Abstract

INTRODUCTION

Strategies to optimize cognitive and brain development, including physical activity, computerized training games, martial arts, yoga, mindfulness, and school curricula, are in the spotlight (Diamond and Lee, 2011). For example, aerobic fitness plays an important role in the brain health of children (see Chaddock et al., 2011 for a review), especially in terms of brain structure and brain function, including gray matter volumes [e.g., hippocampus (Chaddock et al., 2010a), basal ganglia (Chaddock et al., 2010b)] and functional brain networks (e.g., frontal and parietal cortex; Voss et al., 2011; Chaddock et al., 2012a). These fitness-related differences in brain health are often coupled with performance differences, such that higher fit children have been shown to outperform their lower fit peers on tasks of cognitive control (Chaddock et al., 2010b, 2012a; Voss et al., 2011) and memory (Chaddock et al., 2010a) as well as scholastic achievement tests in the classroom (Coe et al., 2006; Castelli et al., 2007; Trudeau and Shephard, 2008; Chomitz et al., 2009).

Still unknown is the relationship between aerobic fitness and the microstructure of white matter tracts during childhood. The structure of white matter is important for efficient transmission of information between gray matter as well as for the integration of brain areas into structural networks to support cognitive function. During childhood, many white matter tracts increase in estimates of microstructure, in parallel with improvements in cognition (Barnea-Goraly et al., 2005; Muetzel et al., 2008; see Schmithorst and Yuan, 2010, for a review), but there has been little research about individual differences in the structure of these pathways, the plasticity of white matter fibers, or their susceptibility to intervention during development. Given that the size and function of several different gray matter regions differ between higher fit and lower fit children (Hillman et al., 2009; Chaddock et al., 2010a,b; Davis et al., 2011; Pontifex et al., 2011; Voss et al., 2011; Chaddock et al., 2012a), it is possible that white matter microstructure that connects gray matter areas also differs as a function of aerobic fitness. In support of such a hypothesis, higher aerobic fitness levels and participation in physical activity are positively associated with white matter structure in older adults (Colcombe et al., 2006; Marks et al., 2007; Johnson et al., 2012; Voss et al., 2012).

Here, for the first time, we used diffusion tensor imaging (DTI) to investigate the microstructural properties of white matter in higher fit and lower fit 9- and 10-year-old children. We analyzed fractional anisotropy (FA), a general index of white matter microstructure, hypothesized to be higher in tightly bundled, structurally compact fibers with high integrity (Basser, 1995; Beaulieu, 2002; Sen and Basser, 2005; Rykhlevskaia et al., 2008). We also explored specific patterns of diffusivity [radial diffusivity (RD) and axial diffusivity (AD)], hypothesized to reflect potential biological properties of white matter microstructure (Basser, 1995; Pierpaoli and Basser, 1996; Pierpaoli et al., 1996; Song et al., 2002). For example, a reduction in RD is observed in the presence of remyelination, causing RD to often be used as a marker of myelination (Song et al., 2002, 2003, 2005; Nair et al., 2005; Budde et al., 2007; Rykhlevskaia et al., 2008). Additionally, AD is said to be sensitive to changes in axonal fibers including axonal diameter, loss or damage (Song et al., 2003; Budde et al., 2007). By exploring the associations among aerobic fitness and measures of diffusivity (FA, RD, AD), we can investigate microstructural white matter properties influenced by aerobic fitness during child development.

We predicted that higher fit children would show greater estimates of white matter microstructure than lower fit children. Because the structure and function of brain circuits important for cognitive control have been found to be sensitive to aerobic fitness (Bunge and Crone, 2009; see Chaddock et al., 2012b for a review), we focused on white matter tracts that relate to these cognitive skills (Liston et al., 2006; Madsen et al., 2010; Tamnes et al., 2012; Chaddock-Heyman et al., 2013), and that connect frontal, striatal and parietal brain areas. In particular, in this first exploration of the association between aerobic fitness and white matter microstructure in children, we examined a priori white matter tracts (included in Chaddock-Heyman et al., 2013). These tracts included the corpus callosum, which connects the left and right cerebral hemispheres (Olesen et al., 2003; Nagy et al., 2004; Tamnes et al., 2012), the corona radiata which carries ascending and descending information throughout the brain (Olesen et al., 2003; Bendlin et al., 2010; Niogi et al., 2010; Chaddock-Heyman et al., 2013), and the superior longitudinal fasciculus, a tract that provides bidirectional information transfer between the frontal and parietal cortex (Nagy et al., 2004; Chaddock-Heyman et al., 2013). Additionally, we examined the posterior thalamic radiation, nerve fibers connecting the thalamus and caudate nucleus with the cerebral cortex (Olson et al., 2009; Chaddock-Heyman et al., 2013), and the cerebral peduncle, part of the brainstem, which includes nerve tracts carrying motor information to and from the brain to the rest of the body (Olson et al., 2009; Chaddock-Heyman et al., 2013). The present study contributes to our understanding of plasticity of the microstructure of white matter that allows efficient communication between gray matter areas in the brain and integrates the brain into networks important for cognition and academic achievement.

MATERIALS AND METHODS

REGION-OF-INTEREST ANALYSIS

Diffusion values (FA, RD, AD) were calculated for each participant within a priori ROIs, created from the JHU ICBM-DTI-81 white matter labels atlas² (Mori et al., 2005; Wakana et al., 2007; Hua et al., 2008). Tract ROIs were created in the genu, body and splenium of the corpus callosum, and the left and right corona radiata, superior longitudinal fasciculus, posterior thalamic radiation, and cerebral peduncle (Figure ​Figure11). An FSL command, fslmaths, was used to create each ROI (e.g., fslmaths JHUAtlas –uthr 16 –thr 16 RCerebralPeduncle).

Open in a separate window

FIGURE 1

Illustrations of the white matter tract ROIs in the corpus callosum (white = genu; yellow = body; purple = splenium), corona radiata (anterior = green; superior = pink; posterior = peach), superior longitudinal fasciculus (blue), posterior thalamic radiation (light blue), and cerebral peduncle (red).

RESULTS

DISCUSSION

The present study is the first to demonstrate that aerobic fitness is positively related to the microstructure of white matter fiber tracts in the brain during childhood. Higher fit 9- and 10-year-old children (>70th percentile VO₂max) showed greater FA in sections of the corpus callosum, corona radiata, and superior longitudinal fasciculus, compared to lower fit children (<30th percentile VO₂max). An exploration of the components of the diffusion tensor showed that the relationships between aerobic fitness and FA were primarily characterized by fitness group differences in RD. Because axonal caliber and thickness of the myelin sheath determine conduction velocity (Paus, 2010), our results raise the possibility that children with higher fitness levels may have faster neural conduction between brain regions important for cognitive control. White matter microstructure may be another potential neural mechanism of aerobic fitness that assists in efficient communication between gray matter and the integration of regions into networks to support cognitive function and academic achievement.

The analysis of different patterns of diffusivity allows us to speculate about the biological properties of white matter structure that may differ as a function of aerobic fitness in children. First, our data suggest that higher fit children have greater FA, or more tightly bundled and structurally compact fibers, in the corpus callosum, specifically in the body of the white matter tract. This effect in children supports and extends the association between aerobic fitness and the body of the corpus callosum in an aging population (Johnson et al., 2012). Whereas the fitness-FA association in the corpus callosum in older adults was driven by negative fitness-RD associations (Johnson et al., 2012), we did not observe aerobic fitness group differences in RD or AD in this tract in the child sample. Thus, it is possible that the increased estimate of fiber integrity in this tract was not solely a function of greater myelination (RD) or axonal fiber diameter (AD). Other tissue differences may have played a role, such as increased neurotrophic factors or increased vascularization, two cellular effects of exercise which are observed in rodents that could influence white matter (Ding et al., 2006). Regardless of the biological mechanism, because the corpus callosum connects the left and right cerebral hemispheres and facilitates interhemispheric communication and the exchange of cognitive, motor and sensory information between the hemispheres (Gazzaniga, 1995), structurally compact fibers may help lead to superior cognitive performance in higher fit children, as suggested in previous work (Olesen et al., 2003; Nagy et al., 2004; Tamnes et al., 2012).

We also demonstrate significant effects of aerobic fitness on FA in the superior corona radiata and superior longitudinal fasciculus, driven by associations with RD. We suggest that aerobic fitness during childhood may influence fiber structural integrity and fiber alignment in these tracts perhaps via increased myelination (reductions in RD; Song et al., 2003, 2005; Sun et al., 2006, 2008). To link RD to myelination, Song et al. (2005) used an experimental model of demyelination and remyelination in a mouse brain and demonstrated that RD increased with demyelination and subsequently decreased with the progression of remyelination. The RD changes were found to be specific to changes in myelin integrity, and distinct from axonal injury (Song et al., 2005). Like the corpus callosum, no aerobic fitness differences were found in terms of AD in the superior corona radiata or superior longitudinal fasciculus, which suggests that axonal diameter, loss or damage may not relate to individual differences in fitness (Song et al., 2003; Budde et al., 2007). This is not unexpected given that axonal information is not influenced by myelination (Song et al., 2003; Budde et al., 2007). Because the corona radiata has ascending and descending tracts from the cerebral cortex that integrate information throughout the brain, and cognitively, the tract is known to play a role in processing speed, cognitive control and memory (Olesen et al., 2003; Bendlin et al., 2010; Niogi et al., 2010; Chaddock-Heyman et al., 2013), it is important to investigate its plasticity. Similarly, as the superior longitudinal fasciculus provides bidirectional information transfer between the frontal and parietal cortex (Petrides and Pandya, 1984; Schmahmann and Pandya, 2006), and these connections have been found to correlate with performance on tasks of attentional and interference control (Chaddock-Heyman et al., 2013), and working memory in children (Nagy et al., 2004), our data have promising implications for the influence of aerobic fitness on brain health during childhood.

Whereas we demonstrate that aerobic fitness is associated with white matter microstructure across a number of tracts throughout the brain, the results also demonstrate some specificity. The fitness groups did not differ in the structure of the posterior thalamic radiation or cerebral peduncle, two tracts involved in motor function. Given our recent report of an association between interference control and the microstructure of the posterior thalamic radiation and cerebral peduncle (Chaddock-Heyman et al., 2013), and our reports that higher fit children show less interference RT than lower fit peers (Chaddock et al., 2010b), it is surprising that higher fit children did not have greater estimates of structure in these lower level motor tracts. However, some research suggests that cognitive gains with physical activity and fitness are specific to tasks that require increased cognitive control (Kramer et al., 1999; Colcombe and Kramer, 2003; Colcombe et al., 2004; Chaddock et al., 2011), so it is possible that aerobic fitness would not influence lower level motor tracts, as observed herein.

This study provides a first step in identifying the relationship between aerobic fitness and white matter microstructure during childhood, but the results should be evaluated in the context of their limitations. Studies with larger sample sizes, whole-brain exploratory analyses, a physical activity intervention, and a number of cognitive tasks should be conducted to confirm the complex relationship among physical activity, aerobic fitness, white matter structure, and cognition. Because we approached the study with a priori hypotheses, additional investigations should employ whole-brain analyses corrected for multiple comparisons. Also, it is important to remember that DTI does not measure tissue parameters (e.g., fiber integrity, myelination) directly, but measures the displacement of water molecules. Thus, indices of underlying microstructural properties can only be inferred from this displacement. Furthermore, because there are many axons within a voxel and TBSS only selects one axon per voxel, a number of different tissue parameters can result in differences in FA, RD, and AD.

These results arrive at an important time, as children become increasingly unfit and sedentary, and educators reduce or eliminate opportunities for physical activity during the school day in favor of academic topics (Centers for Disease Control and Prevention, 2010). In fact, white matter microstructure in the corona radiata has been found to relate to mathematics performance in the classroom (van Eimeren et al., 2008), which raises the possibility that the removal of fitness training opportunities during the school day may unintentionally have deleterious effects on white matter tracts, potentially reducing scholastic achievement. In fact, our study raises the possibility that greater white matter integrity, perhaps via greater myelination, is one pathway by which higher fit children excel in cognitive and academic performance compared to their less fit peers. Hopefully these findings will reinforce the importance of aerobic fitness during development and lead to additional physical activity opportunities in and out of the school environment.

Conflict of Interest Statement

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Acknowledgments

Footnotes

REFERENCES