Hello, my name is Arca Malela, I am a neurosurgery resident at the University of Pittsburgh Medical Center Department of Neurosurgery, and this is my presentation, Different Principles Govern Multiple Scales of Brain Folding. We have no disclosures to report. Here's an outline of what we will be discussing today, starting with the motivation of what we investigated in this project, our hypothesis, our methods, our findings, and a discussion of our results. Our study is primarily motivated by the question, how does the brain fold? Historically, the most common explanation put forth was that the brain folds because it is constrained by the skull, and in order to form more cortical surface area and volume, the brain folds. However, as we and others have shown, this is unlikely to be the correct explanation. The two most common explanations currently in the literature are displayed on the slide. On the right is the genetic hypothesis that differential genetic expression in the ventricular zone leads to different patterns of radial and tangential growth in cortical progenitor cells causing focal areas of growth and expansion leading to gyri and intervening areas that lead to sulci. In contrast, Tallin et al. and other authors have suggested that there is not an underlying genetic patterning, but rather it's a purely mechanical phenomenon. Tallin et al., for example, suggested that the different tissue properties in a bilayer, such as at the cortical surface, in the setting of volume expansion, leads to a tissue buckling, causing the formation of gyri and sulci, as seen in the figure on the left. However, both of these explanations essentially assume that one explanation explains all of cortical development. When we examined malformations of cortical development, we observed that different malformations exhibited different forms and different alterations in folding patterns. For example, on the right, we see a patient with alobar holoprosencephaly, in which you can observe gyri and sulci, however there is a complete absence of the formation of normal lobes of the brain, and there is not a sylvean fissure. In contrast, in the middle, we see a patient with leasencephaly, where there is the formation of normal lobes, you can see a temporal lobe and a frontal lobe, and a sylvean fissure, but there is the gross absence of the majority of sulci and gyri in the brain. On the left, there are sulci and gyri in this patient with polymicrogyra, but their configuration in nature is completely abnormal. However, you can still see a sylvean fissure. This suggests that different processes may be driving these folding patterns. Finally, we identified the fact that the sylvean fissure separates two areas that are quite close in 3D space, but are separated in terms of cortical distance. We developed something called the convergence index, which essentially compares these two numbers and demonstrated that the temporal lobe and the frontal lobe, specifically the opercula, the banks of the sylvean fissure, have significant convergence, that they start far apart from one another, but they end up close to each other. This developmental fact also explains the fact that no MCA branches cross the sylvean fissure, even though arteries routinely cross other sulci. That fact also allows us to actually split the sylvean fissure, because no vascular, no arterial branches are crossing that fissure. This collectively leads to the hypothesis that brain folding is not one process. It's actually multiple different processes that are temporally and spatially distinct. To test this hypothesis, we performed an analysis on a fetal MRI brain atlas developed from 81 subjects published by Golipour et al., ranging from weeks 21 to 38. We performed week-to-week nonlinear deformable registration to identify specific areas of volumetric growth or change, and how these related to the formation of the sulci and gyri. We segmented out these said sulci and then performed a global volumetric analysis, volumetric analysis at the level of sulci, and an overall volume expansion analysis using the Jacobian determinant. As you can see here, the Golipour atlas has a really high-resolution atlas of all weeks of gestation, from essentially the middle of the second trimester all the way to the end of gestation, and really identifies the key points of folding that are relevant for this study. In our study, we identified that the components of the intracranial space, the cerebrum, the subarachnoid space, etc., all significantly expand throughout gestation, as we can see here. We were specifically interested in comparing the subarachnoid space that invests the gyri and sulci of the brain, what we term the sulco-subarachnoid space, and then the subdural-subarachnoid space that bathes the brain. The difference is that one of these is intimately tied to the formation of sulci and gyri, whereas the other is reflective of the degree to which the brain occupies the intracranial volume. To calculate relative intracranial volume, we divided each component by total intracranial volume at each gestational age. For example, the cerebrum expands from about 55% of intracranial volume at week 21 to greater than 60%. In contrast, the total subarachnoid space decreases over that period, a decrease largely driven by a loss in volume in the subdural subarachnoid space. However, at the end of gestation, it still occupies about 20% of intracranial volume. This consists largely of CSF, suggesting that there's not an external volumetric constraint to brain growth during development. Interestingly, the sulco-subarachnoid space remains relatively constant throughout gestation, suggesting that an equal amount of sulco-subarachnoid space is created by and occupied by newly forming sulci and gyri. The growth of the cerebrum can be modeled as a logistic curve, where a clear slowdown in growth can be seen after week 34, in the absence of an external volumetric constraint, suggesting that there may be internal mechanisms that constrain the growth of the brain. Our second analysis focuses on the relationship of a given sulcus, such as the sylvian fissure of the parietal occipital sulcus, with its surrounding gyri, and understanding the relationship between their growth throughout gestation. For the sylvian fissure, we can see that the borders of the sylvian fissure, the opercula, grow significantly throughout gestation, as we can see in the blue, but the sylvian fissure itself actually shrinks over gestation in terms of relative volume throughout gestation. In contrast, for the parietal occipital sulcus, both the sulcus and the surrounding gyri increase in volume. Expressing this in a scatterplot, you can see there's a strong negative correlation between the relative volume of the sylvian fissure and its borders, whereas there's a strong positive correlation between the relative volume of the parietal occipital sulcus and its borders. And again, this suggests that there's different mechanisms of folding. Essentially, the sylvian fissure starts as a flat depression, as you can see here on the right, and then has a progressive closure where the gyri overgrow the sylvian fissure in the underlying insula. This closes the sylvian fissure and decreases its relative volume and progresses throughout gestation. In contrast, the parietal occipital sulcus and many other sulci progress from a shallow depression that then deepens over time, and both the sulcus and the surrounding gyri grow together in a correlated fashion. This is a different form of folding than that observed in the sylvian fissure. To understand the timing and localization of these different forms of folding, we used a method called Jacobian Determinant Analysis to identify focal hotspots or areas of volumetric expansion, as delineated in the red, in comparison to relative areas of shrinkage, as highlighted in the blue. So from weeks 21 to 25, specifically 23 to 25, we see focal volumetric expansion along the sylvian fissure that correspond to the closure process that we described previously on the left. In a different phenomenon, we see expansion along the cingulate sulcus on the medial surface. From weeks 25 to 34, the expansion around the sylvian fissure slowly dies down, but then we see a progressive band of volumetric expansion flanked by areas of volumetric constriction that correspond to the deepening process that we see in other sulci, and then we see this particularly from weeks 29 onward. On the medial hemisphere, we see progressive volumetric expansion along the cingulate sulcus, which morphologically looks very different than what's going on in the lateral hemisphere. Finally, from weeks 34 onwards, we don't really see clear patterning of folding, and there are a couple of hotspots here and there, but they don't correspond to clear sulci or gyri. And the same can be said about the medial hemisphere. So this analysis demonstrates that there are at least three different forms of folding that we can identify that occur at different places and times in development. On the left, we see interlobar folding, the closure process that folds and closes the sylvian fissure. In the middle, we see interlobar folding, the deepening process that forms the paradoxical sulcus, the central sulcus, and medial sulci, as we previously discussed. And then finally, on the left, we don't see a clear patterning of volumetric growth, suggesting that there may not be an underlying patterning at this phase, and maybe a purely mechanical or random phenomenon. In conclusion, a single folding mechanism cannot explain all of brain growth. There appear to be multiple spatially and temporally separated processes that fold the brain. For example, the sylvian fissure does not behave like other sulci embryologically and forms through a closure process in contrast to the deepening process observed in the paradoxipital sulcus, the calcrine sulcus, and other sulci in the brain. Our work can be used to spatially and temporally localize genetic changes and identify genetic issues that may drive malformations of cortical development, and it can be used to constrain mechanical models of the brain and challenge the assumption that a single mechanism creates all gyri and sulci in the brain. Our work was recently published in Cerebral Cortex, which we urge you all to check out. And if you have any questions, feel free to reach out. Thank you for listening to my presentation.

Arca Malela

neurosurgery resident

brain folding

cortical progenitor cells

gyri and sulci formation