White matter growth as a mechanism of cognitive development in children

Abstract

We examined the functional role of white matter growth in cognitive development. Specifically, we used hierarchical regression analyses to test the unique contributions of age versus white matter integrity in accounting for the development of information processing speed. Diffusion tensor imaging was acquired for 17 children and adolescents (age range 6–17 years), with apparent diffusion coefficient (ADC) and fractional anisotropy (FA) calculated for 10 anatomically defined fiber pathways and 12 regions of hemispheric white matter. Measures of speeded visual–spatial searching, rapid picture naming, reaction time in a sustained attention task, and intelligence were administered. Age-related increases were evident across tasks, as well as for white matter integrity in hemispheric white matter. ADC was related to few measures. FA within multiple hemispheric compartments predicted rapid picture naming and standard error of reaction time in sustained attention, though it did not contribute significantly to the models after controlling for age. Independent of intelligence, visual–spatial searching was related to FA in a number of hemispheric regions. A novel finding was that only right frontal–parietal regions contributed uniquely beyond the effect of age in accounting for performance: age did not contribute to visual–spatial searching when FA within these regions was first included in the models. Considering we found that both FA in right frontal–parietal regions and speed of visual–spatial searching increased with age, our findings are consistent with the growth of regional white matter organization as playing an important role in increased speed of visual searching with age.

Introduction

Despite recent evidence that white matter plays a critical role in physiological mechanisms of brain maturation and neural signaling (Helmuth, 2001, Reed et al., 2004, Tsuda et al., 2003, Tsuda et al., 2005, Ullian et al., 2001) the evaluation of analogous brain/behavior relations in human cognitive development is limited. Information speed is the general rate at which a person can complete cognitive operations: age-related increases in information processing speed are robust and are recognized as a mechanism of cognitive development and intellectual outcome (Kail, 2000, Kail and Park, 1994, Luciano et al., 2004). To test the functional role of white matter growth in cognitive development we examined age-related changes in white matter integrity for children and adolescents using diffusion tensor imaging, and related these to information processing speed. Documenting the connection between structural brain growth and cognitive function is a necessary first step in integrating molecular and physiological mechanisms of brain maturation with behavior change. Such integration is important for understanding how brain maturation may mediate functional change, has wide-ranging applications for brain/behavior models in neuroscience, and may ultimately yield novel information for characterizing and treating developmental neurological disorders.

White matter growth is the main source of increased brain volume during child development and continues well into the second decade for some regions (Casey et al., 2000, Giedd et al., 1999, Paus et al., 2001). Diffusion tensor imaging (DTI) provides quantitative indices of the diffusion of water within tissue and is an excellent technique for measuring age-related changes in the biological properties of white matter in vivo (Beaulieu, 2002, Pfefferbaum et al., 2000, Schmithorst et al., 2002). Such information is necessary to quantify subtle changes in white matter organization with maturation and relate those changes to behavior. Both the magnitude of water diffusion, expressed as apparent diffusion coefficient, and the directionality, expressed as the degree of anisotropy provide indices of white matter organization (Beaulieu, 2002). Decreased magnitude and increased directionality of water diffusion across multiple white matter pathways are associated with increased age in children and adolescents (Barnea-Goraly et al., 2005, Ben Bashat et al., 2005, Li, 2002, McGraw et al., 2002, Mukherjee et al., 2001, Schmithorst et al., 2002, Schneider et al., 2004, Snook et al., 2005, Suzuki et al., 2003). White matter is likely important in the ontogeny of information processing speed as it facilitates the rate of transmission of electrical signals along axons (Aboitiz et al., 1992, Schmithorst et al., 2002) which is a primary means of neural communication. Further, damage to white matter yields slow processing speed (Kail, 1998). Finally, white matter integrity is related to information processing speed in adults (Madden et al., 2004, Tuch et al., 2005). Anisotropy within right parietal and occipital hemispheric white matter predicts reaction time in young healthy adults (Tuch et al., 2005). Changes are also observed with aging: anisotropy in the splenium of the corpus callosum is related to reaction time for young adults while anisotropy in the anterior internal capsule is most relevant for older adults (Madden et al., 2004).

Although DTI has been used widely to document compromised white matter in clinical pediatric populations (Ashtari et al., 2005, Barnea-Goraly et al., 2004, Filippi et al., 2003, Khong et al., 2003, Khong et al., 2005, Khong et al., 2006, Mabbott et al., 2006, Molko et al., 2004, Peng et al., 2004), the functional implications of white matter growth for normal development have received less attention. General intelligence has been related to white matter integrity within bilateral association areas involving frontal and occipital–parietal areas (Schmithorst et al., 2005). Intelligence measures are not sufficient for explaining the development of brain/behavior relations however, as they are composite measures of multiple cognitive processes (Kail, 2000, Neisser et al., 1996). In terms of specific functions, increased anisotropy within left temporal–parietal white regions is related to proficiency in reading ability in children and adults (Beaulieu et al., 2005, Deutsch et al., 2005, Klingberg et al., 2000, Niogi and McCandliss, 2006). To account for development in examining brain/behavior relations, researchers have controlled for age when calculating correlations between DTI indices and behavioral performance (Liston et al., 2005, Nagy et al., 2004, Olesen et al., 2003, Schmithorst et al., 2005). Independent of age (a) faster reaction time in cognitive control is associated with increased organization of white matter tracts from the caudate to frontal grey matter (Liston et al., 2006) and (b) visual–spatial working memory is related to white matter organization in the anterior corpus callosum, left frontal lobe, and left temporal–occipital regions (Nagy et al., 2004, Olesen et al., 2003). Though such findings support the role of white matter maturation in the development of cognitive function they are incomplete. White matter integrity and age are correlated: because of this multi-collinearity, the use of partial correlations to control for age only may not reflect the shared variance in the model. Age is simply a surrogate index of maturation and experience.

A more robust approach is to test the unique contribution of both age and white matter integrity in predicting cognitive performance. To do this we examined the relations between age, white matter organization, and individual differences in information processing speed using hierarchical regression. We acquired DTI indices of white matter integrity for 17 children and adolescents ranging in age from 6 to 17 years. Apparent diffusion coefficient (ADC) and fractional anisotropy (FA) were calculated bilaterally for (a) large commissural and projection fiber pathways including the corpus callosum, internal capsule, and external capsule (Fig. 1), and (b) hemispheric white matter compartments including inferior frontal, frontal, frontal–parietal, temporal, parietal–occipital, and occipital regions (Fig. 2). Information processing speed was measured using visual–spatial searching, rapid picture naming, and sustained attention reaction time. First, the presence of age-related changes in white matter integrity was assessed. If developmental changes are present in white matter integrity, then decreases in ADC and increases in FA will be associated with increasing age. Second, the specificity of relations between the regions of white matter and different measures of information speed were examined. If explicit white matter pathways are related to specific cognitive functions, then ADC and FA for different brain regions should differentially predict performance on the various tasks of information processing speed. Third, for white matter regions where age-related effects were observed, we employed multiple hierarchical regression analyses to predict information processing speed and compared the relative increase in the variance accounted for by the model when age versus DTI indices was entered first. If white matter growth accounts for development of information processing speed, then DTI indices should reduce the contributions of age in predicting information processing speed.