy/specialized teaching services/quality control/security management
Job 11 (GNS) Engineering technology/physical sciences/medical sciences
Job 12 (GVN) Social research/law/social service/mathematics and statistics
Job 13 (GVQ) General office work/educational and library services
Job 14 (SP) Design and visual arts
Job 15 (GQK) Clerical machine operation/communicator
Job 16 (GM) Nursing‐care services
Job 17 (GNQ) General sales/accounting
Job 18 (GQ) Hospitality/security services
Job 19 (QK) Clerical handling
Open in a new tab

Note. Abbreviations: F = finger dexterity; G = general intelligence; K = motor coordination; M = manual dexterity; N = numerical aptitude; P = form perception; Q = clerical perception; S = spatial aptitude; V = verbal aptitude.

2. METHODS

2.1. Participants

Two groups of participants were included in this study. The first group (for the training of classifiers) included 112 university students in their freshman and sophomore years (28 males, 84 females; mean age + standard deviation (SD): 20.14 ± 0.69 years). The second group (for the testing of classifiers) included 35 students also in their freshmen and sophomore years from the same university as the first group (5 males, 30 females; mean age + SD: 20.15 ± 0.75 years). The volunteers in these two groups participated in both the psychometric and MRI experiments. None of the participants had a history of neurological disease or any other medical conditions (e.g., pregnancy, arrhythmia, or claustrophobia). The participants were given a complete description of the study, and written informed consent was obtained in accordance with the Declaration of Helsinki. This study was approved by the Institutional Review Board of Tohoku Fukushi University in Japan.

2.2. Psychometric measurements of vocational aptitudes

GATB was created by the US government based on a vast amount of occupational analyses (United States Department of Labor, 1970). In the analysis, the government classified more than 75,000 jobs into 20 job groups and then found nine aptitudes needed for each job group. In Japan, the Ministry of Health, Labour and Welfare created the Japanese version of GATB and has updated it over the years; it is still widely used today. The Japanese version of GATB, in the present day, has been recognized as having higher reliability and validity than any other test battery and has been through five revisions to track the development of Japanese lifestyle and culture (Muroyama, 2016).

The Japanese version of GATB classifies people into 40 subdivisions of jobs based on weighted combinations of the scores of the nine abilities. Additionally, 19 independent job divisions, which are larger categories than the 40 subdivisions, are derived from a linear combination of the nine ability scores without weighting and are linearly divided from low to high levels with even intervals of scores such as low, middle, and high, according to GATB (Employment Security Bureau, Ministry of Health, Labour & Welfare, 2013). The process of grouping into the 40 subdivisions is nonlinear but grouping into the 19 divisions is linear. Therefore, we attempted to evaluate the 19 job divisions, because one of the purposes of this study is to directly provide information about the vocational aptitudes of individuals rather than just to provide information related to the nine abilities. The two student groups participated in psychometric measurements of vocational aptitudes for the 19 job divisions. The job divisions are listed in Table 1.

In the measurements of vocational aptitudes for the job divisions, there were no participants who reported any pre‐experience or pre‐knowledge about GATB until they underwent the testing. The measurements were administrated in the testing environment as follows: in a quiet, distraction‐free, comfortable room and by well‐trained testers according to the instructions (Shokugyo Tekisei Kenkyukai, 2011). Our testers cautiously checked whether participants were in a comfortable state before starting each subtest of GATB. All testers had experienced various other psychological tests before the GATB measurements. The measurements were mostly performed after MRI measurements, and the mean interval from the fMRI measurements to GATB was 3.27 months (SD = 1.67). For three of the participants, GATB was conducted before the MRI measurements.

2.3. MRI measurements

We used a 3‐Tesla MRI scanner (Skyra‐fit; Siemens Co., Erlangen, Germany), and all participants were scanned in two sessions to obtain structural (T1) and functional image measurements (resting‐state fMRI). Sagittal structural images were acquired using the following parameters: repetition time = 1,900 ms, echo time = 2.52 ms, matrix size = 256 × 256, in‐plane resolution = 1 × 1 mm2, slice thickness = 1 mm, and number of slices = 192. Resting‐state fMRI data were acquired using the following parameters: repetition time = 1,000 ms, echo time = 24 ms, matrix size = 64 × 64, in‐plane resolution = 3.4 × 3.4 mm2, slice thickness = 3.4 mm, and number of volumes = 480. During the resting‐state fMRI session, participants were asked to lie on the bed with their eyes open, to lightly focus on the center of their visual field, and to not think about anything as much as possible. The lights in the room were turned off for all MRI scans.

2.4. Procedure for MRI data processing

The analysis of MRI data for the two groups was performed according to the processing pipeline described below.

2.5. Preparation of ROIs for functional networks

We used ROIs based on the two templates from Dosenbach and Harvard (Desikan et al., 2006; Dosenbach et al., 2010) that were reported in the DPABI software (see below) (Chao‐Gan & Yu‐Feng, 2010; Yan, Wang, Zuo, & Zang, 2016) and defined brain areas of 272 ROIs (Supporting Information, Table 1) by removing redundant ROIs from the two templates. Each ROI was a sphere with a radius of 5 mm around the center coordinate. We used the Dosenbach template first and then added ROIs from the Harvard template, but only those that did not overlap with the Dosenbach ROIs. Originally, the Dosenbach template consisted of 160 ROIs, and the other 112 ROIs were added from the Harvard template to make our 272 ROIs.

2.6. Preprocessing of fMRI data

To preprocess the resting‐state fMRI data for the 112 training participants, we used the data processing assistant for a part of resting‐state fMRI preprocessing software known as DPABI (Chao‐Gan & Yu‐Feng, 2010; Yan et al., 2016). The preprocessing included slice‐scan time correction, 3D motion correction (maximum head motion: 111 participants passed the maximum threshold of 1.5 mm and 1.5°, and the remaining participant passed the maximum threshold of 3.0 mm and 3°), band‐pass temporal filtering (between 0.01 and 0.1 Hz), and artifact rejection based on the CSF signal. To control for head motion confounds, the Friston 24‐parameter model was used to regress out head motion effects. These functional images were co‐registered with each corresponding structural image.

2.7. Identification of functional networks

To identify functional networks for the job divisions, we performed a simple regression analysis with the vocational aptitude scores measured by GATB for each job division and the resting‐state fMRI signals from the 272 predefined ROIs of the 112 training participants using network‐based statistics (NBS) for multiple comparison correction (Zalesky, Fornito, & Bullmore, 2010) for each of the 19 job divisions (Figure 1a) as follows. (c) For the 272 time courses obtained from the 272 ROIs, the correlation was calculated between ROIs for each participant, which produced a coefficient matrix (272 × 272); (b) the process was performed for all 112 training participants, through which 112 correlation coefficient matrices (272 × 272) were created; (c) a simple regression analysis was performed with the 112 correlation matrices and a vector of vocational aptitude scores for the 112 training participants for each job division to identify functional brain networks in which randomly shuffled correlation matrices were generated for a permutation test, and then the regression analysis on the true correlation matrix with the vocational aptitude vector and a nonparametric permutation test on the random correlation matrices with more than 5,000 iterations were performed (Figure 1a); (4) NBS were used for multiple comparisons. For all 19 job divisions, the processing steps (a)–(d) were repeated. Correction for multiple comparisons based on NBS statistics was performed at an initial threshold of p < .005, and the number of edges in the network was limited to as close to 15 as possible, because a previous study has shown that a small number of brain connections can predict a psychiatric disorder (Yahata et al., 2016).

Figure 1.

Figure 1

Open in a new tab

MRI data processing workflow. (a) Diagrams to illustrate the data flow for the primary data processing; identification of functional networks and derivation of multi‐class classifiers (MCCs) using multi‐class support vector machine (SVM). Step 1: resting‐state functional MRI (rs‐fMRI) signals are extracted from the predefined 272 ROIs, and functional connectivity (FC) matrices (correlation matrices) are constructed for the 112 training participants. Step 2: a simple regression and network‐based statistics (NBS) analysis are performed for the FC matrices of the 112 participants, and functional brain networks (BNs) reflecting 19 job divisions are identified. Step 3: an input dataset (MRI data) to the MCCs—SVM classifiers—is constructed from the edges of the identified functional BNs in addition to another input dataset (vocational aptitude levels), which is constructed from the General Aptitude Test Battery (GATB). Step 4: the MCCs are trained by the input datasets. Step 5: the vocational aptitude (VA) levels of the job divisions estimated by the classifiers are compared with the VA levels determined by GATB to calculate accuracy. (b) Verification of MCCs for the test group: after making FC matrices for the 35 test participants, Steps 4 and 5 were carried out (see panel a). [Color figure can be viewed at http://wileyonlinelibrary.com]

2.8. Multiple SVM classifier design (training)

We attempted to design three‐class multiple SVM classifiers to evaluate the vocational aptitudes, because the Japanese GATB classifies the job divisions into three classes as described in the previous section (Figure 1a). Datasets for the design were prepared from the results of the identified brain network analysis for the 112 training participants and the vocational aptitudes scores from GATB, which consisted of a data matrix with features and a label matrix for each of the 19 job divisions. Data matrices were composed of correlation coefficients (features) of selected network edges. Label matrices were created based on the scores of the corresponding vocational aptitude of GATB, and the vocational aptitude scores were normalized and divided into three classes (levels) according to the GATB classification: “low,” “middle,” and “high,” for which the GATB scores were divided into evenly with the same size intervals as those of the scores of vocational aptitudes from the minimum to the maximum; that is, the classes were defined using the equation {(x − min)/(max − min)} × 3, where x was the score of each vocational aptitude. In the calculation, the number after the decimal point was rounded up. With the datasets of network edges and all vocational aptitude scores grouped into three classes, multiclass (three classes) classifiers were designed using the “classification learner app” in MATLAB 2017b (Mathworks Co., USA), where a one‐versus‐one design method was used with 10‐fold cross‐validation, and the predictors were standardized in such a way that the software centered and scaled each column of the predictor data (network edges) by the weighted column mean and standard deviation. All support vector machine models for a job division were designed (using the “all SVM” option), among which an SVM model that showed the best performance was chosen as the classifier for the job division. This process was performed for all 19 job divisions.

2.9. Verification of SVM classifiers (test)

The MRI data for the second student group of 35 participants were processed through the same pipeline as the data from the first student group except for the identification of functional networks, and the vocational aptitude levels of the 19 job divisions were also estimated for the second student group (Figure 1b).

3. RESULTS

The measured vocational aptitude scores for the 19 job divisions in GATB for the 112 training participants were evaluated for normality, and the scores were revealed to have come from a normal distribution for all 19 job divisions (i.e., the hypothesis that aptitude scores of each job division came from other distributions than from a normal distribution was rejected with the minimum p = .22 and the maximum p = .93, chi‐square test: goodness‐of‐fit). The histograms are summarized in Supporting Information, Figure 2. We compared the mean values of the vocational aptitude scores from our training sample (112 participants) with those of the original (true) population of Japan (Employment Security Bureau, Ministry of Health, Labour and Welfare, 2013) to determine how they differed for each of the 19 vocational aptitudes. A two‐sample t test was performed using mean values and standard deviations of the populations (Figure 2a,b), and six and eleven vocational aptitudes were found to be significantly different (p = .01 and .05, respectively). Comparison of the standard deviations of the 112‐participant training sample and the original (true) population of Japan showed that the standard deviation of the sample data was more than 50% of the original population, with a maximum of 85.2% and a minimum of 50.7% (Figure 2c). For the 35‐participant test sample, 17 job divisions revealed significant normal distribution (i.e., the hypothesis that the scores came from other distributions than from a normal distribution was rejected, with minimum p = .09 and the maximum p = .89; chi‐square test: goodness‐of‐fit), and two job divisions were revealed not to have a normal distribution (p = .03). The histograms are summarized in Supporting Information, Figure 3.

Figure 2.

Figure 2

Open in a new tab

Data distribution of GABT scores as the mean value and standard deviation of vocational aptitude scores of 19 job divisions for the 112‐participant training sample (a) and for the original (true) population of Japan (b), and the ratio of standard deviations of the two populations (c). [Color figure can be viewed at http://wileyonlinelibrary.com]

A simple regression analysis of the resting‐state connectivity and vocational aptitude scores measured by GATB gave 19 functional brain networks (p = .005, corrected by NBS); each network corresponded to one job division out of the 19 total GATB job divisions (Figure 3). The average number of edges of the networks was 11 ± 3.97. Each network had a different configuration from the others and showed that specific brain areas that were distinct in the network configuration were contributing to the aptitude required for that job division (Table 1 and Figure 3). Overall, 121 brain areas were involved in the 19 functional networks (Table 2 and Figure 3).

Figure 3.

Figure 3

Open in a new tab

Functional networks identified for the 19 job divisions. Each functional network corresponds to a job division. The numbers assigned to each node (ROIs) of the networks appear in Table 2. (a) Functional networks for jobs 1, 2, 3, and 4. (b) Functional networks for jobs 5, 6, 7, and 8. (c) Functional networks for jobs 9, 10, 11, and 12. (d) Functional networks for jobs 13, 14, 15, and 16. (e) Functional networks for jobs 17, 18, and 19. [Color figure can be viewed at http://wileyonlinelibrary.com]

Table 2.

Brain areas involved in the functional networks associated with the job divisions shown in Figure 3 (121 regions of interest and the coordinates of the regions in MNI coordinates)

MNI coordinates
# ROI numbers in 19 networks ROI‐belonging job network Brain region Brodmann x y z
1 1 Job04,10 vmPFC_R Brodmann area 10 6 64 3
2 2 Job02,11,13,15 mPFC Brodmann area 6 29 57 18
3 5 Job06,14 vmPFC_L Brodmann area 32 −25 51 27
4 6 Job10,17 vmPFC_L Brodmann area 9 9 51 16
5 8 Job04,07,09 ACC_R Brodmann area 32 27 49 26
6 13 Job15 Inf temporal_L Brodmann area 21 8 42 −5
7 14 Job03,15 Post cingulate_R Brodmann area 23 9 39 20
8 15 Job10,17,18 Fusiform_R Brodmann area 36 46 39 −15
9 20 Job01,10 Post cingulate_L Brodmann area 23 −16 29 54
10 26 Job02,03 Post cingulate_L Brodmann area 30 38 21 −1
11 27 Job05 Angular gyrus_R Brodmann area 39 9 20 34
12 28 Job02,05,06 Angular gyrus_L Brodmann area 39 −36 18 2
13 29 Job10 Precuneus_R Brodmann area 7 40 17 40
14 30 Job05 IPS_L Brodmann area 19 −6 17 34
15 32 Job05 Occipital_R Brodmann area 39 58 11 14
16 34 Job06 Occipital_L Brodmann area 39 44 8 34
17 35 Job05 aPFC_R Brodmann area 10 60 8 34
18 38 Job15 Vent aPFC_L Brodmann area 46 −20 6 7
19 41 Job06 ACC_L Brodmann area 32 10 5 51
20 43 Job02,05,06 vPFC_L Brodmann area 45 0 −1 52
21 49 Job05 IPL_R Brodmann area 40 −44 −6 49
22 50 Job11 Post parietal_L Brodmann area 40 −26 −8 54
23 52 Job10 IPL_L Brodmann area 40 −54 −9 23
24 57 Job05 vPFC_R Brodmann area 13 −12 −12 6
25 58 Job05 ACC_L Brodmann area 32 11 −12 6
26 70 Job07,08,09 Thalamus_L Caudate Body 42 −24 17
27 76 Job08 Post insula_L Brodmann area 13 −30 −28 9
28 77 Job07 Temporal_R Brodmann area 41 −24 −30 64
29 78 Job11 Post cingulate_L Pulvinar 51 −30 5
30 79 Job19 Fusiform_R Brodmann area 20 −41 −31 48
31 81 Job14 Parietal_R Brodmann area 13 54 −31 −18
32 83 Job19 Parietal_L Brodmann area 40 −53 −37 13
33 86 Job19 Temporal_L Brodmann area 21 34 −39 65
34 87 Job12 TPJ_L Brodmann area 39 8 −40 50
35 88 Job11,19 Frontal_R Brodmann area 44 −41 −40 42
36 89 Job02,03,04,07,09,14 dFC_R Brodmann area 6 58 −41 20
37 90 Job03 vFC_L Brodmann area 9 −8 −41 3
38 96 Job10,17,18 Mid insula_L Brodmann area 13 54 −44 43
39 98 Job04,07,11,16 Parietal_L Brodmann area 6 −28 −44 −25
40 100 Job02,11,13 Precentral gyrus_L Brodmann area 43 42 −46 21
41 101 Job17 Precentral gyrus_R Brodmann area 4 −48 −47 49
42 103 Job04,07,08,12,14,16 Mid insula_R * −59 −47 11
43 104 Job10,17,18 Mid insula_L Brodmann area 13 −53 −50 39
44 105 Job02,11,13,14,15 Temporal_R Brodmann area 41 5 −50 33
45 107 Job11 Parietal_L Brodmann area 2 44 −52 47
46 109 Job04,07,08,09,12,14,16 Parietal_L Brodmann area 2 −24 −54 −21
47 112 Job13 Parietal_R Brodmann area 3 −6 −56 29
48 113 Job07 Post insula_R Brodmann area 13 −34 −57 −24
49 115 Job10 Parietal_L Brodmann area 3 −11 −58 17
50 116 Job03 Parietal_L Brodmann area 40 32 −59 41
51 117 Job04,10,17 Post parietal_L Brodmann area 40 51 −59 34
52 119 Job04 Temporal_L Brodmann area 41 36 −60 −8
53 123 Job09,12 Occipital_R * 46 −62 5
54 124 Job17,18 Temporal_R Brodmann area 37 −48 −63 35
55 129 Job12 Occipital_R Brodmann area 31 19 −66 −1
56 130 Job05,08 Occipital_L Brodmann area 31 1 −66 −24
57 131 Job10,17,18 Occipital_L Brodmann area 18 −34 −67 −29
58 132 Job03 Occipital_R Brodmann area 23 11 −68 42
59 133 Job12,16 Occipital_R Brodmann area 18 17 −68 20
60 135 Job02,04,07 Post occipital_L Brodmann area 18 39 −71 13
61 136 Job03 Post occipital_R Brodmann area 17 −9 −72 41
62 137 Job04 Post occipital_R Brodmann area 18 45 −72 29
63 138 Job07 Post occipital_L Brodmann area 19 −11 −72 −14
64 139 Job11,16 Post occipital Brodmann area 18 29 −73 29
65 142 Job03 Post occipital Brodmann area 17 −29 −75 28
66 143 Job04,14 Lat cerebellum_L * 5 −75 −11
67 146 Job01,06,10 Lat cerebellum_L * −42 −76 26
68 152 Job12,16 Med cerebellum_R * −5 −80 9
69 153 Job04,07,08,09,16 Inf cerebellum_L * 29 −81 14
70 156 Job04,16 Med cerebellum_R * −37 −83 −2
71 157 Job04 Med cerebellum_R * −29 −88 8
72 160 Job07,09 Inf cerebellum_R * −4 −94 12
73 164 Insular cortex_R 38 3 0
74 165 Job01,03,04,10 Superior frontal gyrus_L −14 19 57
75 167 Job02,11,13,15 Middle frontal gyrus_L −38 19 42
76 172 Job06 Inferior frontal gyrus, pars opercularis_R 52 15 16
77 176 Job08,09,12,14 Temporal pole_R 41 13 −30
78 177 Job11 Superior temporal gyrus, anterior division_L −56 −4 −8
79 179 Job19 Superior temporal gyrus, posterior division_L −61 −27 2
80 180 Job08,09 Superior temporal gyrus, posterior division_R 60 −23 2
81 181 Job17,18 Middle temporal gyrus, anterior division_L −58 −4 −22
82 184 Job03 Middle temporal gyrus, posterior division_R 61 −22 −13
83 186 Job06 Middle temporal gyrus, temporooccipital part_R 58 −49 2
84 187 Job10 Inferior temporal gyrus, anterior division_L −48 −5 −39
85 189 Job04,07,15 Inferior temporal gyrus, posterior division_L −54 −29 −26
86 192 Job19 Inferior temporal gyrus, temporooccipital part_R 54 −50 −17
87 196 Job16,19 Superior parietal lobule_R 29 −48 59
88 197 Job08,11,13 Supramarginal gyrus, anterior division_L −57 −33 37
89 198 Job11 Supramarginal gyrus, anterior division_R 59 −27 38
90 201 Job07,13,14 Angular gyrus_L −50 −56 30
91 202 Job04,10 Angular gyrus_R 52 −52 32
92 203 Job08 Lateral occipital cortex, superior division_L −32 −73 38
93 205 Job11 Lateral occipital cortex, inferior division_L −45 −76 −2
94 206 Job09,16 Lateral occipital cortex, inferior division_R 46 −74 −2
95 207 Job12,14 Intracalcarine cortex_L −10 −75 8
96 211 Job05,06 Juxtapositional lobule cortex (formerly supplementary motor cortex)_L −6 −3 56
97 212 Job05,06 Juxtapositional lobule cortex (formerly supplementary motor cortex)_R 6 −3 58
98 215 Job15 Paracingulate gyrus_L −6 37 21
99 217 Job02 Cingulate gyrus, anterior division_L −5 18 25
100 221 Job13,15 Precuneous cortex_L −8 −60 37
101 222 Job02,11,13,15 Precuneous cortex_R 9 −58 39
102 223 Job04,12,16 Cuneal cortex_L −8 −80 27
103 226 Job15 Frontal orbital cortex_R 29 23 −16
104 227 Job14 Parahippocampal gyrus, anterior division_L −22 −9 −30
105 228 Job03 Parahippocampal gyrus, anterior division_R 22 −8 −30
106 230 Jon07 Parahippocampal gyrus, posterior division_R 23 −30 −17
107 231 Job16 Lingual gyrus_L −13 −65 −5
108 236 Job03 Temporal fusiform cortex, posterior division_R 36 −24 −28
109 237 Job07 Temporal occipital fusiform cortex_L −34 −54 −16
110 241 Job02,05 Frontal operculum cortex_L −40 18 5
111 242 Job02,06 Frontal operculum cortex_R 41 19 5
112 243 Job05 Central opercular cortex_L −48 −8 12
113 244 Job05 Central opercular cortex_R 49 −6 11
114 246 Job19 Parietal operculum cortex_R 49 −27 21
115 247 Job07,08,14 Planum polare_L −47 −6 −7
116 251 Job19 Planum temporale_L −53 −30 11
117 253 Job12 Supracalcarine cortex_L −10 −72 15
118 257 Job07 Brain stem −7 −31 −34
119 258 Job14 Brain stem 8 −31 −34
120 261 Job03 Caudate −13 9 10
121 270 Job14 Amygdala 　 23 −4 −18
Open in a new tab

Classifiers corresponding to each of the functional networks were designed using multi‐class (three classes: high, middle, and low) SVM, and the training accuracies of the classifiers for all the 19 job divisions were greater than chance (defined as 33.3%), as verified by cross‐validation (Figure 4). The average accuracy of the 19 classifiers was 58.0% ± 8.6%, with a maximum of 70.5% and a minimum of 42.9%. The classifiers were then tested by applying them to the 35‐participant test sample, whose data were not used for training the classifiers (multiclass SVM; see the Methods section). The prediction accuracy for the second student group was more than the level of chance for all classifiers in the 19 job divisions (Figure 5). The average accuracy of the 19 classifiers was 51.9% ± 8.1%, with a maximum accuracy of 68.5% and a minimum of 37.1%. The sensitivity and specificity of the classifiers were calculated and are summarized in Table 3. To evaluate the difference between the distribution of the GATB scores and that of the predicted scores of the 35‐participant test sample, we compared the two data score sets for each vocational aptitude, which revealed no significant differences in any of the 19 job divisions (minimum p = .17; maximum p = .99; two‐sample Kolmogorov–Smirnov test). For additional classifier performance evaluation, we examined the classes which the mispredicted classes of a classifier belonged to, that is, the opposite or the neighboring class. For the classifiers of eight job divisions, all mispredicted classes were found to belong to the neighboring class, and for the other 11 classifiers, <8.5% of the mispredicted classes belonged to the opposite classes (Table 4).

Figure 4.

Figure 4

Open in a new tab

Training accuracies of the 19 multiclass classifiers. The training accuracies using 112 participants were acquired by the 10‐fold cross‐validation method. Accuracies for all 19 classifiers were better than chance (defined here as 33.3%)

Figure 5.

Figure 5

Open in a new tab

Test accuracies of the classifiers. The classifiers were tested for 35 participants whose data were not used in the training. Accuracy was determined by comparing the VA levels estimated by the classifiers and those measured by GATB. Accuracies for all 19 classifiers were better than chance

Table 3.

Sensitivity (Sen) and specificity (Spe) of the classifiers of the 19 job divisions for the test group

Job divisions Sen Spe
Job01 35.10 69.30
Job02 35.20 63.20
Job03 32.00 65.60
Job04 32.00 65.50
Job05 32.10 65.70
Job06 30.00 65.20
Job07 34.30 71.40
Job08 42.00 75.00
Job09 33.30 66.70
Job10 41.00 75.30
Job11 33.50 67.30
Job12 32.40 63.30
Job13 37.70 74.20
Job14 31.00 64.40
Job15 35.50 73.40
Job16 60.70 76.40
Job17 31.40 67.00
Job18 28.90 70.60
Job19 41.20 71.10
Open in a new tab

Table 4.

Ratio of mispredicted classes that actually belonged to the opposite classes (Opp) [%]

Job divisions Opp
Job01 2.80
Job02 0.00
Job03 0.00
Job04 0.00
Job05 0.00
Job06 8.50
Job07 8.50
Job08 5.70
Job09 0.00
Job10 8.50
Job11 2.80
Job12 0.00
Job13 8.50
Job14 0.00
Job15 5.70
Job16 5.70
Job17 8.50
Job18 8.50
Job19 0.00
Open in a new tab

4. DISCUSSION

This study aimed to map brain functions that intrinsically represent vocational aptitudes and to determine the utility of this approach in assessing these aptitudes. We identified functional networks that reflected 19 distinct job divisions and used them to derive multiple SVM classifiers for each division. We validated that the SVM classifiers could significantly and accurately estimate the vocational aptitudes of the participants. The data indicated that functional networks intrinsically represent vocational aptitudes and that the measurement of these functional networks can assess these vocational aptitudes.

Measuring vocational aptitude in this manner can reveal what brain functions are involved in specific aptitudes, which may be used to develop methods to enhance an individual's career‐specific abilities. Furthermore, the confounding effects of learning during repetition of psychometric aptitude tests prohibit the use of such tests for the evaluation of aptitude development over the course of an individual's training. Measuring brain functionality can circumvent this learned testing bias and allow for longitudinal evaluation over time.

Previous studies examining psychological scores and brain function or structural changes have identified brain areas associated with the intelligence quotient (Haier, Jung, Yeo, Head, & Alkire, 2004; Colom, Jung, & Haier, 2006; Royle et al., 2013; MacDonald, Ganjavi, Collins, Evans, & Karama, 2014). The assessment of vocational aptitudes by intelligence alone is limited because although it reflects a general cognitive ability, it only correlates with certain aspects of vocational abilities (Gottfredson, 2003). A previous study examined brain morphology and vocational aptitude by comparing subcortical brain regions and psychological test scores. However, it did not completely investigate all the traits required to assess aptitude (Burgaleta et al., 2014). To the best of our knowledge, this study is the first to use brain function to assess specific vocational aptitudes.

Task‐based fMRI and resting‐state fMRI studies have indicated that changes in functional networks are more sensitive than structural changes alone (Choi, Sung, Hong, Chung, & Ogawa, 2015; Ogawa, Lee, Kay, & Tank, 1990). Numerous fMRI studies have also demonstrated specific changes in functional networks associated with learning or training (Albert, Robertson, & Miall, 2009; Baldassarre et al., 2012; Guerra‐Carrillo, Mackey, & Bunge, 2014; Dong et al., 2015; Sidarta, Vahdat, Bernardi, & Ostry, 2016; Yang et al., 2016; Wang et al., 2016; Nasrallah, To, Chen, Routtenberg, & Chuang, 2016; Rjosk et al., 2017). Consistent with the results of these studies, we found that we can evaluate and assess vocational aptitudes by measuring brain function that might occur during learning or training.

The accuracies of the classifiers were better than chance for the training and test groups for all job divisions, supporting the idea that functional networks can reliably reflect vocational aptitudes. There is some variation in this accuracy across the job divisions, and some characteristics of functional networks and vocational characteristics of GATB might contribute to this variation. Therefore, it may be difficult to determine why this accuracy varies across the job divisions, but classifiers with higher accuracy can be used to more reliably predict the corresponding job aptitudes.

The specificity of job aptitudes estimated by brain function is limited by the specificity of GATB, because the brain function networks were constructed based on the job divisions defined by GATB and covered several subdivisions. A major goal of this study is to make practically useful classifiers that can predict a job aptitude with a sufficiently high accuracy so that an individual's aptitudes can be evaluated by MRI. The performance of the classifiers in our study, 58.0% and 51.9% accuracy on average for the training and test, respectively, is not high enough for this purpose, although with this accuracy, we can still apply the classifiers to evaluate vocational aptitudes at the group level or even to individuals by performing repeated MRI measurements (e.g., three times for an individual by allowing the measurement time to be prolonged). Therefore, we need to improve the accuracy to increase the method's utility and practicality; future studies should aim to obtain more specific brain information or to engineer features to improve classifier performance.

In addition, the evaluation of professionals in each job division and measurements of longitudinal changes in vocational aptitudes on the way to becoming a professional will be needed for the further development of our approach.

Another thing we have to consider in relation to our results is the distribution of the population of the training data set, because the functional networks and classifiers for vocational aptitudes were constructed using information about the variance among the participants. For some job divisions, the distribution of the 112‐participant training sample explains just over half of the distribution of the original population of Japan, as seen in Figure 1c, which may limit the external validity of our functional networks and classifiers. To devise a way to control data sets to have similar variance as the original population would be another aim for our future study, alongside the efforts for the improvement of classifier quality.

Taken together, our results demonstrate that vocational aptitudes are intrinsically represented in the brain and that we can assess them by measuring brain function. We were able to identify functional networks that reflected GATB job divisions and design classifiers that corresponded to these networks.

5. CONCLUSIONS

We found that the vocational aptitudes are represented intrinsically in the brain and that they can be measured by fMRI. We also designed classifiers that can estimate vocational aptitudes. The evaluation of vocational aptitudes by brain function will allow us to make objective assessment of job aptitudes and follow the development of these aptitudes with training or learning, both of which will be useful in assisting with career development.

Supporting information

Additional Supporting Information may be found online in the supporting information tab for this article.

Supporting Information

Click here for additional data file. (79.4KB, docx)

Supporting Information

Click here for additional data file. (116.9KB, docx)

Supporting Information

Click here for additional data file. (13KB, docx)

Supporting Information

Click here for additional data file. (53.4KB, docx)

ACKNOWLEDGMENT

This study was supported by the MEXT‐Supported program for the Strategic Research Foundation at Private Universities, 2014–2018.

Sung Y‐W, Kawachi Y, Choi U‐S, et al. Estimation of vocational aptitudes using functional brain networks. Hum Brain Mapp. 2018;39:3636–3651. 10.1002/hbm.24200

REFERENCES

  1. Aimone, J. B. , Wiles, J. , & Gage, F. H. (2009). Computational influence of adult neurogenesis on memory encoding. Neuron, 61(2), 187–202. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Albert, N. B. , Robertson, E. M. , & Miall, R. C. (2009). The resting human brain and motor learning. Current Biology, 19(12), 1023–1027. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Andrew, D. M. , Paterson, D. G. , & Longstaff, H. P. (1979). Minnesota clerical test manual, 2nd ed San Antonio, TX: The Psychological Corporation. [Google Scholar]
  4. Baldassarre, A. , Lewis, C. M. , Committeri, G. , Snyder, A. Z. , Romani, G. L. , & Corbetta, M. (2012). Individual variability in functional connectivity predicts performance of a perceptual task. Proceedings of the National Academy of Sciences of the United States of America, 109(9), 3516–3521. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Barrick, M. R. , Mount, M. K. , & Strauss, J. P. (1993). Conscientiousness and performance of sales representatives: Test of the mediating effects of goal setting. Journal of Applied Psychology, 78(5), 715. [Google Scholar]
  6. Bennett, G. K. , Seashore, H. G. , & Wesman, A. G. (1974). Manual for the differential aptitude test, 5th ed New York, NY: Psychological Corporation. [Google Scholar]
  7. Bermudez, P. , & Zatorre, R. J. (2005). Differences in gray matter between musicians and nonmusicians. Annals of the New York Academy of Sciences, 1060, 395–399. [DOI] [PubMed] [Google Scholar]
  8. Brown, S. D. , & Lent, R. W. (1996). A social cognitive framework for career choice counseling. Career Development Quarterly, 44(4), 354–366. [Google Scholar]
  9. Burgaleta, M. , MacDonald, P. A. , Martínez, K. , Roman, F. J. , Álvarez‐Linera, J. , González, A. R. , … Colom, R. (2014). Subcortical regional morphology correlates with fluid and spatial intelligence. Human Brain Mapping, 35(5), 1957–1968. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Chan, D. (1997). Racial subgroup differences in predictive validity perceptions on personality and cognitive ability tests. Journal of Applied Psychology, 82(2), 311. [DOI] [PubMed] [Google Scholar]
  11. Chao‐Gan, Y. , & Yu‐Feng, Z. (2010). DPARSF: A MATLAB toolbox for “pipeline” data analysis of resting‐state fMRI. Frontiers in System Neuroscience, 4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Choi, U.‐S. , Sung, Y.‐W. , Hong, S. , Chung, J.‐Y. , & Ogawa, S. (2015). Structural and functional plasticity specific to musical training with wind instruments. Frontiers in Human Neuroscience, 9, [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Colom, R. , Jung, R. E. , & Haier, R. J. (2006). Distributed brain sites for the g‐factor of intelligence. NeuroImage, 31(3), 1359–1365. [DOI] [PubMed] [Google Scholar]
  14. Decety, J. , & Michalska, K. J. (2010). Neurodevelopmental changes in the circuits underlying empathy and sympathy from childhood to adulthood. Developmental Science, 13(6), 886–899. [DOI] [PubMed] [Google Scholar]
  15. Deng, W. , Saxe, M. D. , Gallina, I. S. , & Gage, F. H. (2009). Adult‐born hippocampal dentate granule cells undergoing maturation modulate learning and memory in the brain. Journal of Neuroscience, 29(43), 13532–13542. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Desikan, R. S. , Segonne, F. , Fischl, B. , Quinn, B. T. , Dickerson, B. C. , Blacker, D. , … Killiany, R. J. (2006). An automated labeling system for subdividing the human cerebral cortex on MRI scans into gyral based regions of interest. NeuroImage, 31(3), 968–980. [DOI] [PubMed] [Google Scholar]
  17. Dong, M. , Li, J. , Shi, X. , Gao, S. , Fu, S. , Liu, Z. , … Tian, J. (2015). Altered baseline brain activity in experts measured by amplitude of low frequency fluctuations (ALFF): A resting state fMRI study using expertise model of acupuncturists. Frontiers in Human Neuroscience, 9, [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Dong, W. K. , & Greenough, W. T. (2004). Plasticity of nonneuronal brain tissue: Roles in developmental disorders. Mental Retardation and Developmental Disabilities Research Reviews, 10(2), 85–90. [DOI] [PubMed] [Google Scholar]
  19. Dosenbach, N. U. , Nardos, B. , Cohen, A. L. , Fair, D. A. , Power, J. D. , Church, J. A. , … Lessov‐Schlaggar, C. N. (2010). Prediction of individual brain maturity using fMRI. Science (New York, N.Y.), 329(5997), 1358–1361. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Draganski, B. , Gaser, C. , Kempermann, G. , Kuhn, H. G. , Winkler, J. , Büchel, C. , & May, A. (2006). Temporal and spatial dynamics of brain structure changes during extensive learning. Journal of Neuroscience, 26(23), 6314–6317. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Employment Security Bureau, Ministry of Health, Labour and Welfare (2013). Kousei Roudoushou Hen Ippan Shokugyou Tekisei Kensa Tebiki Shinro Shidou Syokugyou Shidou You Kaitei 2 Han (The manual of Japanese version of general aptitude test battery revised 2nd edition). Tokyo: Employment Research Corporation. [Google Scholar]
  22. Fan, F. , Zhu, C. , Chen, H. , Qin, W. , Ji, X. , Wang, L. , … Yu, C. (2013). Dynamic brain structural changes after left hemisphere subcortical stroke. Human Brain Mapping, 34(8), 1872–1881. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Frey, M. C. , & Detterman, D. K. (2004). Scholastic Assessment or g? The Relationship Between the Scholastic Assessment Test and General Cognitive Ability. Psychological Science, 15(6), 373–378. [DOI] [PubMed] [Google Scholar]
  24. Golestani, N. , Paus, T. , & Zatorre, R. J. (2002). Anatomical correlates of learning novel speech sounds. Neuron, 35(5), 997–1010. [DOI] [PubMed] [Google Scholar]
  25. Gottfredson, L. S. (1998). The general intelligence factor. Scientific American, Incorporated. [Google Scholar]
  26. Gottfredson, L. S. (2003). The challenge and promise of cognitive career assessment. Journal of Career Assessment, 11(2), 115–135. [Google Scholar]
  27. Guerra‐Carrillo, B. , Mackey, A. P. , & Bunge, S. A. (2014). Resting‐state fMRI: A window into human brain plasticity. Neuroscientist, 20(5), 522–533. [DOI] [PubMed] [Google Scholar]
  28. Hackman, D. A. , & Farah, M. J. (2009). Socioeconomic status and the developing brain. Trends in Cognitive Sciences, 13(2), 65–73. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Haier, R. J. , Jung, R. E. , Yeo, R. A. , Head, K. , & Alkire, M. T. (2004). Structural brain variation and general intelligence. NeuroImage, 23(1), 425–433. [DOI] [PubMed] [Google Scholar]
  30. Hunter, J. E. (1986). Cognitive ability, cognitive aptitudes, job knowledge, and job performance. Journal of Vocational Behavior, 29(3), 340–362. [Google Scholar]
  31. Jung, R. E. , Wertz, C. J. , Meadows, C. A. , Ryman, S. G. , Vakhtin, A. A. , & Flores, R. A. (2015). Quantity yields quality when it comes to creativity: A brain and behavioral test of the equal‐odds rule. Frontiers in Psychology, 6, [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Kwallek, N. , Lewis, C. M. , Lin‐Hsiao, J. W. D. , & Woodson, H. (1996). Effects of nine monochromatic office interior colors on clerical tasks and worker mood. Color Research & Application, 21(6), 448–458. [Google Scholar]
  33. Legree, P. J. (1995). Evidence for an oblique social intelligence factor established with a Likert‐based testing procedure. Intelligence, 21(3), 247–266. [Google Scholar]
  34. Lent, R. W. , Brown, S. D. , & Hackett, G. (1994). Toward a unifying social cognitive theory of career and academic interest, choice, and performance. Journal of Vocational Behavior, 45(1), 79–122. [Google Scholar]
  35. MacDonald, P. A. , Ganjavi, H. , Collins, D. L. , Evans, A. C. , & Karama, S. (2014). Investigating the relation between striatal volume and IQ. Brain Imaging and Behavior, 8(1), 52–59. [DOI] [PubMed] [Google Scholar]
  36. Maguire, E. A. , Woollett, K. , & Spiers, H. J. (2006). London taxi drivers and bus drivers: A structural MRI and neuropsychological analysis. Hippocampus, 16(12), 1091–1101. [DOI] [PubMed] [Google Scholar]
  37. May, A. (2011). Experience‐dependent structural plasticity in the adult human brain. Trends in Cognitive Sciences, 15(10), 475–482. [DOI] [PubMed] [Google Scholar]
  38. McClearn, G. E. , Johansson, B. , Berg, S. , Pedersen, N. L. , Ahern, F. , Petrill, S. A. , & Plomin, R. (1997). Substantial genetic influence on cognitive abilities in twins 80 or more years old. Science (New York, N.Y.), 276(5318), 1560–1563. [DOI] [PubMed] [Google Scholar]
  39. McGee, M. G. (1979). Human spatial abilities: Psychometric studies and environmental, genetic, hormonal, and neurological influences. Psychological Bulletin, 86(5), 889. [PubMed] [Google Scholar]
  40. Muroyama, H. (2016). Shokugyo Nouryoku no Hyouka: GATB wo Mochiita 13 Nenkan no Data no Kentou. (The assessment of vocational aptitudes using the GATB date for 13 years. JILPT Reference (Shiryo) Series No.169.) Tokyo: The Japan Institute for Labour Policy and Training. [Google Scholar]
  41. Nasrallah, F. A. , To, X. V. , Chen, D.‐Y. , Routtenberg, A. , & Chuang, K.‐H. (2016). Functional connectivity MRI tracks memory networks after maze learning in rodents. NeuroImage, 127, 196–202. [DOI] [PubMed] [Google Scholar]
  42. Ogawa, S. , Lee, T.‐M. , Kay, A. R. , & Tank, D. W. (1990). Brain magnetic resonance imaging with contrast dependent on blood oxygenation. Proceedings of the National Academy of Sciences of the United States of America, 87(24), 9868–9872. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Pereira, A. C. , Huddleston, D. E. , Brickman, A. M. , Sosunov, A. A. , Hen, R. , McKhann, G. M. , … Small, S. A. (2007). An in vivo correlate of exercise‐induced neurogenesis in the adult dentate gyrus. Proceedings of the National Academy of Sciences, 104(13), 5638–5643. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Plomin, R. , Owen, M. J. , & McGuffin, P. (1994). The genetic basis of complex human behaviors. Science (New York, N.Y.), 264(5166), 1733. [DOI] [PubMed] [Google Scholar]
  45. Rhyu, I. J. , Bytheway, J. A. , Kohler, S. J. , Lange, H. , Lee, K. J. , Boklewski, J. , … Greenough, W. T. (2010). Effects of aerobic exercise training on cognitive function and cortical vascularity in monkeys. Neuroscience, 167(4), 1239–1248. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Rjosk, V. , Lepsien, J. , Kaminski, E. , Hoff, M. , Sehm, B. , Steele, C. J. , … Ragert, P. (2017). Neural correlates of mirror visual feedback‐induced performance improvements: A resting‐state fMRI study. Frontiers in Human Neuroscience, 11, [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Royle, N. A. , Booth, T. , Hernández, M. C. V. , Penke, L. , Murray, C. , Gow, A. J. , … Deary, I. J. (2013). Estimated maximal and current brain volume predict cognitive ability in old age. Neurobiology of Aging, 34(12), 2726–2733. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Sidarta, A. , Vahdat, S. , Bernardi, N. F. , & Ostry, D. J. (2016). Somatic and reinforcement‐based plasticity in the initial stages of human motor learning. Journal of Neuroscience, 36(46), 11682–11692. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Shokugyou Tekisei Kensa Kenkyukai (Vocational aptitude test research corporation) (2011). Kousei Roudoushou Hen Ippan Shokugyou Tekisei Kensa Tebiki Shinro Shidou Syokugyou Shidou You Q & A Syuu. (The Japanese version of general aptitude test battery for career and vocational guidance Q & A.) Tokyo: Employment Research Corporation. [Google Scholar]
  50. Teipel, S. J. , Bokde, A. L. , Meindl, T. , Amaro, E. , Soldner, J. , Reiser, M. F. , … Hampel, H. (2010). White matter microstructure underlying default mode network connectivity in the human brain. NeuroImage, 49(3), 2021–2032. [DOI] [PubMed] [Google Scholar]
  51. Thompson, P. M. , Cannon, T. D. , Narr, K. L. , Van Erp, T. , Poutanen, V.‐P. , Huttunen, M. , … Khaledy, M. (2001). Genetic influences on brain structure. Nature Neuroscience, 4(12), 1253. [DOI] [PubMed] [Google Scholar]
  52. Tremblay, M.‐È. , Lowery, R. L. , & Majewska, A. K. (2010). Microglial interactions with synapses are modulated by visual experience. PLoS Biology, 8(11), e1000527. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Turkheimer, E. , Haley, A. , Waldron, M. , D'Onofrio, B. , & Gottesman, I. I. (2003). Socioeconomic status modifies heritability of IQ in young children. Psychological Science, 14(6), 623–628. [DOI] [PubMed] [Google Scholar]
  54. United States Department of Defense (1984). Test manual for the armed services vocational aptitude battery. North Chicago, IL: U.S. Military Entrance Processing Command. [Google Scholar]
  55. United States Department of Labor (1970). Manual for the USES general aptitude test battery section III: Development. Washington, DC: Government Printing Office. [Google Scholar]
  56. Voss, H. U. , & Schiff, N. D. (2009). MRI of neuronal network structure, function, and plasticity. Progress in Brain Research, 175, 483–496. [DOI] [PubMed] [Google Scholar]
  57. Wang, S. , Wang, G. , Lv, H. , Wu, R. , Zhao, J. , & Guo, W. (2016). Abnormal regional homogeneity as potential imaging biomarker for psychosis risk syndrome: A resting‐state fMRI study and support vector machine analysis. Scientific Reports, 6(1), http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4897690/. [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. Wonderlic, E. F. (1983). Wonderlic personnel test manual. Northfield, IL: Wonderlic EF and Associates, INC. [Google Scholar]
  59. Yahata, N. , Morimoto, J. , Hashimoto, R. , Lisi, G. , Shibata, K. , Kawakubo, Y. , … Kawato, M. (2016). A small number of abnormal brain connections predicts adult autism spectrum disorder. Nature Communications, 7, 11254 10.1038/ncomms11254. [DOI] [PMC free article] [PubMed] [Google Scholar]
  60. Yan, C.‐G. , Wang, X.‐D. , Zuo, X.‐N. , & Zang, Y.‐F. (2016). DPABI: Data processing & analysis for (resting‐state) brain imaging. Neuroinformatics, 14(3), 339–351. [DOI] [PubMed] [Google Scholar]
  61. Yang, C.‐C. , Barrós‐Loscertales, A. , Pinazo, D. , Ventura‐Campos, N. , Borchardt, V. , Bustamante, J.‐C. , … Ávila, C. (2016). State and training effects of mindfulness meditation on brain networks reflect neuronal mechanisms of its antidepressant effect. Neural Plasticity, 2016, 1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  62. Yurgelun‐Todd, D. A. , Killgore, W. D. , & Cintron, C. B. (2003). Cognitive correlates of medial temporal lobe development across adolescence: A magnetic resonance imaging study. Perceptual and Motor Skills, 96(1), 3–17. [DOI] [PubMed] [Google Scholar]
  63. Zalesky, A. , Fornito, A. , & Bullmore, E. T. (2010). Network‐based statistic: Identifying differences in brain networks. NeuroImage, 53(4), 1197–1207. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Additional Supporting Information may be found online in the supporting information tab for this article.

Supporting Information

Click here for additional data file. (79.4KB, docx)

Supporting Information

Click here for additional data file. (116.9KB, docx)

Supporting Information

Click here for additional data file. (13KB, docx)

Supporting Information

Click here for additional data file. (53.4KB, docx)

Articles from Human Brain Mapping are provided here courtesy of Wiley

RESOURCES