Learning-related contraction of gray matter in rodent sensorimotor cortex is associated with adaptive myelination
Curation statements for this article:-
Curated by eLife
Evaluation Summary:
This is a useful study employing a well-established one-pawed reaching/grasping paradigm for fine-motor skill learning to assess if learning is associated with cortical structural changes as assessed by longitudinal MRI measurements in mice. The authors report a non-linear time course of MRI signal changes representing a decrease in grey matter and an increase in white matter volumes in the cerebral cortex and other regions. The authors ascribe these changes to increased myelination, a conclusion that is supported by quantitative immunolabelling for the myelin protein MBP. These results represent an interesting addition to the literature around myelination changes associated with learning/activity (adaptive myelination). Additional histological analysis of changes in myelination would bolster support for the authors' conclusions.
(This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. The reviewers remained anonymous to the authors.)
This article has been Reviewed by the following groups
Listed in
- Evaluated articles (eLife)
Abstract
From observations in rodents, it has been suggested that the cellular basis of learning-dependent changes, detected using structural MRI, may be increased dendritic spine density, alterations in astrocyte volume, and adaptations within intracortical myelin. Myelin plasticity is crucial for neurological function, and active myelination is required for learning and memory. However, the dynamics of myelin plasticity and how it relates to morphometric-based measurements of structural plasticity remains unknown. We used a motor skill learning paradigm in male mice to evaluate experience-dependent brain plasticity by voxel-based morphometry (VBM) in longitudinal MRI, combined with a cross-sectional immunohistochemical investigation. Whole-brain VBM revealed nonlinear decreases in gray matter volume (GMV) juxtaposed to nonlinear increases in white matter volume (WMV) within GM that were best modeled by an asymptotic time course. Using an atlas-based cortical mask, we found nonlinear changes with learning in primary and secondary motor areas and in somatosensory cortex. Analysis of cross-sectional myelin immunoreactivity in forelimb somatosensory cortex confirmed an increase in myelin immunoreactivity followed by a return towards baseline levels. Further investigations using quantitative confocal microscopy confirmed these changes specifically to the length density of myelinated axons. The absence of significant histological changes in cortical thickness suggests that nonlinear morphometric changes are likely due to changes in intracortical myelin for which morphometric WMV in somatosensory cortex significantly correlated with myelin immunoreactivity. Together, these observations indicate a nonlinear increase of intracortical myelin during learning and support the hypothesis that myelin is a component of structural changes observed by VBM during learning.
Article activity feed
-
-
Author Response
Reviewer #1 (Public Review):
This paper describes longitudinal MRI measurements of "grey matter volume" (GMV) and "white matter volume" (WMV) in the brains of mice that were trained in a well-established one-pawed reaching/grasping paradigm for fine-motor skill learning. GMV/WMV ratio is presumed to reflect the extent to which axons in the region of interest are ensheathed by water-poor myelin membrane ("myelinated"). The conclusion is that WMV increases during learning in several task-related brain regions such as the primary motor cortex and somatosensory cortex, as well as a number of regions that are not so obviously task-related. Parallel decreases in GMV were observed. No change in overall cortical volume was detected so the conclusion is that some intra-cortical axons become myelinated in response to motor …
Author Response
Reviewer #1 (Public Review):
This paper describes longitudinal MRI measurements of "grey matter volume" (GMV) and "white matter volume" (WMV) in the brains of mice that were trained in a well-established one-pawed reaching/grasping paradigm for fine-motor skill learning. GMV/WMV ratio is presumed to reflect the extent to which axons in the region of interest are ensheathed by water-poor myelin membrane ("myelinated"). The conclusion is that WMV increases during learning in several task-related brain regions such as the primary motor cortex and somatosensory cortex, as well as a number of regions that are not so obviously task-related. Parallel decreases in GMV were observed. No change in overall cortical volume was detected so the conclusion is that some intra-cortical axons become myelinated in response to motor learning - supporting the idea of "adaptive myelination" proposed by others. Supporting histochemical evidence is provided (quantitative myelin immunolabelling). The MRI changes observed did not occur in a simple linear or cumulative fashion during learning, but rather increased in a non-linear asymptotic way, or even peaked and decreased again during training ("quadratic"). This is an interesting and useful study that takes us a little closer to understanding what is going on in the brain during learning and memory formation and continues the development of MRI as a useful non-invasive tool for studying the contribution of myelin to these processes.
Specific points:
- "Grey matter" and "white matter" are normally used to describe spatially distinct brain regions that are sparsely myelinated (grey) or heavily myelinated (white), for example, the cerebral cortex (grey) and underlying subcortical axon tracts (white). However, most or all regions are described here as white matter within the classical grey matter - within the motor cortex, for example. Classical white matter regions such as corpus callosum do not get a mention. Presumably, the authors' use of the terms grey and white matter refer to specific MRI signals that are designed to pick up relatively water-rich or water-poor domains that are presumed to reflect the abundance of myelinated versus unmyelinated fibers, not necessarily the classic anatomical grey or white matter. However, this is confusing. Is it possible to change the terminology from grey and white matter to myelin-rich and myelin-poor, water-poor and water-rich, or something similar? At the very least it requires a better explanation.
We thank this reviewer for bringing up this point and apologize for the confusion. In the revised version of the manuscript, we now present higher-magnification of the images that were used to quantify MBP immunoreactivity (densitometry) (see Main Figure 5-Supplementary Figure 3 in the revised version of the manuscript). In addition, new immunohistochemical experiments were performed and a second method was used to investigate myelinated axons within the cortex. Coronal sections were immunolabeled for myelin basic protein (MBP) and high-resolution confocal imaging was performed on a subset of trained mice (n=12 mice, n=108 probes, 9 probes per animal, represented in Main Figure 6-Supplementary Figure 1 in the revised version of the manuscript). We acquired Z-stacks with a minimum of 30 optical sections and performed an analysis of fibers based on a quantitative 3D immunohistochemical method (3D-QICH) to reconstruct and analyze length density, diameter and volumetric fraction of myelinated axons. This method of analysis of fibers was first implemented to measure vascularity (Fouard et al., 2006) which was later developed further and validated for the systematic analysis of axons (Hamodeh et al., 2010; Hamodeh et al.,2014; Hamodeh et al., 2017). The method employed for the 3D-reconstruction and analysis of myelinated axons is explained in detail in the Material and Methods section of the revised manuscript. There a significant increase in the length density of myelinated axons from baseline to experimental day 6 followed by a significant decrease towards baseline levels at experimental day 14 (one-way ANOVA, F2,7 = 8.249, P < .05; Fig. 6B), following a quadratic model rather than a linear one (AIC > 2).
Fouard, C., Malandain, G., Prohaska, S., & Westerhoff, M. (2006). Blockwise processing applied to brain microvascular network study. IEEE Trans.Med Imaging, 25(10), 1319-1328.
Hamodeh, S., Eicke, D., Napper, R. M. A., Harvey, R. J., & Sultan, F. (2010). Population based quantification of dendrites: evidence for the lack of microtubule-associate protein 2a,b in Purkinje cell spiny dendrites. Neuroscience, 170(4), 1004-1014. doi:10.1016/j.neuroscience.2010.08.021
Hamodeh, S., Sugihara, I., Baizer, J., & Sultan, F. (2014). Systematic analysis of neuronal wiring of the rodent deep cerebellar nuclei reveals differences reflecting adaptations at the neuronal circuit and internuclear level. J Comp Neurol, 522, 2481-2497.
Hamodeh, S., Bozkurt, A., Mao, H., & Sultan, F. (2017). Uncovering specific changes in network wiring underlying the primate cerebrotype. Brain Struct Funct, 222(7), 3255-3266. doi:10.1007/s00429-017-1402-6
- Several previous studies of motor learning in rodents, both MRI- and histology-based, have identified structural alterations and/or changes to oligodendrocytes and myelin in the corpus callosum underlying the motor cortex. In general, those white matter alterations were proportionally greater than those detected within the cortex itself. However, the present study apparently did not find significant MRI signal changes in sub-cortical white matter, which is surprising. Was this because the MRI sequences were not optimized for classical "white matter", or because the white matter was specifically excluded from the analysis (masked out)? If the latter, why was sub-cortical white matter excluded from the analysis? This needs discussion and explanation.
We thank this reviewer for bringing up this critical point. As mentioned above in point #4 to the Editor, significant increases in WMV were observed on the whole-brain level in many areas of WM in the brain (also see Main Figure 2-Supplemnetary Figure 3). For whole-brain analyses, all subcortical white matter regions were included in the analysis of WMV. Table 1 in the revised version of the manuscript indicate the significant changes and the direction of these changes: decreases in GMV (Main Figure 2A) and increases in WMV (Main Figure 2B). Significant changes were found in WMV, but these were not represented in the Figures originally presented. Instead, we chose to depict significant changes at PFDR corr < 0.01 for increases in WMV and PFDR corr < 0.001 for decreases in GMV, due to the high number of significant voxels at PFDR corr < 0.05, for both WMV and GMV. The Figure in point #4 to the Editor (new Main Figure-Supplementary Figure 4) depicts significant increases in WMV according to the asymptotic model at PFDR corr < 0.05. Clear changes are observed in subcortical WMV, however, we chose to present higher thresholded results (PFDR corr < 0.01) to present the more discrete clusters of increases in WMV together with the more discrete clusters of decreases in GMV at PFDR corr < 0.001.
- The quantitative MBP immunolabelling is a crucial piece of supporting evidence for the suggestion that MRI signal changes reflect adaptive myelination. What was the baseline against which immunoreactivity was measured? What did the fluorescence labelling look like at higher magnification - can individual myelin sheaths be distinguished, for example, and could these sheaths be counted, to complement and reinforce densitometry? Higher-mag images should be included in a revision.
We thank this reviewer for these questions. Baseline measurements of myelin immunoreactivity were quantified in brain sections from food-restricted mice that never underwent behavioral training, represented as experimental day 0 in Main Figure 5C, Main Figure 6A-C, Main Figure 5-Supplementary Figure 2B. We also evaluated myelin immunoreactivity in non-trained control mice; mice that were food-restricted and placed into the training cage during the 15 experimental days, yet the daily ration of food pellets was provided on the floor of the cage rather than the shelf of the training cages. These data are represented in Main Figure 5-Supplementary Figure 2A and 2B.
In the revised version of the manuscript, we have included a higher magnification image of a representative section (see below and as Main Figure 5-Supplementary Figure 3) to depict the area for which MBP-immunoreactivity was quantified. Individual myelinated axons can be appreciated in areas of cortex or striatum with limited myelinated axons. Yet due to the dense plexus of myelinated axons in cortical areas where significant VBM clusters were observed, it was not possible to identify and count individual myelinated axons within 20-micron thick histological sections using fluorescence light microscopy. To complement and reinforce our observations from MBP densitometry, we performed additional immunohistochemical labeling in subsequent coronal brain sections and used confocal laser scanning microscopy to be able to distinguish individual myelinated axons. As mentioned in answer #1 to editor, we acquired Z-stacks with a minimum of 30 optical sections and performed an analysis of fibers based on a quantitative 3D immunohistochemical method (3D-QICH) to reconstruct and analyze length density, diameter and volumetric fraction of myelinated axons. The method employed for the 3D-reconstruction and analysis of myelinated axons is explained in detail in the Material and Methods section of the revised manuscript. There a significant increase in the length density of myelinated axons from baseline to experimental day 6 followed by a significant decrease towards baseline levels at experimental day 14 (one-way ANOVA, F2,7 = 8.249, P < .05; Fig. 6B), following a quadratic model rather than a linear one (AIC > 2). This new data is now presented in Main figure 6 in the revised version of the manuscript and confirm our observations from densitometry of adaptive myelination during learning.
Reviewer #2 (Public Review):
This study uses a well-established reaching task to assess the effect of learning on cortical structures as assessed by MRI in mice. The results show a decrease in grey matter (GM) and an increase in white matter (WM) volumes that appear to peak at experimental day 8, falling slightly thereafter.
This is an interesting addition to the literature around myelination changes associated with learning/activity (adaptive myelination). However, it requires significant additional analysis. The correlation between imaging and histology is critical, but the only measure used here is MBP immunoreactivity. This is insufficient, as MBP can be expressed by newly-formed oligodendrocyte cell bodies, by their processes, and by the myelin sheath they form; but only the latter is relevant to function. So, a much more detailed analysis of oligodendrocyte morphology and myelin sheath number/size is required. This analysis needs to distinguish different layers of the cortex. This is easy for the superficial layers where myelination is sparse but much more difficult in the more heavily myelinated deeper layers. Here, counting nodes of Ranvier by Caspr immunostaining provides a good proxy. Ideally, both sheath number and sheath length would be analysed, but I accept that most studies point to number rather than changes in length as being the key changes in adaptive myelination. Then, the critical precise correlation of imaging changes with myelin sheath number can be made and the conclusion that the MRI changes represent physiologically significant changes in myelination becomes more solid.
We thank this reviewer for bringing up their suggestions to improve our manuscript. In the revised manuscript, we have now addressed which cortical layers demonstrate significant changes in GMV and WMV (new Main Figure 4 in the revised manuscript) and we have now included an additional series of experiments to further quantitate myelinated axons in somatosensory cortex for the forelimb.
We acknowledge that MBP can be expressed in newly-formed oligodendrocyte cell bodies, by their processes, and by the myelin sheath they form. For this reason, we complemented the densitometry now presented in Main Figure 5 of the revised manuscript with a confocal-based analysis of myelin sheath/myelinated axons. The latter is presented in Main Figure 6 of the revised manuscript and further supports adaptive changes in intracortical myelin during learning. Using confocal microscopy, in combination with the quantitative analysis of fibers by using a function in Amira software for fiber skeleton reconstruction, significant changes were observed in length density. In the revised discussion we have stated that changes in the length density of myelinated axons reflect both changes in length and in number, or density, of myelinated axons in somatosensory cortex for the forelimb. Our analysis also quantitated the diameter of myelinated axons, for which we observed a decrease at experimental day 6 followed by an increase in diameter at experimental day 14, albeit these changes did not reach significance. We added in the revised discussion a paragraph hypothesizing that an increase in length density combined with a putative decrease in the diameter of myelinated axons at experimental day 6 could indicate the appearance of new myelinated axons (novel candidate circuits). Afterwards, during the consolidation phase of learning, optimal candidate circuits may be selected and refined, for which putative increases in myelin sheath diameter may occur. However, to further understand changes in myelinated axons with learning, future studies should focus on a longitudinal in vivo evaluation of individual myelinated axons.
Due to the dense plexus of myelinated axons in cortical areas where significant VBM clusters were observed, these deeper layers are challenging to quantitate adaptive changes in individual myelinated axons and/or nodes of Ranvier by the use of Caspr immunoreactivity in the 20-µm thick histological sections generated by our dataset. These heavily myelinated deeper layers are also challenging to quantitate adaptive changes in myelination using, for example, longitudinal in vivo measurements by two-photon microscopy since this technique is typically limited to imaging more superficial depths (300–400 µm) of cortex. The focus of this manuscript was to demonstrate that white matter volume in somatosensory cortex significantly correlates with myelin immunoreactivity, to support the hypothesis that myelin is a component of non-linear structural changes observed by longitudinal voxel-based morphometry during learning. We are planning a future study are to determine a physiological correlate to the changes we present in this manuscript using fiber photometry and multielectrode recordings during learning.
-
Evaluation Summary:
This is a useful study employing a well-established one-pawed reaching/grasping paradigm for fine-motor skill learning to assess if learning is associated with cortical structural changes as assessed by longitudinal MRI measurements in mice. The authors report a non-linear time course of MRI signal changes representing a decrease in grey matter and an increase in white matter volumes in the cerebral cortex and other regions. The authors ascribe these changes to increased myelination, a conclusion that is supported by quantitative immunolabelling for the myelin protein MBP. These results represent an interesting addition to the literature around myelination changes associated with learning/activity (adaptive myelination). Additional histological analysis of changes in myelination would bolster support for the authors' …
Evaluation Summary:
This is a useful study employing a well-established one-pawed reaching/grasping paradigm for fine-motor skill learning to assess if learning is associated with cortical structural changes as assessed by longitudinal MRI measurements in mice. The authors report a non-linear time course of MRI signal changes representing a decrease in grey matter and an increase in white matter volumes in the cerebral cortex and other regions. The authors ascribe these changes to increased myelination, a conclusion that is supported by quantitative immunolabelling for the myelin protein MBP. These results represent an interesting addition to the literature around myelination changes associated with learning/activity (adaptive myelination). Additional histological analysis of changes in myelination would bolster support for the authors' conclusions.
(This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. The reviewers remained anonymous to the authors.)
-
Reviewer #1 (Public Review):
This paper describes longitudinal MRI measurements of "grey matter volume" (GMV) and "white matter volume" (WMV) in the brains of mice that were trained in a well-established one-pawed reaching/grasping paradigm for fine-motor skill learning. GMV/WMV ratio is presumed to reflect the extent to which axons in the region of interest are ensheathed by water-poor myelin membrane ("myelinated"). The conclusion is that WMV increases during learning in several task-related brain regions such as the primary motor cortex and somatosensory cortex, as well as a number of regions that are not so obviously task-related. Parallel decreases in GMV were observed. No change in overall cortical volume was detected so the conclusion is that some intra-cortical axons become myelinated in response to motor learning - supporting …
Reviewer #1 (Public Review):
This paper describes longitudinal MRI measurements of "grey matter volume" (GMV) and "white matter volume" (WMV) in the brains of mice that were trained in a well-established one-pawed reaching/grasping paradigm for fine-motor skill learning. GMV/WMV ratio is presumed to reflect the extent to which axons in the region of interest are ensheathed by water-poor myelin membrane ("myelinated"). The conclusion is that WMV increases during learning in several task-related brain regions such as the primary motor cortex and somatosensory cortex, as well as a number of regions that are not so obviously task-related. Parallel decreases in GMV were observed. No change in overall cortical volume was detected so the conclusion is that some intra-cortical axons become myelinated in response to motor learning - supporting the idea of "adaptive myelination" proposed by others. Supporting histochemical evidence is provided (quantitative myelin immunolabelling). The MRI changes observed did not occur in a simple linear or cumulative fashion during learning, but rather increased in a non-linear asymptotic way, or even peaked and decreased again during training ("quadratic"). This is an interesting and useful study that takes us a little closer to understanding what is going on in the brain during learning and memory formation and continues the development of MRI as a useful non-invasive tool for studying the contribution of myelin to these processes.
Specific points:
1. "Grey matter" and "white matter" are normally used to describe spatially distinct brain regions that are sparsely myelinated (grey) or heavily myelinated (white), for example, the cerebral cortex (grey) and underlying subcortical axon tracts (white). However, most or all regions are described here as white matter within the classical grey matter - within the motor cortex, for example. Classical white matter regions such as corpus callosum do not get a mention. Presumably, the authors' use of the terms grey and white matter refer to specific MRI signals that are designed to pick up relatively water-rich or water-poor domains that are presumed to reflect the abundance of myelinated versus unmyelinated fibers, not necessarily the classic anatomical grey or white matter. However, this is confusing. Is it possible to change the terminology from grey and white matter to myelin-rich and myelin-poor, water-poor and water-rich, or something similar? At the very least it requires a better explanation.2. Several previous studies of motor learning in rodents, both MRI- and histology-based, have identified structural alterations and/or changes to oligodendrocytes and myelin in the corpus callosum underlying the motor cortex. In general, those white matter alterations were proportionally greater than those detected within the cortex itself. However, the present study apparently did not find significant MRI signal changes in sub-cortical white matter, which is surprising. Was this because the MRI sequences were not optimized for classical "white matter", or because the white matter was specifically excluded from the analysis (masked out)? If the latter, why was sub-cortical white matter excluded from the analysis? This needs discussion and explanation.
3. The quantitative MBP immunolabelling is a crucial piece of supporting evidence for the suggestion that MRI signal changes reflect adaptive myelination. What was the baseline against which immunoreactivity was measured? What did the fluorescence labelling look like at higher magnification - can individual myelin sheaths be distinguished, for example, and could these sheaths be counted, to complement and reinforce densitometry? Higher-mag images should be included in a revision.
-
Reviewer #2 (Public Review):
This study uses a well-established reaching task to assess the effect of learning on cortical structures as assessed by MRI in mice. The results show a decrease in grey matter (GM) and an increase in white matter (WM) volumes that appear to peak at experimental day 8, falling slightly thereafter.
This is an interesting addition to the literature around myelination changes associated with learning/activity (adaptive myelination). However, it requires significant additional analysis. The correlation between imaging and histology is critical, but the only measure used here is MBP immunoreactivity. This is insufficient, as MBP can be expressed by newly-formed oligodendrocyte cell bodies, by their processes, and by the myelin sheath they form; but only the latter is relevant to function. So, a much more detailed …
Reviewer #2 (Public Review):
This study uses a well-established reaching task to assess the effect of learning on cortical structures as assessed by MRI in mice. The results show a decrease in grey matter (GM) and an increase in white matter (WM) volumes that appear to peak at experimental day 8, falling slightly thereafter.
This is an interesting addition to the literature around myelination changes associated with learning/activity (adaptive myelination). However, it requires significant additional analysis. The correlation between imaging and histology is critical, but the only measure used here is MBP immunoreactivity. This is insufficient, as MBP can be expressed by newly-formed oligodendrocyte cell bodies, by their processes, and by the myelin sheath they form; but only the latter is relevant to function. So, a much more detailed analysis of oligodendrocyte morphology and myelin sheath number/size is required. This analysis needs to distinguish different layers of the cortex. This is easy for the superficial layers where myelination is sparse but much more difficult in the more heavily myelinated deeper layers. Here, counting nodes of Ranvier by Caspr immunostaining provides a good proxy. Ideally, both sheath number and sheath length would be analysed, but I accept that most studies point to number rather than changes in length as being the key changes in adaptive myelination. Then, the critical precise correlation of imaging changes with myelin sheath number can be made and the conclusion that the MRI changes represent physiologically significant changes in myelination becomes more solid.
-
