Resting-state connectivity and its association with cognitive performance, educational attainment, and household income in UK Biobank (N = 3,950)

Cognitive ability is an important predictor of lifelong physical and mental well-being and its impairments are associated with many psychiatric disorders. Higher cognitive ability is also associated with greater educational attainment and increased household income. Understanding neural mechanisms underlying cognitive ability is therefore of crucial importance for determining the nature of these associations. In the current study, we examined the spontaneous activity of the brain at rest to investigate its relationships with not only cognitive ability, but also educational attainment and household income. We used a large sample of resting-state neuroimaging data from UK Biobank (N=3,950). Firstly, analysis at the whole-brain level showed that connections involving the default mode network (DMN), fronto-parietal network (FPN) and cingulo-opercular network (CON) were significantly positively associated with levels of cognitive performance assessed by a verbal-numerical reasoning test (standardised β ranged from 0.054 to 0.097). Connections associated with higher levels of cognitive performance were also significantly positively associated with educational attainment (r=0.48, N=4,160) and household income (r=0.38, N=3,793). Further, analysis on the coupling of functional networks showed that better cognitive performance was associated with more positive DMN-CON connections, decreased cross-hemisphere connections between homotopic network in CON and FPN, and stronger CON-FPN connections (absolute β ranged from 0.034 to 0.063). The present study finds that variation in brain resting state functional connectivity associated with individual differences in cognitive ability, largely involving DMN and lateral prefrontal networks. Additionally, we provide further evidence of shared neural associations of cognitive ability, educational attainment, and household income.


Abstract:
Cognitive ability is an important predictor of lifelong physical and mental well-being and its impairments are associated with many psychiatric disorders. Higher cognitive ability is also associated with greater educational attainment and increased household income.
Understanding neural mechanisms underlying cognitive ability is therefore of crucial importance for determining the nature of these associations. In the current study, we examined the spontaneous activity of the brain at rest to investigate its relationships with not only cognitive ability, but also educational attainment and household income. We used a large sample of resting-state neuroimaging data from UK Biobank (N=3,950). Firstly, analysis at the whole-brain level showed that connections involving the default mode network (DMN), fronto-parietal network (FPN) and cingulo-opercular network (CON) were significantly positively associated with levels of cognitive performance assessed by a verbal-numerical reasoning test (standardised β ranged from 0.054 to 0.097).
Connections associated with higher levels of cognitive performance were also significantly positively associated with educational attainment (r=0.48, N=4,160) and household income (r=0.38, N=3,793). Further, analysis on the coupling of functional networks showed that better cognitive performance was associated with more positive DMN-CON connections, decreased cross-hemisphere connections between homotopic network in CON and FPN, and stronger CON-FPN connections (absolute β ranged from 0.034 to 0.063). The present study finds that variation in brain resting state functional connectivity associated with individual differences in cognitive ability, largely involving DMN and lateral

Introduction
General cognitive ability is positively associated with a longer duration of education 1 , better examination performance in school 2 and better workplace performance 3 . Cognitive ability in youth is also positively associated with higher socioeconomic status in adulthood 3 , and with reduced risk of several mental and physical diseases [3][4][5][6] . Thus, cognitive ability is a key trait associated with many educational, social and health outcomes 7 , and therefore identifying the associated neural mechanisms will help better understand the causes of these associations.
Structural and event-related fMRI studies have consistently identified prefrontal brain regions as having the strongest associations with general cognitive ability 8,9 . These regions play a crucial role in executive control 10 and multisensory integration 11 , and can be assessed using various task-based paradigms [12][13][14] . However, it has recently been demonstrated that the brain is highly active in the absence of experimental stimuli, i.e. when it is in 'resting state'. The activity of the brain under resting state is metabolically demanding and topologically efficient; it has been proposed that this actively maintains neural signalling in preparation for quick adaptions 15,16 . Such spontaneous modulations at rest are temporally correlated between distant brain regions, forming the linkage known as functional connectivity.
The spatial patterns of functional connectivity are known as resting state In order to address the above issues, we first examined associations between whole-brain connectivity measures and cognitive performance using the VNR task. Second, to compare overlapping networks involved in cognitive performance, educational attainment and household income, we conducted similar whole-brain tests or each trait separately, which were then compared with the results for cognitive performance. Finally, we examined the integrative coupling between these resting state functional network connections using a network-of-interest (NOI) approach, focussing on networks identified by the previous two whole brain analyses. Educational attainment showed positive association with cognitive performance, with age, age 2 and sex controlled (β=0.457, p<2Χ10 -16 ).
The proportion of people who reported having household income at each level is shown in Figure 1.  Table S1). There were 18 significant connections that showed a significant positive association between connection strength and cognitive functioning in VNR; and 8 connections were negatively associated (Supplementary Table   S1). The positive connections largely involved the DMN, which includes   Full lists of regions of the maps for the above results are presented in Supplementary Table S4.
The spatial maps for the results of cognitive performance in VNR, educational attainment, and household income overlapped substantially (Figures 2 and 3).
By performing correlation analysis at the standardised effect sizes of whole brain (see Methods, Statistical methods), we found a correlation of r=0.47 (df=1,483, p<2Χ10 -16 ) between the global effect sizes for cognitive performance and educational attainment. The correlation between the effect sizes of cognitive performance and household income was r=0.38 (df=1,483, p<2Χ10 -16 ) ( Figure 4).

Network-of-interest (NOI) test on VNR, educational attainment, and household income
The whole-brain tests showed that the connections associated with cognitive performance in VNR, educational attainment and household income were predominantly located within the DMN (covering medial PFC, PCC and TPJ), cingulo-opecular network (CON, covering ventral lateral PFC, and dorsal ACC) and fronto-parietal network (FPN, covering dorsal lateral PFC and posterior parietal cortex). Therefore, DMN, CON, and FPN were selected as NOI from another group-ICA of lower resolution so these networks could be fully extracted (see Methods). The pairwise between-network coupling of these five networks (DMN was unilateral, and CON and FPN were separately extracted on each hemisphere) was tested to determine their association with cognitive performance, educational attainment, and/or household income. The spatial maps for the above components can be viewed in Figure 5. The valence and values for the coupling of the above NOI were shown in Table 1. Similar with the analyses at whole-brain connectivity, the values of the connections were transformed into connection strength before they were fed into the model.

Discussion
In the present study, we utilized a large population-based sample of ~4,000 We used a large sample and provided evidence that, in addition to the broadly suggested idea of lateral PFC playing a crucial role in cognitive processing, DMN was also associated with cognitive performance (β of connections positively associated with cognitive ability ranged from 0.054 to 0.097) 8,24,38 .
Previous studies showed that DMN serves as a hub for the whole brain 28 . In comparison with other functional networks, DMN showed a higher metabolic rate in resting-state 27 , stronger connections with the rest of the whole brain in both task-free and task-engaging situations 39 , and a key role in maintaining basic levels of wakefulness/alertness in the brain 40  DMN, as a communicative hub, contributes to functional efficiency over the whole brain 44 , potentially producing better integration and cooperation in core regions that are important for cognitive tasks.
Additionally, the present study tested the coupling between networks of interest. More positive DMN-CON coupling was associated with better cognitive ability (absolute β >0.045). In addition to the well-recognised task-positive lateral prefrontal cortex (therefore anti-correlated with the DMN), our findings in this large single-scanner sample lend substantial credence to increasing evidence that the CON itself 18,46 , and its positive coupling with the DMN (in both resting-state 47 and event-related studies 48 ) is highly pertinent for important aspects of cognitive performance. The role of the CON was related with maintaining task-engaging status 18, 49 and flexibly switching between the DMN and central executive network based on experimental context 21,50 . The experimental context in which CON and DMN were found to be simultaneously activated was often about goal-directed cognition 21 , which involves self-driven retrieval of memory or learned experience and self-regulatory planning 29 . As the DMN is associated with self-referential processing 28 and self-driven cognition like retrieval of personal experience 51 and planning 29,52 , positive coupling of the CON and DMN may indicate recruitment of self-referential and goal-oriented activity. Therefore successful DMN-CON coupling may be useful in maintaining internal mechanisms that support cognitive processing and long-term learning 21 .
The coupling of networks involving lateral PFC showed that better cognitive performance was associated with stronger CON-FPN connections (absolute β >0.034). This result is consistent with previous structural and functional findings that support the key role of prefrontal areas on cognitive performance 8,53 . We also found that better cognitive performance was related to between-hemisphere dissociation within networks (absolute β >0.040).
Whereas this is the first time to our knowledge that this has been examined in a study of large sample, such reduced structural coupling between left and right lateral PFC has been observed in schizophrenic patients with impaired cognitive performance 54 . More lateralisation of the brain is associated with better cognitive performance 55,56 , whereas, less lateralisation, especially in prefrontal cortex, is related with reduced specialisation of brain functions across hemispheres, therefore the advantageous de-coupling we report here potentially denotes increased brain efficiency 55,57 .
The whole-brain connection map for cognitive performance overlaps substantially with those from educational attainment and household income.
Further analyses showed that there were global correlations of cognitive ability with educational attainment (r=0.47) and with household income (r=0.38).
GWAS studies found that cognitive performance and educational attainment  58,59 ) and educational duration show partially overlapping associations with some structural brain measures in older age 60 . Taken together, one interpretation of these data is that the functional hallmarks of a more 'intelligent' and better educated brain are related to income by virtue of these temporally preceding factors. It could equally be the case that income confers additional lifestyle benefits that also influence these cerebral characteristics; the causal direction that gives rise to the highly overlapping functional connectivity reported here would be more adequately addressed with longitudinal multi-modal data.
A limitation for the current study is that the verbal-numerical reasoning test, as connections between brain components derived in two different resolutions, giving us another strength of studying both the connections over the whole brain and the coupling of bulk intrinsic functional networks within a single dataset. Finally, in addition to visual checking of overlapping regions of the significant connections, we statistically compared the functional connectivity associated with cognitive ability, educational attainment and household income over the whole brain, giving a magnitude of neural associations among them.

Conclusion
The present study used a large, population-based sample, who provided multi-dimensional rs-fMRI data, and found substantial evidence for functional neural associations cognitive ability (verbal-numerical reasoning) both in whole-brain dynamics and the coupling of intrinsic functional networks. The findings also characterised the degree of rs-fMRI overlap between cognitive ability and educational and socioeconomic level, providing evidence of the overlapping biological associations on the neurological level.

Imaging data
We used the network matrices from the IDPs (imaging-derived phenotypes) which were processed by the UK Biobank imaging project team 30 . All the processing of resting-state data described in this section was performed by the UK Biobank team, including acquisition and pre-processing of resting-state data and estimation of functional connectivity. Quality check was conducted by UK Biobank following the standard protocol 61 . The detailed methods of the UK Biobank imaging processing can be found in a previous protocol paper 30 . For clarification, these processes are described briefly below.
All imaging data were obtained on a Siemens Skyra 3.0 T scanner (Siemens Medical Solutions, Germany). The fMRI scans employed a single-shot The final 55*55 and 21*21 partial correlation matrices were used as measurements of functional connections. Both types of correlation matrices were used as they addressed different principles. Since the dimensionality of the 55*55 matrix is higher, it gives a higher resolution of the whole-brain functional connectome. We therefore used this for the first two steps of whole-brain analysis. The lower dimensionality matrix, on the other hand, is better able to identify large functional networks, such as the DMN, frontal networks, and primary and higher level visual networks 23 . Hence, the functional networks that were found in the whole-brain analysis were selected from the 21*21 matrix as NOI, and the partial correlations between the NOI were used as the proxy of the coupling of these functional networks. distribution of the scores in the sample analysed here is presented in Figure 1.

Educational attainment and household income
Educational attainment and household income phenotypes were self-reported in a touchscreen-questionnaire session, the details of which are provided on the study website (http://biobank.ctsu.ox.ac.uk/crystal/refer.cgi?id=100471, http://biobank.ctsu.ox.ac.uk/crystal/refer.cgi?id=100256). Descriptive statistics of educational attainment and household income are presented in Figure 1.
For educational attainment, participants could choose at least one of the to determine the level of household income (<£18,000 as 1, >£100,000 as 5).

Statistical methods
The associations between brain connections and cognitive performance, educational attainment, and household income were tested using the linear GLM function in R (https://stat.ethz.ch/R-manual/R-devel/library/stats/html/glm.html). All imaging analyses were adjusted for age, age 2 and sex.
As the L2-regularised partial correlation represents more accurately the direct connections, as described in the section for imaging data above, we used the

Author contributions
XS developed the design of the study and conducted the analyses. XS, AMM, and HCW drafted the manuscript. AMM and HCW supervised and contributed to the design of the study. IJD, SRC, SJR, DMH, SML, and MEB were involved in overseeing analysis methodology and editing the paper. MJA was involved in curating the data. UK Biobank collected all data and was involved in the preprocessing of imaging data. All authors discussed and commented on the manuscript.

Conflicts of interest
The authors declare no competing financial interests.      The spatial maps for the ICA nodes that involved in the significant connections were multiplied by their effect sizes, then the spatial map in (b) was generated by summing up the weighted maps. To better illustrate the regions involving in significant connections, a threshold of 50% of the highest intensity was applied, so the regions with intensity higher than the threshold would show on the map.  Figure 2. The categorisation of components of brain regions in the circular brain network illustration is identical with Figure 2. Again like Figure  2, A threshold of 50% of the highest value was applied for better illustration of the projection of brain regions on MNI template.   Table 1.  Table 1). An orange arrow means positive association between cognitive ability with the absolute strength of a connection, whilst a blue arrow indicates decreased absolute strength of a connection with better cognitive performance. Solid arrows are positive connections and dashed ones are negative. (b) and (c) represent the association of cognitive performance in verbal-numerical reasoning and the connection between left/right CON (β=0.061 and -0.045 respectively for left/right CON) and DMN (β=-0.045). Y-axis represent the normalised correlation coefficient between temporal modulations of networks. Better cognitive performance was associated with more positive connections between DMN and bilateral CON. The spatial maps of the functional networks can be found in Figure 5.  R  e  s  t  i  n  g  -s  t  a  t  e  f  M  R  I  ,  c  o  g  n  i  t  i  o  n  ,  e  d  u  c  a  t  i  o  n  a  l  a  t  t  a  i  n  m  e  n  t  a  n  d  h  o  u  s  e  h  o  l  d  i  n  c  o  m  e  f  r  o  m  U  K  B i o b a n k 3 9