Age-related decrements in cortical gyrification: Evidence from an accelerated longitudinal dataset

Cortical gyrification has been found to decrease due to aging, but thus far this has only been examined in cross-sectional samples. Interestingly, the topography of these age-related differences in gyrification follow a distinct gradient along the cortex relative to age effects on cortical thickness, likely suggesting a different underlying neurobiological mechanism. Here I examined several aspects of gyrification in an accelerated longitudinal dataset of 280 healthy adults aged 45-92 with an interval between first and last MRI session of up to 10 years (total of 815 MRI sessions). Results suggest that age changes in sulcal morphology underlie these changes in gyrification.


8
While these folding patterns change due to aging, the underlying process of this change is not 4 space using FreeSurfer's spherical surface co-registration (mris_preproc;Fischl et al., 1999).
This registration allows for vertex-wise comparisons in gyrification across participants, as 1 6 9 shown in Figure 2A.
This common surface space was then divided into 200 coronal sections, analogous to 1 7 1 the contour procedure used by Zilles et al. (1988). This allows for mean gyrification to be 1 7 2 simplified into an anterior-posterior plot, as demonstrated in Figure 2B. Note that, however, procedure that relies on 25-mm radius local region of interest. This results in a gyrification 1 7 5 index value for each vertex of the surface mesh, referred to as a local gyrification index. 1 7 6 Gyrification for each section is spatially autocorrelated with adjacent sections. To-date there 1 7 7 do not appear to be any publications that have examined the anterior-posterior gyrification 1 7 8 pattern using the FreeSurfer gyrification calculation implementation (i.e., the Schaer et al. 1 7 9 [2012] approach). Age-related changes in global gyrification were examined as the slope of decline in 1 9 2 gyrification with age. In subsequent analyses, the relative contribution of sulcal width and 1 9 3 depth, in explaining the age decrements in gyrification are evaluated using a mediation 1 9 4 analysis. 1 9 5

9 6
Global gyrification 1 9 7 Before examining the topography of gyrification, I evaluated age changes in global 1 9 8 gyrification, which itself is yet to be examined in a longitudinal dataset. The decline in 1 9 9 gyrification was estimated using a linear mixed effects model to estimate the slope (allowing 2 0 0 for random intercepts for each participant; i.e., different starting points, but a common slope). Anterior-posterior gradient in gyrification 2 1 3 The results of this analysis are shown in Figure 4. Looking from anterior to posterior, the 2 1 4 gradient gradually rises to a peak mid-way through the frontal lobe (approx. percentile 75), 2 1 5 followed by a relative plateau through the section that subtends the temporal lobe, with a 2 1 6 trough as the anterior-posterior section is increasingly represented by the parietal lobe 2 1 7 (percentile 37). A higher plateau peak subtends the parietal lobe, and then gradually declines 2 1 8 as gradient transitions into the occipital lobe (beginning from around percentile 15). 2 1 9 The middle and lower rows of Figure 4 show that while gyrification decreases 2 2 0 globally, they are most pronounced in the parietal lobe. Further examination of the surface 2 2 1 topography can better differentiate gyrification in parietal and temporal regions. These aging 2 2 2 results demonstrate that the overall distribution of gyrification does not change with age, it 2 2 3 merely diminishes in magnitude. 2 2 4 2 2 5 Topography of gyrification 2 4 1 As gyrification is calculated as the ratio of areas of the cortical surface to an enclosing 2 4 2 surface, the gyrification index is highest at the insula-as shown Figure 5A. Here I found a 2 4 3 similar pattern in the present accelerated longitudinal dataset, as shown in Figure 5B. The 2 4 4 decline in gyrification is most pronounced in the parietal lobe and posterior aspects of the 2 4 5 frontal lobe. However, on individual cortical surfaces, even examining changes over nearly a 2 4 6 decade, changes in the surfaces are only barely visible (see Figure 6). From the global gyrification and anterior-posterior analyses, it is clear that the age decreases 2 9 3 in gyrification are gradual. Moreover, the anterior-posterior analysis was designed to evaluate 2 9 4 if these age changes in gyrification were related to shifts in the underlying distribution of 2 9 5 cortical folding, but this does not appear to be the case. Instead, there is a global decrease, 2 9 6 with some regions more affected than others, as highlighted in the topography analyses. 2 9 7 Converging with prior cross-sectional studies, regions most affected by age related 2 9 8 gyrification changes are distinct from those regions affected by changes in cortical thickness.

9 9
While studies examining cross-sectional data suggest that global gyrification involves 3 0 0 a large degree of age-independent variability, the results of the longitudinal analyses here 3 0 1 show a relatively consistent age decline, distinct from differences in the y-intercept of the 3 0 2 gyrification index (i.e., Figure 3). Though changes in gyrification due to aging have not 3 0 3 previously been investigated using an anterior-posterior gradient approach, the overall published results (e.g., Schaer et al., 2008;Cao et al., 2017;Lamballais et al., 2020). As Despite numerous prior studies reporting decreases in gyrification with age, these 3 1 2 studies provide little towards explaining the underlying mechanism. While a mechanism is 3 1 3 not presented here either, the current results provide some insight into a more specific