Relationships between brain activation and brain structure in normally developing children

Abstract

Dynamic changes in brain structure, activation, and cognitive abilities co-occur during development, but little is known about how changes in brain structure relate to changes in cognitive function or brain activity. By using cortical pattern matching techniques to correlate cortical gray matter thickness and functional brain activity over the entire brain surface in 24 typically developing children, we integrated structural and functional magnetic resonance imaging data with cognitive test scores to identify correlates of mature performance during orthographic processing. Fast-naming individuals activated the right fronto-parietal attention network in response to novel fonts more than slow-naming individuals, and increased activation of this network was correlated with more mature brain morphology in the same fronto-parietal region. These relationships remained even after effects of age or general cognitive ability were statistically controlled. These results localized cortical regions where mature morphology corresponds to mature patterns of activation, and may suggest a role for experience in mediating brain structure-activation relationships.

Figure 1.

Model that guided the present study. Regions where there is overlap between activation and skill improvement reflect regions where activation intensity changes with better skill (striped). Activation of such regions is therefore defined as “mature.” Activation is then correlated with morphology. We looked for regions where “mature” activation corresponded with morphological growth trajectory (i.e., thinning in dorsal fronto-parietal regions; thickening in perisylvian regions). If there are regions of overlap (checker board), then morphological maturation is associated with maturation of functional activation.

Figure 2.

Orthographic processing task. Subjects were instructed to press a button with their right index finger if an ascender was present, and a button with their right middle finger if there was no ascender. In the examples given here, the correct response for “alarm” and its corresponding false font string was “yes,” and the correct response for “sauce” and its corresponding false font string was “no.” False font strings matched real words for length, size, and location of ascenders and descenders. Words containing the letters i and j were excluded to avoid confusion. Scan repetition time was 4 s. Ten volumes were acquired from each experimental block and 4 from each rest period. The initial 2 volumes corresponding to the instruction (“GET READY!”) were discarded to exclude measurements preceding T1 equilibrium. The experiment lasted 7 min, 52 s and yielded 116 usable volumes.

Figure 3.

Regions of interest used in the permutation analyses. Lateral regions are color coded as follows: ventral frontal, yellow; dorsal frontal, pink; temporal, dark blue; occipital, green; parietal, light blue; perisylvian, brick red (created from a statistical map published previously; Sowell, Thompson, Leonard, et al. 2004). Medial regions are color coded as follows: dorsal frontal, purple; ventral frontal, olive green; parietal, dark blue; occipital, red; callosal subcortical area (not tested in permutations), white.

Figure 4.

Functional activation of response to novel print and implicit reading. (a) An uncorrected statistical map of response to novel print (false font-word) is rendered on the cortical surface, and activation that passed correction for multiple comparisons (cluster threshold set at Z > 1.7, P = 0.05) is rendered on axial slices. Regions responding to novel print more than real words included right parietal (BA 7, 19), right frontal (BA 9, 10), right inferior temporal (BA 37), and bilateral occipital (BA 18) regions. (b) An uncorrected statistical map of implicit reading (word-false font) is rendered on the cortical surface, and activation that passed correction for multiple comparisons (cluster threshold set at Z > 1.7, P = 0.05) is rendered on axial slices. Regions responding more to real words than novel print included the left inferior frontal gyrus (BA 44, 45, 47). Left posterior temporal and inferior parietal activation (BA 22, 40) evident on the surface rendering but not on the axial slices indicated subthreshold activation in those regions.

Figure 5.

Relationships between activation, gray matter thickness, and naming speed. For all surface maps (a, c, d, and e), cortical surface points with statistically significant Pearson's r correlation coefficients are differentiated from points with nonsignificant values (gray) by color coding, and 3 different levels of statistical significance (P ≤ 0.05, 0.01, and 0.005) are rendered. Relationship between activation and gray matter thickness. (a) Negative correlations (i.e., red, orange, yellow) indicate that those with thinner cortex activated the fronto-parietal network more in response to novel print. (b) Scatter plots of 3 right parietal surface points (averaged) show that response to novel print (false font-word) is mainly driven by the false font string condition (false-rest) and not by the word condition (word-rest). Relationship between activation and naming speed. (c) Negative correlations indicate that individuals with faster-naming speed activated diverse bilateral regions in response to novel print, including the fronto-parietal attention network. After effects of age (d) and general cognitive development (e) were statistically removed, naming speed still correlated significantly with the fronto-parietal network's response to novel print. Naming speed predicted activation of the fronto-parietal attention network independently of age or IQ.

Figure 6.

Correlation between age and gray matter thickness. Cortical surface points with statistically significant Pearson's r correlation coefficients are differentiated from points with nonsignificant values (gray) by color coding, and 3 different levels of statistical significance (P ≤ 0.05, 0.01, and 0.005) are rendered. Results conform to documented pattern of cortical thinning in the dorsal fronto-parietal region among larger subject groups with more extended age ranges than in the present study; however, the present study sample did not show cortical thickening in the perisylvian region as others have reported (Sowell et al. 2003; Gogtay et al. 2004; Sowell, Thompson, Leonard, et al. 2004).