See 1 citation found by title matching your search:
See comment in PubMed Commons belowCereb Cortex. 2015 Jun;25(6):1477-89. doi: 10.1093/cercor/bht333. Epub 2013 Dec 15.
Preferential detachment during human brain development: age- and sex-specific structural connectivity in diffusion tensor imaging (DTI) data.
- 1
- Department of Brain & Cognitive Sciences, Seoul National University, Seoul 151-747, South Korea School of Computing Science and Institute of Neuroscience, Newcastle University, Newcastle upon Tyne NE1 7RU, UK.
- 2
- Department of Brain & Cognitive Sciences, Seoul National University, Seoul 151-747, South Korea Department of Biomedical Engineering, Korea University, Seoul 136-703, South Korea.
- 3
- Institute of Neuroscience and Psychology, University of Glasgow, Glasgow G12 8QB, UK Department of Neurophysiology, Max-Planck Institute for Brain Research, 60438 Frankfurt a. M., Germany Ernst Strüngmann Institute (ESI) for Neuroscience in Cooperation with Max Planck Society, Deutschordenstr. 46, Frankfurt am Main, 60528, Germany.
Abstract
Human brain maturation is characterized by the prolonged development of structural and functional properties of large-scale networks that extends into adulthood. However, it is not clearly understood which features change and which remain stable over time. Here, we examined structural connectivity based on diffusion tensor imaging (DTI) in 121 participants between 4 and 40 years of age. DTI data were analyzed for small-world parameters, modularity, and the number of fiber tracts at the level of streamlines. First, our findings showed that the number of fiber tracts, small-world topology, and modular organization remained largely stable despite a substantial overall decrease in the number of streamlines with age. Second, this decrease mainly affected fiber tracts that had a large number of streamlines, were short, within modules and within hemispheres; such connections were affected significantly more often than would be expected given their number of occurrences in the network. Third, streamline loss occurred earlier in females than in males. In summary, our findings suggest that core properties of structural brain connectivity, such as the small-world and modular organization, remain stable during brain maturation by focusing streamline loss to specific types of fiber tracts.
© The Author 2013. Published by Oxford University Press.
KEYWORDS:
brain connectivity; connectome; maturation; network analysis; tractography
Figure 1.
Overall procedure. From T1-weighted images, we generated 82 regions of interests (ROIs, 34 cortical areas and 7 subcortical areas per hemisphere, on the left). From diffusion tensor images (DTI), we reconstructed streamlines using deterministic tracking (on the right). Combining 2 preprocessing steps, we constructed weighted networks, where the number of streamlines between any pair of ROIs formed the weight of an edge (fiber tract).
Cereb Cortex. 2015 Jun;25(6):1477-1489.
Figure 2.
Topological and spatial network properties. Fitted lines were drawn when there was a significant age effect (red: female, blue: male). When multiple lines were drawn, the lines are parallel unless otherwise noted. Black line represents significant age affect without a sex difference. (A) Total number of streamlines, (B) edge density, (C) streamline count in thick versus thin edges, (E) global efficiency, (F) local efficiency, and (G) streamline count in short versus long streamlines.
Cereb Cortex. 2015 Jun;25(6):1477-1489.
Figure 3.
Modular organization. (A) Modularity Q, (B) Streamline count in within- versus between-module edges, and (C) individual edge analysis (gray: intramodule edges and light gray: intermodule edges, both without changes over age; red: edges with a decreased streamline count, blue: edges with an increased streamline count; and yellow: edges with sex-specific changes). When multiple lines were drawn, the lines are parallel unless otherwise noted. A list of all changes is provided in Table for sex-specific changes and Supplementary Table S4 for age effect.
Cereb Cortex. 2015 Jun;25(6):1477-1489.
Figure 4.
Sex-specific developmental changes in individual edge analysis for male (A–C) and for female subjects (D–F), where red edges represent significant decrease, blue edges indicate significant increases over development, gray edges illustrate the tested edges that all subjects shared in common and the sex-specific changes were emphasized by the thick edges. (A and C) Sagittal views of the left hemisphere, (B and D) transverse view, and (C and F) sagittal views of the right hemisphere, of male and female brains, respectively. (A) Two edges showed age-related changes; one in the temporal lobe lost streamlines and the other edge in the occipital lobe gained streamlines. (C) An edge in the parietal lobe lost streamlines. (D). Two edges in the temporal and the occipital lobes lost streamlines. (F) Two edges in the frontal and parietal lobes lost streamlines.
Cereb Cortex. 2015 Jun;25(6):1477-1489.
Figure 5.
Sex-specific developmental changes. (A–G) Scatter plots of streamline count with relevant fitted lines. Red: female, blue: male. Upper panel: The 4 fiber tracts demonstrating age effects only for females. Lower panel: The 3 fiber tracts displaying age effects only for males. Lh, left hemisphere; rh, right hemisphere. (A) The fiber tract between lh.fusiform and lh.lateraloccipital showing a reduction of streamline counts only for females. (B). The fiber tract between lh.lingual and lh.pericalcarine with a decreased number of streamlines for females, (C). rh.medialorbitofrontal–rh.rostralanteriorcingulate, (D) rh.precuneus–rh.superiorparietal, (E) The fiber tract between lh.transversetemporal and lh.insula with a reduced number of streamlines over age only for males, (F) rh.postcentral–rh.insula, (G) lh.cuneus–lh.pericalcarine. The rate of change per year and corresponding P value is included in the figure and FDR-adjusted P values can be found in Table .
Cereb Cortex. 2015 Jun;25(6):1477-1489.
Figure 6.
Individual edge slopes representing age effect per year with FDR-adjusted confidence intervals. (A) Individual edge age effect for both genders. x-axis: indices of edges, y-axis: coefficients for age effect per year with FDR-adjusted confidence intervals. The last 2 edges with positive slopes and confidence interval ranges are the edges with an increased streamline count and the others are the fiber tracts characterized by a decreased number of streamlines. (B) Age-related sex effect. x-axis: indices of edges, First 4 edges show decreasing rate of streamline count for females and the rest 3 edges displays age effect for males, y-axis: coefficients for age effect per year with FDR-adjusted confidence intervals.
Cereb Cortex. 2015 Jun;25(6):1477-1489.
Figure 7.
(A and B) The schematic summary of the preferential reduction of thick, short, and within-module streamlines over age. (A) Location of change: 2 ellipses represent left and right hemispheres and small circles inside hemispheres indicate ROIs. Lines connecting ROIs illustrate fiber tracts between ROIs. Red lines are where the reduction of streamlines occurred; thick, short or intramodule edges were mostly affected. (B) Magnitude of change: Short, thick, or intramodule edges lost more streamlines than long, thin, or intermodule edges. x-axis: either long, thin, or intermodule streamline count (SC), y-axis: either short, thick, or intramodule SC. (C and D) Hypothetical developmental curves for males (blue) and females (red). (C) For the total streamline count based on the observation of our data (Fig. A): a longer lasting and higher peaked increase and a delayed decrease in males. (D) For individual edges: we observed sex-specific development (Fig. C), which can be explained by 3 representative cases: if the 2 curves strongly overlap they show similar decreasing patterns (case 1), if one of the curves peaks later, one curve shows a decreasing pattern while the other curve is still increasing (case 3) or simply not decreasing yet (case 2). Therefore, depending on the time scale of the developmental trajectory, males and females may show different patterns.
Cereb Cortex. 2015 Jun;25(6):1477-1489.
Publication type
MeSH terms