1. Neuroscience
Download icon

An interactive meta-analysis of MRI biomarkers of myelin

  1. Matteo Mancini  Is a corresponding author
  2. Agah Karakuzu
  3. Julien Cohen-Adad
  4. Mara Cercignani
  5. Thomas E Nichols
  6. Nikola Stikov
  1. Department of Neuroscience, Brighton and Sussex Medical School, University of Sussex, United Kingdom
  2. NeuroPoly Lab, Polytechnique Montreal, Canada
  3. CUBRIC, Cardiff University, United Kingdom
  4. Functional Neuroimaging Unit, CRIUGM, Université de Montréal, Canada
  5. Neuroimaging Laboratory, Fondazione Santa Lucia, Italy
  6. Wellcome Centre for Integrative Neuroimaging (WIN FMRIB), University of Oxford, United Kingdom
  7. Big Data Institute, University of Oxford, United Kingdom
  8. Montreal Heart Institute, Université de Montréal, Canada
Research Article
  • Cited 10
  • Views 7,425
  • Annotations
Cite this article as: eLife 2020;9:e61523 doi: 10.7554/eLife.61523

Abstract

Several MRI measures have been proposed as in vivo biomarkers of myelin, each with applications ranging from plasticity to pathology. Despite the availability of these myelin-sensitive modalities, specificity and sensitivity have been a matter of discussion. Debate about which MRI measure is the most suitable for quantifying myelin is still ongoing. In this study, we performed a systematic review of published quantitative validation studies to clarify how different these measures are when compared to the underlying histology. We analyzed the results from 43 studies applying meta-analysis tools, controlling for study sample size and using interactive visualization (https://neurolibre.github.io/myelin-meta-analysis). We report the overall estimates and the prediction intervals for the coefficient of determination and find that MT and relaxometry-based measures exhibit the highest correlations with myelin content. We also show which measures are, and which measures are not statistically different regarding their relationship with histology.

Introduction

Myelin is a key component of the central nervous system. The myelin sheaths insulate axons with a triple effect: allowing fast electrical conduction, protecting the axon, and providing trophic support (Nave and Werner, 2014). The conduction velocity regulation has become an important research topic, with evidence of activity-dependent myelination as an additional mechanism of plasticity (Fields, 2015; Sampaio-Baptista and Johansen-Berg, 2017). Myelin is also relevant from a clinical perspective, given that demyelination is often observed in several neurological diseases such as multiple sclerosis (Höftberger and Lassmann, 2018).

Given this important role in pathology and plasticity, measuring myelin in vivo has been an ambitious goal for magnetic resonance imaging (MRI) for more than two decades (MacKay et al., 1994; Rooney et al., 2007; Stanisz et al., 1999). Even though the thickness of the myelin sheath is in the order of micrometres, well beyond the MRI spatial resolution, its presence influences several physical properties that can be probed with MRI, from longitudinal and transversal relaxation phenomena to water molecule diffusion processes.

However, being sensitive to myelin is not enough: to study how and why myelin content changes, it is necessary to define a specific biomarker. Interestingly, the quest for measuring myelin has evolved in parallel with an important paradigm shift in MRI research, where MRI data are no longer treated as just ‘pictures’, but as actual 3D distributions of quantitative measures. This perspective has breathed new life into an important field of research, quantitative MRI (qMRI), that encompasses the study of how to measure the relevant electromagnetic properties that influence magnetic resonance phenomena in biological tissues (Cercignani et al., 2018; Cohen-Adad and Wheeler-Kingshott, 2014). From the very definition of qMRI, it is clear that its framework applies to any approach for non-invasive myelin quantification.

Similarly to other qMRI biomarkers, MRI-based myelin measurements are indirect, and might be affected by other microstructural features, making the relationship between these indices and myelination noisy. Assessing the accuracy of such measurements, and their sensitivity to change, is essential for their translation into clinical applications. Validation is therefore a fundamental aspect of their development (Cohen-Adad, 2018). The most common approach is based on acquiring MR data from in vivo or ex vivo tissue and then comparing those data with the related samples analyzed using histological techniques. Despite being the most realistic approach, this comparison involves several methodological choices, from the specific technique used as a reference to the quantitative measure used to describe the relationship between MRI and histology. So far, a long list of studies have looked at MRI-histology comparisons (Cohen-Adad, 2018; Laule and Moore, 2018; MacKay and Laule, 2016; Petiet et al., 2019), each of them focusing on a specific pathology and a few MRI measures.

Despite these numerous studies, there is still an ongoing debate on what MRI measure should be used to quantify myelin and as a consequence there is a constant methodological effort to propose new measures. This debate would benefit from a quantitative analysis of all the findings published so far, specifically addressing inter-study variations and prospects for future studies, something that is currently missing from the literature.

In this study, we systematically reviewed quantitative MRI-histology comparisons and we used meta-analysis tools to address the following question: how different are the modalities for myelin quantification in terms of their relationship with the underlying histology?

Results

Literature survey

The screening process is summarized in the flowcharts in Figure 1 and Appendix 1—figure 1. The keywords as reported in the appendix returned 688 results on PubMed (last search on 03/06/2020). These results included 50 review articles. From the 50 review articles, six were selected as relevant for both the topics of myelin and related MRI-histology comparisons (Cohen-Adad, 2018; Laule and Moore, 2018; Laule et al., 2007; MacKay and Laule, 2016; Petiet et al., 2019; Turner, 2019). After the assessment, 58 original research studies were considered eligible, as shown in Appendix 1—table 1 (in the appendix) and Figure S2. All the data collected are available in the supplementary materials (Source data 1).

Sankey diagram representing the screening procedure (PRISMA flow chart provided in the appendix).

In terms of specific modalities, the survey shows that the most common MRI approach compared with histology was diffusion-weighted imaging (used in 28 studies), followed by magnetization transfer (MT, 27 studies), T2 relaxometry (19 studies) and T1 relaxometry (10 studies). Only 20 studies considered more than one approach: among the others, 20 focused exclusively on diffusion, 12 on MT, and six on T2 relaxometry.

From these 58 studies, we then focused only on brain studies and we further excluded studies not reporting either the number of subjects or the number of ROIs per subject. We also excluded one single-subject study that relied on voxels as distinct samples, whereas the other studies in this review are based on ROIs (i.e. including more than one voxel). In the end, 43 suitable studies were identified for the subsequent analyses.

Meta-analysis

To compare the studies of interest, we first organized them according to the MRI measure used. Figure 2 and Figure 3 (and also Figure S3-S4) show the R2 values for the selected studies across measures: the highest values (R2 >0.8) are obtained mostly from MT measures, but they are associated with small sample sizes (with an average of 32 sample points). The studies with largest sample sizes are associated with R2 values between 0.6 and 0.8 for MT and T2 relaxometry, but with lower values for T1 relaxometry and other approaches.

Bubble chart of R2 values between a given MRI measure and histology for each study across MRI measures, with the area proportional to the number of samples.
Treemap chart of the studies considered for the meta-analysis, organized by MRI measure.

The color of each box represents the reported R2 value while the size box is proportional to the sample size. To see the interactive figure: https://neurolibre.github.io/myelin-meta-analysis/02/closer_look.html#figure-4.

To combine the results for each measure, we then used a mixed-effect model: in this way we were able to express the overall effect size in terms of a range of R2 values within a confidence interval, but also to assess prediction intervals and inter-study differences. The results are shown as forest plots in Figure 4 (and also Figure S5).

Forest plots showing the R2 values reported by the studies and estimated from the mixed-effect model for each measure.

The hourglasses and the dotted lines in the mixed-effect model outcomes represent the prediction intervals. To see the interactive figure: https://neurolibre.github.io/myelin-meta-analysis/03/meta_analysis.html#figure-5.

Apart from MPF and MWF, all the measures showed R2 overall estimates in the range 0.21–0.53. To investigate the significance of the differences between measures, we conducted a repeated measures meta-regression on every R2 estimate recorded (98 in total over 43 studies). As shown in Figure 5 (and also Figure S6), the measures can be roughly subdivided in two groups: MT- and relaxometry-based measures gave significantly higher R2 estimates compared to diffusion-based measures. Within the diffusion-based measures, FA shows slightly higher estimates than the others, with marginal significance over RD and AD or no significance in case of MD.

Results from the repeated measures meta-regression, displayed in terms of z-scores (left) and p-values (right) for each pairwise comparison across all the MRI measures.

In the z-score heatmap, each element refers to the comparison between the measure on the x axis with the one on the y axis. For example, MPF and FA (z-score = 7.14; p-value<0.0001) are statistically different, while MPF and T1 (z-score = 2.51; p-value=0.43) are not statistically different. To see the interactive figure: https://neurolibre.github.io/myelin-meta-analysis/03/meta_analysis.html#figure-6.

Within MT- and relaxometry-based measures, the trends follow those in the forest plots (Figure 4), but most differences are not significant (Figure 5). However, the results in terms of z-score give a measure of distance between the R2 distributions. From this perspective, MPF has higher R2 estimates compared to all the other measures, but it is only marginally higher than MWF (z-score = 0.77; p-value=1) so we cannot claim that one is superior to the other. Following the same reasoning, MTR and T1 are not statistically different (z-score = 0.47; p-value=1).

When considering the prediction intervals calculated using τ2 (the variance of the effect size parameters across the population of studies), for most measures the interval spanned from 0.1 to 0.9 (Figure 4 and Figure S5). This implies that future studies relying on such measures can expect, on the basis of these studies, to obtain any R2 value in this broad interval. The only exceptions were MPF (0.49–1) and MWF (0.45–0.95), whose intervals were narrower than the alternatives. Finally, I2 (a measure of how much of the variability in a typical study is due to heterogeneity in the experimental design) was generally quite high (Table 1). MWF showed the lowest I2 across measures (I2 = 73.19%), but this may be misleading considering that it was based on only four studies, while the other measures included around 10 studies. Excluding MWF, MPF also showed a relatively low I2 (I2 = 83.18%). Qualitative comparisons across experimental conditions and methodological choices highlighted differences across pathology models, targeted tissue types and reference techniques (Figure 6 and Figure S7). Other factors such as magnetic field, co-registration, specific tissue and the related conditions (Figure S8) showed comparable distributions.

Experimental conditions and methodological choices influencing the R2 values (top: reference techniques; middle: pathology model; bottom: tissue types).
Table 1
Results from the mixed-effect models: for each measure we reported the number of studies, the estimate and standard error of the overall R2 distribution, the τ2 and the I2.
MeasureNumber of studiesEstimateStandard errorTau2I2
MTR160.5080.06910.0796.03%
MPF100.76570.04550.012883.18%
FA170.37660.06630.065287.49%
RD150.33640.06790.061592.30%
MD120.26390.06790.04487.35%
T180.53210.06920.032886.51%
AD90.20950.08020.04897.69%
T270.39380.10230.065184.49%
MWF40.69970.04320.004173.19%

Discussion

Indirect measures are the most popular (for better or worse)

The literature survey offers an interesting perspective on popular research trends (Figure S2). The first consideration one can make is that every myelin imaging technique achieves myelin sensitivity through different means. A clear example is offered by the two most common approaches in this meta-analysis, DWI and MT: the MT effect is driven by saturation pulses interacting with myelin macromolecules that transfer their magnetization to water, whereas in diffusion experiments myelin is just not part of the picture. Diffusion acquisitions are blind to direct myelin measurement because the TEs used are too long (~100 ms) to be influenced by the actual macromolecules – with T2 of ~ 10 us (Stanisz et al., 1999) – or even the water molecules trapped in the myelin sheath – with T2 of ~ 30 ms (MacKay et al., 1994). To infer myelin content, one needs to rely on the interaction between intracellular and extracellular water compartments. The majority of diffusion studies included in this analysis used tensor-based measures (with fractional anisotropy being the most common), but some also used kurtosis-based analysis. The main issue with this approach is that other factors affect those measures (Beaulieu, 2002; Beaulieu, 2009), making it difficult to specifically relate changes in water compartments to changes in myelin.

Despite this issue, the use of diffusion as a proxy for myelin is quite widespread, specifically outside the field of quantitative MRI. This is probably a consequence of how popular DWI has become and how widely available are the related acquisition sequences. MT, the second most popular technique for quantifying myelin, estimates myelin by acquiring data with and without saturating the macromolecular proton pool. The simplest MT measure, MT ratio (MTR), incorporates non-myelin contributions in the final measurement. Recent acquisition variations include computing MTR from acquisitions with ultra-short echo times (Du et al., 2009; Guglielmetti et al., 2020; Wei et al., 2018) or relying on inhomogeneous MT (Duhamel et al., 2019; Varma et al., 2015). More complex experiments, for example quantitative MT, are based on fitting two compartments to the data, the free water and the macromolecular compartments, or pools. In this way, one is able to assess myelin through MPF with higher specificity, although still potentially including contributions from other macromolecules. Additional measures have also been considered (including the T2 of each pool, the exchange rate between the pools). The drawback of qMT is the requirement for a longer and more complex acquisition. Recently, there have been alternative techniques to estimate only MPF, resulting in faster acquisitions with similar results (Khodanovich et al., 2019; Khodanovich et al., 2017; Yarnykh, 2012). Despite being focused on macromolecular contributions, these approaches are not strictly specific to myelin (Sled, 2018): in this sense, an important limitation is that MT effects are sensitive to the pH of the targeted tissue and therefore changes in the pH (caused for example by inflammation processes) will affect MT-based measures of myelin (Stanisz et al., 2004).

Following diffusion and MT, the most popular approach is T2 relaxometry. Unlike diffusion and MT, in T2 relaxometry experiments one can directly observe the contribution from the water trapped between the myelin bilayers, and can therefore estimate the myelin water fraction. A simpler but less specific approach consists in estimating the transverse relaxation time considering the decay to be mono-exponential. A historical and practical drawback of these approaches is that they require longer acquisitions, although faster alternatives have been developed (Does and Gore, 2000; Prasloski et al., 2012). A more subtle but nevertheless important limitation lies in the multi-compartment model used in multi-exponential T2 relaxometry (Does, 2018): this model generally assumes slow water exchange between compartments, but it has been showed that water exchange actually contributes to T2 spectra variations (Dula et al., 2010; Harkins et al., 2012).

Finally, other studies used a diverse collection of other measures, including T1 relaxometry, apparent transversal relaxation rate (R2*), proton density (PD), macromolecular tissue volume (MTV), relaxation along a fictitious field (RAFF), and quantitative susceptibility mapping (QSM).

After this general overview, it is clear that each modality could be a suitable candidate for a quantitative myelin biomarker. To then make a choice informed by the studies here reported, it becomes necessary to consider not only effect sizes in terms of correlation, but also sample sizes and acquisition times.

There is no myelin MRI measure true to histology

When looking at the R2 values across the different measures, the first detail that catches one’s eye is how most measures present a broad range of values (Figure 2 and Figure 3). When taking into account the sample size, the largest studies show higher correlations for MT and T2 relaxometry studies than any other approach (Figure S3 and Figure S4). In quantitative terms, the meta-analysis corroborates this idea, showing that MPF and MWF tend to be more specific to myelin compared to the other measures (respectively with R2 = 0.7657 and R2 = 0.6997), in line with the underlying theory. Notably, diffusion-based measures show the lowest overall estimates (with values between R2 = 0.3766 for FA and R2 = 0.2095 for AD): this could be due to the fact, as already mentioned, that DWI does not specifically measure myelin properties, and despite FA and RD being influenced by the myelin content, they are also influenced by other factors that make them unsuitable as measures of myelin. The repeated measure meta-regression confirms this overall picture, clearly distinguishing MT- and relaxometry-based measures from diffusion-based ones (Figure 5).

Despite these considerations on the advantages of MPF and MWF, one should refrain from concluding that they are the ‘true’ MRI measures of myelin. The reason for this caution is given not by the overall effect sizes observed here, but by the collateral outcomes of the meta-analysis. The first one is given by the prediction intervals: most measures exhibit large intervals (Figure 4), not supporting the idea of them being robust biomarkers. MPF and MWF seem to be again the most suitable choices for future studies, but a range between 0.5 and 1 is still quite large.

The second important aspect to consider is given by the differences across studies: the meta-analysis showed how such differences strongly limit inter-study comparisons for a given measure (Figure 6). This result should be expected, given that the studies here examined are inevitably influenced by the specific experimental constraints and methodological choices. Given the limited number of studies, it is not possible to quantitatively study interactions between MRI measures and the other factors (e.g. modality used as a reference, tissue types, magnetic field strength). For further qualitative insights, we invite the reader to explore the interactive figures S7-S8. A first important factor to consider is the validation modality used as a reference, which will be dictated by the equipment availability and cost. However, such a choice has an impact on the actual comparison: histology and immunochemistry, despite being specific to myelin, do not offer a volumetric measure of myelin, but rather a proxy based on the transmittance of the histological sections. So far, the only modality able to give a volumetric measure would be electron microscopy, which is an expensive and resource-consuming approach. Also, electron microscopy has several limitations, including tissue shrinkage, degradation of the myelin sheath structure due to imperfect fixation, imperfect penetration of the osmium stain, polishing, keeping focus over large imaging regions. All these effects contribute to the lack of precision and accuracy when quantifying myelin content with EM-based histology (Cohen-Adad, 2018). Another important observation is that none of the studies here reviewed considered histology reproducibility, which is hard to quantify as a whole given that a sample can be processed only once: collateral factors affecting tissue processing (e.g. sectioning distortions, mounting and staining issues) constitute an actual limitation for histology-based validation. A further example of influential factor often dictated by equipment availability is the magnetic field strength of the MRI scanner: figure S8 shows that most studies were conducted at 7T and 9.4T, with some pioneering studies at 1.5T and even fewer ones at other field strengths.

In addition to differences in experimental and methodological designs, there are also several considerations that arise out of the lack of shared practices in MRI validation studies. The first evident one is the use of correlations: despite being a simple measure that serves well the purpose of roughly characterizing a relationship, Pearson correlation is not the right tool for quantitative biomarkers, as it does not characterize the actual relationship between histology and MRI. Linear regression is a step forward but has the disadvantage of assuming a linear relationship. Despite Pearson correlation and linear regression being the most common measures used in the studies here reviewed, it is still not clear if the relationship is actually linear. Only one study among the considered ones computed both Pearson and Spearman correlation values (Tardif et al., 2012), and reported higher Spearman correlations, pointing out that non-linear relationships should actually be considered. One last consideration regarding the use of correlation measures for validating quantitative biomarkers is about the intercept in the MRI-histology relationship. Notably, only MWF is expected to assume a value equal to zero when myelin is absent (West et al., 2018). For the other measures, it would be necessary to estimate the intercept, which leads to the calibration problem in the estimate of myelin volume fraction. Notably, calculating Pearson correlation does not provide any information for such calibration. Another arbitrary practice that would benefit from some harmonization is the choice of ROIs. The studies reported here examined a diverse list of ROIs, in most cases hand-drawn on each modality, encompassing different types of tissue, and the most common approach is to report a single, pooled correlation. This is problematic, as different types of tissue (e.g. grey matter and white matter) will show different values for MRI-based measures but also for histology-based ones, making linearity assumptions about the two modalities. However, with this approach gross differences between tissues drive the observed correlation, without actually showing if the MRI-based measure under analysis is sensitive to subtle differences and therefore a suitable quantitative biomarker for myelin. The effect of considering different types of tissues is showed in Figure 6 and Figure S7, where correlation ranges change when considering different types of tissue. However, the large correlation range in white matter, the most common tissue studied, suggests that other factors also affect the correlation.

It should be clear at this point that any debate about a universal MRI-based measure of myelin is pointless, at least at the moment, as the overall picture provided by previous studies does not point to any such ideal measure. Nevertheless, is debating about a universal measure helpful for future studies?

Better biomarkers require more reproducibility studies

We hope this meta-analysis convinces the reader that a holy grail of myelin imaging does not exist, at least as long as we consider histology to be the ground truth. Given that we all have to pick our poison, the upside is that measures based on MT and relaxometry are not statistically different, and therefore, future studies have an actual choice among candidate measures. For further progress, rather than debating about a perfect measure, we would argue that what is missing at the moment is a clear picture of what can be achieved with each specific MRI modality. The studies examined here focus on a large set of different measures, and more than half of them considered at most two measures, highlighting how the field is mostly focused on formulating new measures. While it is understood that novel measures can provide new perspectives, it is also fundamentally important to understand the concrete capabilities and limitations of current measures. From this meta-analysis, what the literature clearly lacks is reproducibility studies, specifically answering two main questions: (1) what is the specificity of each measure? We should have a practical validation of our theoretical understanding of the relevant confounds; (2) what is the ‘parameter sensitivity’ of each measure? Here, we refer to parameter sensitivity in a broad sense, that includes also experimental conditions and methodological choices. The results here presented show how certain conditions (e.g. pathology) seem to affect the coefficient of determination more than others but given the limited number of studies for each modality, we refrained from additional analyses to avoid speculation. A warning message that is evident from these results is the inherent limitation of DWI for estimating myelin content: this is not by any means a novel result (Beaulieu, 2002; Beaulieu, 2009), but it is nevertheless worth reiterating given the outcomes of our analysis. If estimating myelin content is relevant in a diffusion study, it is important to consider complementing the diffusion measure with one of the modalities here reviewed; in this way, it would be possible to decouple the influence of myelin content from the many other factors that come into play when considering diffusion phenomena.

Finally, an important factor to take into account when choosing a biomarker of myelin is the actual application. For animal research, long acquisitions are not a major issue. However, when considering biomarkers for potential clinical use, the acquisition time can become a relevant issue. An example is the well-established multi-echo spin-echo implementation of MWF, that can only be used for a specific slice in a hypothetical clinical scenario. Faster techniques have been proposed for estimating it with gradient- and spin-echo (GRASE) sequences (Does and Gore, 2000; Feinberg and Oshio, 1991; Prasloski et al., 2012). Even in this case, the acquisition time still reaches 15 min for acquiring roughly the whole brain with an isotropic resolution of 2 mm. Complex MT acquisitions such as qMT suffer from the same problem, although it is possible to use optimized and faster protocols to focus specifically on MPF (Khodanovich et al., 2019; Khodanovich et al., 2017; Yarnykh, 2012).

Conclusions

Several MRI measures are sensitive to myelin content and the current literature suggests that most of them are not statistically different in terms of their relationship with the underlying histology. Measures highly correlated with histology are also the ones with a higher expected specificity. This suggests that future studies should try to better address how specific each measure is, for the sake of clarifying suitable applications.

Materials and methods

Review methodology

Request a detailed protocol

The Medline database (https://pubmed.ncbi.nlm.nih.gov) was used to retrieve the articles. The keywords used are specified in the appendix. We followed the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines for record screening and study selection. The results were first screened to remove unrelated work. Specifically we discarded: work relying only on MRI; work relying only on histology or equivalent approaches; work reporting only qualitative comparisons. After this first screening, the remaining papers were assessed. At this stage, we discarded: studies using MRI-based measures in arbitrary units (e.g. T1-weighted or T2-weighted data); studies using measures of variation in myelin content (defined either as the difference between normal and abnormal myelin content) either for MRI or for histology; studies using arbitrary assessment scales; studies comparing MRI-based absolute measures of myelin with histology-based relative measures (e.g. g-ratio); studies reporting other quantitative measures than correlation or R2 values; studies comparing histology from one dataset and MRI from a different one. As an additional source for potential candidate studies, we screened the review articles in the initial results, and we selected the relevant studies that were not already present in the studies already selected.

From the final papers, we collected first the following details: the DOI; which approach was used (diffusion, MT, T1 relaxometry, T2 relaxometry, or other); which specific MRI measures were compared to histology or equivalent techniques; the magnetic field; the technique used as a reference (histology, immunochemistry, microscopy, electron microscopy); the focus of the study in terms of brain, spinal cord or peripheral nerve; if the subjects were humans or animals, and if the latter which animal; if the tissue under exam was in vivo, in situ or ex vivo, and in the latter case if the tissue was fixed or not; if the tissue was healthy or pathological, and if the latter which pathology; the specific structures examined for correlation purposes; which comparison technique was used (e.g. Pearson correlation, Spearman correlation, linear regression); the number of subjects; the number of ROIs per subject; the male/female ratio; if registration procedures were performed to align MRI and histology; in case of pathological tissue, if control tissue was considered as well; other relevant notes. If before calculating the correlations the data were averaged across subjects, the number of subjects was considered to be one. The same consideration was made for averaging across ROIs. This is because the numbers of subjects and ROIs were used to take into account how many sample points were used when computing the correlation. We set each of those numbers to one for all the studies where the data were averaged respectively across subjects and across ROIs. Finally, in those cases where the number of ROIs or the number of subjects were given as a range rather than specific values, we used the most conservative value and added the related details to the notes.

We then proceeded to collect the quantitative results reported for each measure and for each study in the form of R2. Given that different studies may rely on a different strategy when reporting correlations, we adopted the following reasoning to limit discrepancies across studies while still objectively representing each of them. In case of multiple correlation values reported, for our analysis we selected the ones referring to the whole dataset and the entire brain if available, and considering each ROI in a given subject as a sample if possible; if only correlation values for specific ROIs were reported, the one for the most common reported structure would be chosen. In the case of multiple subjects, if data were provided separately for each group, the correlation for the control group was used. When different comparison methods were reported (e.g. both Pearson and Spearman correlation) or if the MRI data was compared with multiple references (e.g. both histology and immunohistochemistry), the correlations used were chosen on the basis of the following priority orders (from the most preferable to the least): for multiple comparison methods, linear regression, Spearman correlation, Pearson correlation; for multiple reference techniques, electron microscopy, immunohistochemistry, histology. Finally, in any other case where more than one correlation value was available, the most conservative value was used. Any other additional value was in any case mentioned in the notes of the respective study.

Meta-analysis

Request a detailed protocol

For the quantitative analysis, we restricted our focus on brain studies and only on the ones providing an indication of both the number of subjects and the number of ROIs. For each study, we computed the sample size as the product between the number of subjects and the number of ROIs per subject. In this way, we were able to compare the reported R2 values across measures taking into account the related number of points actually used for correlation purposes. We note that correlation or regression analyses run on multiple ROIs and subjects represents a repeated measures analysis, for which the degrees of freedom computation can be complex; however, most papers neglected the repeated measures structure of the data and thus the sample size computation here represents a very approximate and optimistic view of the precision of each R2 value.

To estimate the variance of each R2 value, we relied on the correlation properties and the delta method (Lehman, 1999). Let us consider the Pearson’s correlation r of two variables X and Y with population correlation ρ. If r is calculated from N random samples, the sampling variance is (1-ρ2)2/N. Applying the delta method, we then approximated the variance of R2 as 4 R2(1 R2)2/N, assuming R2≈ρ2. As we recognise that some papers computed Spearman correlation, this calculation is again optimistic and may underestimate the sampling variability of the squared Spearman correlation.

To estimate the overall effect size in terms of R2, we have to choose how to model the distribution of true effects given by the data collected from the literature. The two most common approaches are fixed-effects and mixed-effects models. While the underlying mathematical model is the same as the one used for linear regression (more details in the appendix), the assumptions are different: fixed-effects models assume that all the studies share a common effect size, while mixed-effects models assume that the effect size across studies is similar but not identical (Raudenbush, 2009). In our case, as the studies have several factors that influence the R2 values (e.g. histology/microscopy reference, magnetic field strength, pathology model), we expect a distribution of effect sizes due to inter-study differences. This is why we proceeded to fit a mixed-effects model to each measure that was featured in more than two studies. Apart from the effect size distributions, we reported two additional measures, I2 and τ2: the former expresses as a percentage how much of variability in a typical study is due to heterogeneity (i.e. the variation in study outcomes between studies) rather than chance (Higgins and Thompson, 2002), while the latter can be used to calculate the prediction interval (Raudenbush, 2009), which gives the expected range for the measure of interest in future studies. We used forest plots to represent the outcomes, and both the mixed effects estimate of the population estimated R2, with both a 95% confidence and a (larger) 95% prediction interval.

For the explicit purpose of comparing the effect sizes between different MRI measures, we conducted a repeated measures meta-regression on every R2 value recorded. We associated each R2 value with three additional details: (i) the related variance, as done in the measure-specific mixed-effects models; (ii) the related study, used as the random intercept (i.e. random variable) to incorporate potential inter-study variability; and (iii) the related MRI measure, used as the moderator (i.e. categorical variable) to estimate the differences between measures. In this way, the meta-regression leads to R2 intervals for each MRI measure, with the same trend as measure-specific mixed-effects models but with subtle differences. This is because the meta-regression makes two additional assumptions: first, R2 estimates within the same study share the same random effects and second, the between-study variance is the same for all observations. We then used the meta-regression R2 estimates to compute every possible pairwise comparison between MRI measures and to identify significantly different pairs using Tukey's test, while controlling the error rate over all the possible comparisons (Bonferroni correction).

This additional model is necessary, as direct comparisons are not possible with measure-specific analyses. While the repeated measures meta-regression makes direct comparisons straightforward, we reported the main R2 estimates based on the measure-specific mixed-effects models, as they make weaker assumptions.

For visual comparisons, we used the Jupyter notebook provided in the supplementary materials. For model fitting, we used the Metafor package, version 2.4–0 (Viechtbauer, 2010).

Appendix 1

Search keywords

(myelin[Title/Abstract] AND ((magnetic[Title/Abstract] AND resonance[Title/Abstract]) OR mr[Title/Abstract] OR mri[Title/Abstract])) AND (histology[Title/Abstract] OR histopathology[Title/Abstract] OR microscopy[Title/Abstract] OR immunohistochemistry[Title/Abstract] OR histological[Title/Abstract] OR histologically[Title/Abstract] OR histologic[Title/Abstract] OR histopathological[Title/Abstract] OR histopathologically[Title/Abstract] OR histopathologic[Title/Abstract]).

Results obtained from the Medline database: 688 (03/06/2020).

Appendix 1—figure 1
PRISMA flowchart for the meta-analysis.

Fixed- and mixed-effects models

While a traditional linear regression model estimates the error variance from residuals, in a fixed effects meta-analysis model, each paper’s response and standard errors, as well as the error variance of the regression model can be directly computed from the supplied response standard deviations. Specifically, for a (non-meta) regression model we have the i-th response yi modeled with covariate values xi, yi=xiβ+εi, where random error has unknown variance Varεi=σ2. In a fixed-effects meta-analysis, we are given yi but also si, the standard error of yi, and the regression model has the same form except the variance is known, Varεi=si2, and the weighted least squares regression can be computed, estimating beta and its standard error. A mixed-effects meta-analysis accounts for more variance than what can be ascribed to the sampling error of the reported outcome. The regression model has again the same form, except now the variance is Varεi=si2+τ2, the sum of the reported squared standard error and the unknown between-study variance τ2. Iterative methods are used to estimate τ2 and, once estimated, a weighted least squares regression can be computed. The parameter τ2 can be interpreted as the variance of noise-free (hypothetical, zero standard error) results from the population of all possible studies. The importance of τ2 can also be gauged by I2, the proportion of variance due to random inter-study differences (i.e. 1-I2 is the proportion attributable to random sampling error of each study) (Higgins and Thompson, 2002).

Abbreviations and mathematical symbols

  • AD – axial diffusivity

  • AK – axial kurtosis

  • AWF – axonal water fraction

  • FA – fraction anisotropy

  • ihMTR – inhomogeneous magnetization transfer ratio

  • k_fm – free water-macromolecular exchange rate

  • k_mf – macromolecular-free water exchange rate

  • M0m – macromolecular pool magnetization fraction

  • MD – mean diffusivity

  • MK – mean kurtosis

  • MPF – macromolecular pool fraction

  • MT – magnetization transfer

  • MTR – magnetization transfer ratio

  • MTR-UTE – magnetization transfer ratio (using ultra-short echo time)

  • MTV – macromolecular tissue volume

  • MVF-MT – myelin volume fraction (estimated from MT)

  • MVF-T2 – myelin volume fraction (estimated from T2)

  • MWF – myelin water fraction

  • PD – proton density

  • PN – peripheral nerve

  • PRISMA – Preferred Reporting Items for Systematic Reviews and Meta-Analyses

  • QSM – quantitative susceptibility mapping

  • R1f – free water pool longitudinal relaxation rate

  • R2* – apparent transverse relaxation rate

  • RAFF – relaxation along a fictitious field

  • RD – radial diffusivity

  • RD-DBSI – radial diffusivity (from diffusion basis spectrum imaging)

  • RDe – extra-cellular compartment radial diffusivity

  • RK – radial kurtosis

  • rSPF – relative semi-solid proton fraction

  • SC – spinal cord

  • T1 – longitudinal relaxation time

  • T1p – adiabatic longitudinal relaxation time

  • T1sat – longitudinal relaxation time under magnetization transfer irradiation

  • T2 – transverse relaxation time

  • T2f – free water pool transverse relaxation time

  • T2int – transverse relaxation intermediate component

  • T2m – macromolecular pool transverse relaxation rate

  • T2p – adiabatic transverse relaxation time

Appendix 1—table 1
Selected studies for qualitative analysis.
StudyMRI measure(s)Histology/microscopy measureTissueConditionFocus
Schmierer et al., 2004T1, MTRHistology - LFBHumanMultiple sclerosisBrain
Odrobina et al., 2005T1, T2, T2int, MWF, M0m, MTRMicroscopy - Myelin fractionAnimal - RatDemyelination - TelluriumPN
Pun et al., 2005T1, T2int, MWFMicroscopy - Myelin fractionAnimal - RatDemyelination - TelluriumPN
Laule et al., 2006MWFHistology - LFBHumanMultiple sclerosisBrain
Schmierer et al., 2007aT1, MTR, MPF, T2mHistology - LFBHumanMultiple sclerosisBrain
Schmierer et al., 2007bFA, MDHistology - LFBHumanMultiple sclerosisBrain
Jito et al., 2008FAMicroscopy - Myelin sheath areaAnimal - MouseHealthyBrain
Kozlowski et al., 2008MWF, FA, AD, RD, MDImmunohistochemistry - MBPAnimal - RatInjury - Dorsal columnar transectionSC
Laule et al., 2008MWFHistology - LFBHumanMultiple sclerosisBrain
Schmierer et al., 2008T1, T2, MTR, MPF, MD, FA, AD, RDHistology - LFBHumanMultiple sclerosisBrain
Wu et al., 2008T2Histology - LFBAnimal - MouseDemyelination - CuprizoneBrain
Zaaraoui et al., 2008MTRImmunohistochemistry - MBPAnimal - MouseDemyelination - CuprizoneBrain
Takagi et al., 2009FA, ADEM - Myelin thicknessAnimal - RatDegeneration - Contusive injuryPN
Wang et al., 2009FA, RDHistology - LFBAnimal - RatIschemia - Induced hypoxiaBrain
Zhang et al., 2009RDHistology - LFBAnimal - RatInjury - Dorsal columnar transectionSC
Schmierer et al., 2010MTR, T2Histology - LFBHumanMultiple sclerosisBrain
Fatemi et al., 2011MTRImmunohistochemistry - MBPAnimal - MouseIschemia - Induced hypoxiaBrain
Laule et al., 2011MWFImmunohistochemistry - MBPHumanMultiple sclerosisBrain
Underhill et al., 2011MPFHistology - LFBAnimal - MouseHealthyBrain
Chandran et al., 2012FA, RDImmunohistochemistry - MBPAnimal - MouseDemyelination - CuprizoneBrain
Tardif et al., 2012T1, T2, MTR, PDImmunohistochemistry - MBPHumanMultiple sclerosisBrain
Fjær et al., 2013MTRImmunohistochemistry - PLPAnimal - MouseDemyelination - CuprizoneBrain
Harkins et al., 2013MWF, MPFMicroscopy - Myelin fractionAnimal - RatEdema - HexacloropheneSC
Janve et al., 2013MPF, R1a, k_ba, FA, RD, MD, ADHistology - LFBAnimal - RatDemyelination - LipopolysaccharideBrain
Thiessen et al., 2013MPF, R1f, k_fm, k_mf, T2f, T2m, MD, RD, AD, FA, T1, T2EM - Myelin thicknessAnimal - MouseDemyelination - CuprizoneBrain
Kozlowski et al., 2014MWFImmunohistochemistry - MBPAnimal - RatInjury - Dorsal columnar transectionSC
Wang et al., 2014RD, RD-DBSIImmunohistochemistry - MBPAnimal - MouseDemyelination - Autoimmune encephalomyelitisSC
Fjær et al., 2015MTRImmunohistochemistry - PLPAnimal - MouseDemyelination - Autoimmune encephalomyelitisBrain
Seehaus et al., 2015FA, RD, MDHistology - SilverHumanHealthyBrain
Turati et al., 2015MPFImmunohistochemistry - MBPAnimal - MouseDemyelination - CuprizoneBrain
Wang et al., 2015RD-DBSIHistology - LFBHumanMultiple sclerosisSC
Aojula et al., 2016FA, AD, RD, MDImmunohistochemistry - MBPAnimal - RatHydrocephalusBrain
Hakkarainen et al., 2016T1, T2, MTR, T1p, T2p, RAFFHistology - Gold chlorideAnimal - RatHealthyBrain
Jelescu et al., 2016RD, RK, AWF, Rde, T2, MTREM - Myelin fractionAnimal - MouseDemyelination - CuprizoneBrain
Kelm et al., 2016MD, RD, MK, RK, AWFEM - Myelin fractionAnimal - MouseDemyelination - KnockoutBrain
Reeves et al., 2016T1, T2Immunohistochemistry - MBPHumanEpilepsyBrain
Tu et al., 2016FA, AD, RD, MD, MTRImmunohistochemistry - MBPAnimal - RatTraumatic brain injuryBrain
Chang et al., 2017FA, AD, RD, MDImmunohistochemistry - MBPAnimal - MouseHealthyBrain
Chen et al., 2017MWFEM - Myelin fractionAnimal - RatInjury - Dorsal columnar transectionSC
Khodanovich et al., 2017MPFHistology - LFBAnimal - MouseDemyelination - CuprizoneBrain
Lehto et al., 2017aRAFF, MTR, T1sat, FA, MD, AD, RDHistology - Gold chlorideAnimal - RatDemyelination - LipopolysaccharideBrain
Lehto et al., 2017bMTRHistology - Gold chlorideAnimal - RatTraumatic brain injuryBrain
van Tilborg et al., 2018FAImmunohistochemistry - MBPAnimal - RatWhite matter injuryBrain
Beckmann et al., 2018MTRHistology - LFBAnimal - MouseDemyelination - CuprizoneBrain
Berman et al., 2018MTVEM - Myelin fractionAnimal - MouseDemyelination - KnockoutBrain
Hametner et al., 2018R2*, T1, QSMHistology - LFBHumanVascular diseasesBrain
Praet et al., 2018MK, RK, AK, FA, MD, RD, ADImmunohistochemistry - MBPAnimal - MouseAmyloidosisBrain
Wendel et al., 2018FA, AD, RD, MDImmunohistochemistry - MBPAnimal - MouseTraumatic brain injuryBrain
West et al., 2018MPF, MWF, MVF-T2, MVF-MTEM - Myelin fractionAnimal - MouseDemyelination - KnockoutBrain
Yano et al., 2018FA, RD, MDImmunohistochemistry - PLPAnimal - MouseDemyelination - CuprizoneBrain
Abe et al., 2019FA, RD, ADMicroscopy - Myelin thicknessAnimal - MouseOptogenetic stimulationBrain
Duhamel et al., 2019ihMTR, MTRMicroscopy - FluorescenceAnimal - MouseHealthyBrain
Khodanovich et al., 2019MPFImmunohistochemistry - MBPAnimal - MouseDemyelination - CuprizoneBrain
Mollink et al., 2019FAImmunohistochemistry - MBPHumanAmyotrophic lateral sclerosisBrain
Peters et al., 2019FA, MDHistology - LFBHumanTuberous sclerosis complexBrain
Pol et al., 2019QSM, FA, MDHistology - SolochromeAnimal - MouseHealthyBrain
Soustelle et al., 2019MPF, RD, MWF, rSPFImmunohistochemistry - MBPAnimal - MouseDemyelination - CuprizoneBrain
Guglielmetti et al., 2020MTR, MTR-UTEImmunohistochemistry - MBPAnimal - MouseHealthyBrain

Data availability

All the data collected from the selected studies for this meta-analysis are provided in the spreadsheet file Source data 1.

References

  1. Book
    1. Beaulieu C
    (2009)
    CHAPTER 6 - The Biological Basis of Diffusion Anisotropy
    In: Johansen-Berg H, Behrens T. E. J, editors. Diffusion MRI. Academic Press. pp. 105–126.
  2. Book
    1. Cercignani M
    2. Dowell NG
    3. Tofts PS
    (2018)
    Quantitative MRI of the Brain
    CRC Press.
  3. Book
    1. Cohen-Adad J
    2. Wheeler-Kingshott CA
    (2014)
    Quantitative MRI of the Spinal Cord
    Academic Press.
  4. Book
    1. Höftberger R
    2. Lassmann H
    (2018)
    Chapter 19 - Inflammatory demyelinating diseases of the central nervous system
    In: Kovacs G. G, Alafuzoff I, editors. Handbook of Clinical Neurology. Elsevier. pp. 263–283.
  5. Book
    1. Lehman EL
    (1999)
    Elements of Large-Sample Theory
    Springer.
  6. Book
    1. Raudenbush SW
    (2009)
    Analyzing effect sizes: Random-effects models
    In: Hedges L. V, editors. The Handbook of Research Synthesis and Meta-Analysis,. Russell Sage Foundation. pp. 295–315.

Decision letter

  1. Saad Jbabdi
    Reviewing Editor; University of Oxford, United Kingdom
  2. Chris I Baker
    Senior Editor; National Institute of Mental Health, National Institutes of Health, United States
  3. Saad Jbabdi
    Reviewer; University of Oxford, United Kingdom
  4. Mark Does
    Reviewer

In the interests of transparency, eLife publishes the most substantive revision requests and the accompanying author responses.

Acceptance summary:

This meta-analysis is a great contribution to efforts in the neuroimaging field to better understand the biological underpinnings of MRI techniques. The work looks at published quantitative relationships between multiple MRI measures that are thought to be sensitive to myelin, and multiple measures from histology. The results are presented in an original, highly interactive manner, which will be a great resource for future studies of myelin in the brain.

Decision letter after peer review:

Thank you for submitting your article "An interactive meta-analysis of MRI biomarkers of myelin" for consideration by eLife. Your article has been reviewed by three peer reviewers, including Saad Jbabdi as the Reviewing Editor and Reviewer #1, and the evaluation has been overseen by Chris Baker as the Senior Editor. The following individual involved in review of your submission has agreed to reveal their identity: Mark Does (Reviewer #3).

The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.

Below is a compiled summary of all points raised by the three reviewers.

This article presents a meta-analysis of experimental comparisons between MRI and histological measures of myelin. The three reviewers agree that while this is not a particularly novel piece of work, it is well conducted and is presented in a highly original way using interactive visualisation, which is very useful for this type of complex meta analyses.

The reviewers agreed on the following points that need addressing:

1) Relating to previous literature

The primary shortcoming of the manuscript is that while it does a good job of citing prior experimental MRI-histology studies, it does a relatively poor job of providing references for other claims/assertions. We realize that it is not the authors' intention to review the physics or experimental history of myelin imaging methods, but when the authors provide a reference to support a statement, they ought to make it a suitable one (which may require looking more than a couple years back in history). We've listed a few examples below, but the authors should review the entire manuscript with this in mind.

Examples of questionable referencing:

"demyelination is often observed in several neurological diseases such as multiple sclerosis", cite: Wang Y et al., 2015, a paper reporting diffusion spectrum imaging evaluations MS

"measuring myelin in vivo has been an ambitious goal for magnetic resonance imaging (MRI) for almost two decades", cite: Petiet et al., 2019, a review of ultra-high field MRI measures of myelin.

"Diffusion acquisitions are blind to direct myelin measurement (Campbell et al., 2018)"

"A warning message that is evident from these results is the inherent limitation of DWI for estimating myelin content" again, this was studied extensively more than 20 years ago and has been discussed many time since. It's a good to reiterate, but don't make it sound like a novel finding.

"Faster techniques have been proposed for estimating it with gradient- and spin-echo (GRASE) sequences", cite: Faizy et al., 2018. This approach dates back 20 years and was used by Prasloski in 2012 to generate whole cerebrum MWF imaging.

2) Statistical analyses:

– Are the authors able to assess whether the correlations that they report are driven by tissue type differences or finer changes in the degree of myelination?

– Were there interactions between MR technique used and microscopy technique used in the literature? e.g. in Figure 5, are the R2 values for "myelin thickness" low because they happened to use diffusion measures rather than MT etc.?

– In general there was not a lot of information on interactions between the variables. Another example: were some techniques more likely to have been done at lower field (there was a strong correlation between field-strength and R2)?

– Given that the posed question is "how different are the modalities in their relationship to histology", is there a way to quantify or statistically test the mixed-model findings between modalities to effectively identify if any are better?

– It would be useful to identify the different types of histological techniques alongside the studies for each modality in Figure 4. While this is just one of many factors that is driving the high I2, it would allow for the visualisation of the heterogeneity of histological assessments for each modality. Not all histological techniques are born equal and despite the limitations, that the authors have already discussed, electron microscopy might be arguably the best assessment. I suspect due to the high number of MRI modalities and histological techniques and relatively small number of studies, it's not possible to quantify if any modality has a particularly good correlation with any of the two electron microscopy metrics. Still, if possible might be worth doing as EM is the gold-standard for cellular neuroscientists in the myelin field.

3) Overall message:

– Although the authors' discussion and conclusions present a more nuanced view of the findings, this is not clear from the Abstract. At least two MRI markers (MWF and MPF) have fairly high correlations and moderate prediction intervals. We think this could come across a bit more in the Abstract, as it does in their conclusion. Whether these are a true representation of histologically measured myelin is a different question. The authors have discussed this distinction effectively in their manuscript.

– We encourage the authors to consider leaving the reader on a somewhat more positive note. That is, while there is room for more study, this meta-analysis supports the view that myelin imaging with MRI is, in fact, relatively well substantiated by comparisons to histology. In fact, I would say that myelination is one characteristic of tissue that MRI has proven to be able to measure with a good level of specificity!

https://doi.org/10.7554/eLife.61523.sa1

Author response

This article presents a meta-analysis of experimental comparisons between MRI and histological measures of myelin. The three reviewers agree that while this is not a particularly novel piece of work, it is well conducted and is presented in a highly original way using interactive visualisation, which is very useful for this type of complex meta analyses.

We appreciate the reviewers’ enthusiasm. While we acknowledge that there are already several literature reviews on the topic of MRI-based measures of myelin (as reported in the paper, during the survey we found 50 review articles, with 6 of them specifically discussing comparisons between MRI and histology), to the best of our knowledge no systematic study has been published yet. A search for “myelin meta-analysis” and “myelin systematic review” on PubMed leads respectively to 93 and 91 results (last search: 23/09/20), yet none of these studies focus on MRI validation through histology. We believe that both the systematic and statistical aspects of our work are important, because they give overall quantitative measures of agreement and variability across studies.

The reviewers agreed on the following points that need addressing:

1) Relating to previous literature

The primary shortcoming of the manuscript is that while it does a good job of citing prior experimental MRI-histology studies, it does a relatively poor job of providing references for other claims/assertions. We realize that it is not the authors' intention to review the physics or experimental history of myelin imaging methods, but when the authors provide a reference to support a statement, they ought to make it a suitable one (which may require looking more than a couple years back in history). We've listed a few examples below, but the authors should review the entire manuscript with this in mind.

We thank the reviewers for pointing out this issue in the manuscript. We revised the whole manuscript and provided more suitable references where the previous ones did not properly support our statements (marked through the whole manuscript). We also specifically annotated the manuscript based on the reviewers’ comments 2-6.

Examples of questionable referencing:

"demyelination is often observed in several neurological diseases such as multiple sclerosis", cite: Wang Y et al., 2015, a paper reporting diffusion spectrum imaging evaluations MS

We substituted this reference with an updated review on demyelinating diseases from the Clinical Handbook of Neurology (Höftberger and Lassmann, 2018).

"measuring myelin in vivo has been an ambitious goal for magnetic resonance imaging (MRI) for almost two decades", cite: Petiet et al., 2019, a review of ultra-high field MRI measures of myelin.

We substituted this reference with three landmark studies respectively on myelin water imaging, magnetization transfer and T1 mapping (Mackay et al., 1994; Rooney et al., 2007; Stanisz, Kecojevic, Bronskill, and Henkelman, 1999).

"Diffusion acquisitions are blind to direct myelin measurement (Campbell et al., 2018)"

We substituted this reference with two seminal studies that estimated the transverse relaxation times of myelin water molecules and macromolecules (Mackay et al., 1994; Stanisz et al., 1999).

"A warning message that is evident from these results is the inherent limitation of DWI for estimating myelin content" again, this was studied extensively more than 20 years ago and has been discussed many times since. It's a good to reiterate, but don't make it sound like a novel finding.

We agree that this sentence may be misleading, so we clarified it and added relevant references (Beaulieu, 2002, 2009):

“A warning message that is evident from these results is the inherent limitation of DWI for estimating myelin content: this is not by any means a novel result (Beaulieu, 2002, 2009), but it is nevertheless worth reiterating given the outcomes of our analysis.”

We would still like to keep that sentence, as there are still studies using DWI-based measures for estimating myelin content without addressing the related limitations.

"Faster techniques have been proposed for estimating it with gradient- and spin-echo (GRASE) sequences", cite: Faizy et al., 2018. This approach dates back 20 years and was used by Prasloski in 2012 to generate whole cerebrum MWF imaging.

We substituted the previous reference with the original paper presenting the GRASE sequence, its first application to T2 relaxometry and the suggested study by Prasloski and colleagues (Does and Gore, 2000; Feinberg and Oshio, 1991; Prasloski et al., 2012).

2) Statistical analyses:

– Are the authors able to assess whether the correlations that they report are driven by tissue type differences or finer changes in the degree of myelination?

As we elaborate more in the next comment, we are not able to quantitatively answer this question given the number of potentially different conditions to consider and the limited number of studies. In more qualitative terms, Figure 6 and the interactive version Figure S7 (https://neurolibre.github.io/myelin-meta-analysis/04/other_factors.html#figure-7) show different correlation ranges depending on the types of tissues considered and on the specific pathology model, but in any case the range observed for white matter, the most common tissue studied, is particularly large, suggesting that tissue type differences are not the main factor affecting correlation. We added this consideration in the text:

“The effect of considering different types of tissues is showed in Figure 6 and Figure S7, where correlation ranges change when considering different types of tissue. However, the large correlation range in white matter, the most common tissue studied, suggests that other factors also affect the correlation.”

– Were there interactions between MR technique used and microscopy technique used in the literature? E.g. in Figure 5, are the R2 values for "myelin thickness" low because they happened to use diffusion measures rather than MT etc.?

As noted by the reviewers in this and the following comment, interactions are key elements in this kind of analysis. Unfortunately, given the limited number of studies we are not able to fit a model to study those interactions: specifically, we do not have enough studies to represent each possible MRI/microscopy combination. Despite not being able to tackle this question quantitatively, we believe that the provided interactive visualization is an effective way to qualitatively explore this kind of questions. We added these considerations:

“Given the limited number of studies, it is not possible to quantitatively study interactions between MRI measures and the other factors (e.g. modality used as a reference, tissue types, magnetic field strength). For further qualitative insights, we invite the reader to explore the interactive Figures S7-S8. A first important factor to consider is the validation modality used as a reference, which will be dictated by the equipment availability and cost.”

Regarding the specific issue raised by the reviewers, Figure S7 (https://neurolibre.github.io/myelin-meta-analysis/04/other_factors.html#figure-7) allows the reader to hover over each point in the top plot and get a sense of which MRI measures were investigated for each microscopy technique (screenshot attached). Myelin thickness was used as reference for twelve measures that included diffusion, relaxometry and magnetization transfer.

We believe that having access to the interactive figures directly in an executable research article on the Stencila platform will allow a more immediate exploration of our results.

– In general there was not a lot of information on interactions between the variables. Another example: were some techniques more likely to have been done at lower field (there was a strong correlation between field-strength and R2)?

Following up on the previous comment, Figure S8 (https://neurolibre.github.io/myelin-meta-analysis/04/other_factors.html#figure-8) shows the MRI measures as a function of the related magnetic field strengths, using the same colour coding as Figure 2. One can notice how most studies have been done at 7T and 9.4T, while the first studies (chronologically) were performed at 1.5T. Few measures were studied at other field strengths. We added this consideration in the text:

“A further example of influential factor often dictated by equipment availability is the magnetic field strength of the MRI scanner: Figure S8 shows that most studies were conducted at 7T and 9.4T, with some pioneering studies at 1.5T and even fewer ones at other field strengths.”

– Given that the posed question is "how different are the modalities in their relationship to histology", is there a way to quantify or statistically test the mixed-model findings between modalities to effectively identify if any are better?

Following the reviewers’ suggestion, we performed an additional analysis using a repeated measures meta-regression, explained in the text:

“For the explicit purpose of comparing the effect sizes between different MRI measures, we conducted a repeated measures meta-regression on every R2 value recorded. […] While the repeated measures meta-regression makes direct comparisons straightforward, we reported the main R2 estimates based on the measure-specific mixed-effects models, as they make weaker assumptions.”

From this analysis, we observed both significant differences and comparable R2 estimates, overall subdividing the MRI measures in two groups, with magnetization- and relaxometry-based ones providing higher estimates and diffusion-based measures providing lower estimates. The new results are reported in the text and discussed in subsection “Meta-analysis”:

“To investigate the significance of the differences between measures, we conducted a repeated measures meta-regression on every R2 estimate recorded (98 in total over 43 studies). […] From this perspective, MPF has higher R2 estimates compared to all the other measures, but it is only marginally higher than MWF (z-score=0.77; p-value=1) so we cannot claim that one is superior to the other. Following the same reasoning, MTR and T1 are not statistically different (z-score=0.47; p-value=1).”

The repeated measure meta-regression confirms this overall picture, clearly distinguishing MT- and relaxometry-based measures from diffusion-based ones (Figure 5).

– It would be useful to identify the different types of histological techniques alongside the studies for each modality in Figure 4. While this is just one of many factors that is driving the high I2, it would allow for the visualisation of the heterogeneity of histological assessments for each modality. Not all histological techniques are born equal and despite the limitations, that the authors have already discussed, electron microscopy might be arguably the best assessment. I suspect due to the high number of MRI modalities and histological techniques and relatively small number of studies, it's not possible to quantify if any modality has a particularly good correlation with any of the two electron microscopy metrics. Still, if possible might be worth doing as EM is the gold-standard for cellular neuroscientists in the myelin field.

We agree that it is useful to have a sense of which histological techniques were used for each MRI measure: we added this information in Figure S5 (https://neurolibre.github.io/myelin-meta-analysis/03/meta_analysis.html#figure-5), when hovering on each point (screenshot attached). Although this information is currently not visible in the static figure in the manuscript, it will be immediately accessible in Stencila.

As mentioned in comment 8 and as the reviewers already acknowledge, unfortunately the number of studies using EM is not sufficient to make more definitive statements for each MRI modality.

3) Overall message:

– Although the authors' discussion and conclusions present a more nuanced view of the findings, this is not clear from the Abstract. At least two MRI markers (MWF and MPF) have fairly high correlations and moderate prediction intervals. We think this could come across a bit more in the Abstract, as it does in their conclusion. Whether these are a true representation of histologically measured myelin is a different question. The authors have discussed this distinction effectively in their manuscript.

We agree that the Abstract was not able to present the overall picture as in the discussion. We rephrased part of the Abstract taking into account both this comment and the journal’s word count limit of 150 words:

“We report the overall estimates and the prediction intervals for the coefficient of determination and find that MT and relaxometry-based measures exhibit the highest correlations with myelin content. We also show which measures are, and which measures are not statistically different regarding their relationship with histology.”

– We encourage the authors to consider leaving the reader on a somewhat more positive note. That is, while there is room for more study, this meta-analysis supports the view that myelin imaging with MRI is, in fact, relatively well substantiated by comparisons to histology. In fact, I would say that myelination is one characteristic of tissue that MRI has proven to be able to measure with a good level of specificity!

We added a brief paragraph towards the end of the Discussion to emphasize the positive message of our results:

“We hope this meta-analysis convinces the reader that a holy grail of myelin imaging does not exist, at least as long as we consider histology to be the ground truth. Given that we all have to pick our poison, the upside is that measures based on MT and relaxometry are not statistically different, and therefore future studies have an actual choice among candidate measures.”

https://doi.org/10.7554/eLife.61523.sa2

Article and author information

Author details

  1. Matteo Mancini

    1. Department of Neuroscience, Brighton and Sussex Medical School, University of Sussex, Brighton, United Kingdom
    2. NeuroPoly Lab, Polytechnique Montreal, Montreal, Canada
    3. CUBRIC, Cardiff University, Cardiff, United Kingdom
    Contribution
    Conceptualization, Data curation, Formal analysis, Visualization, Methodology, Writing - original draft, Writing - review and editing
    For correspondence
    ingmatteomancini@gmail.com
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-7194-4568
  2. Agah Karakuzu

    NeuroPoly Lab, Polytechnique Montreal, Montreal, Canada
    Contribution
    Visualization, Writing - review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-7283-271X
  3. Julien Cohen-Adad

    1. NeuroPoly Lab, Polytechnique Montreal, Montreal, Canada
    2. Functional Neuroimaging Unit, CRIUGM, Université de Montréal, Montreal, Canada
    Contribution
    Conceptualization, Writing - review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0003-3662-9532
  4. Mara Cercignani

    1. Department of Neuroscience, Brighton and Sussex Medical School, University of Sussex, Brighton, United Kingdom
    2. Neuroimaging Laboratory, Fondazione Santa Lucia, Rome, Italy
    Contribution
    Conceptualization, Writing - review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-4550-2456
  5. Thomas E Nichols

    1. Wellcome Centre for Integrative Neuroimaging (WIN FMRIB), University of Oxford, Oxford, United Kingdom
    2. Big Data Institute, University of Oxford, Oxford, United Kingdom
    Contribution
    Conceptualization, Methodology, Writing - review and editing
    Contributed equally with
    Nikola Stikov
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-4516-5103
  6. Nikola Stikov

    1. NeuroPoly Lab, Polytechnique Montreal, Montreal, Canada
    2. Montreal Heart Institute, Université de Montréal, Montreal, Canada
    Contribution
    Conceptualization, Methodology, Writing - review and editing
    Contributed equally with
    Thomas E Nichols
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-8480-5230

Funding

Wellcome Trust (213722/Z/18/Z)

  • Matteo Mancini

National Institutes of Health (R01MH096906)

  • Thomas E Nichols

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

MM was funded by the Wellcome Trust through a Sir Henry Wellcome Postdoctoral Fellowship [213722/Z/18/Z]. TEN was supported by NIH grant R01MH096906.

Senior Editor

  1. Chris I Baker, National Institute of Mental Health, National Institutes of Health, United States

Reviewing Editor

  1. Saad Jbabdi, University of Oxford, United Kingdom

Reviewers

  1. Saad Jbabdi, University of Oxford, United Kingdom
  2. Mark Does

Publication history

  1. Received: July 28, 2020
  2. Accepted: October 20, 2020
  3. Accepted Manuscript published: October 21, 2020 (version 1)
  4. Version of Record published: November 6, 2020 (version 2)

Copyright

© 2020, Mancini et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 7,425
    Page views
  • 248
    Downloads
  • 10
    Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF and Executable version)

Download citations (links to download the citations from this article in formats compatible with various reference manager tools)

Open citations (links to open the citations from this article in various online reference manager services)

  1. Further reading

Further reading

    1. Neuroscience
    Charlotte Vanacker et al.
    Research Article Updated

    GnRH neurons are the final central neural output regulating fertility. Kisspeptin neurons in the hypothalamic arcuate nucleus (KNDy neurons) are considered the main regulator of GnRH output. GnRH and KNDy neurons are surrounded by astrocytes, which can modulate neuronal activity and communicate over distances. Prostaglandin E2 (PGE2), synthesized primarily by astrocytes, increases GnRH neuron activity and downstream pituitary release of luteinizing hormone (LH). We hypothesized that glial fibrillary acidic protein (GFAP)-expressing astrocytes play a role in regulating GnRH and/or KNDy neuron activity and LH release. We used adeno-associated viruses to target designer receptors exclusively activated by designer drugs (DREADDs) to GFAP-expressing cells to activate Gq- or Gi-mediated signaling. Activating Gq signaling in the preoptic area, near GnRH neurons, but not in the arcuate, increases LH release in vivo and GnRH firing in vitro via a mechanism in part dependent upon PGE2. These data suggest that astrocytes can activate GnRH/LH release in a manner independent of KNDy neurons.

    1. Neuroscience
    Sahana Sitaraman et al.
    Research Article

    Gap junctions between neurons serve as electrical synapses, in addition to conducting metabolites and signaling molecules. During development, early-appearing gap junctions are thought to prefigure chemical synapses, which appear much later. We present evidence for this idea at a central, glutamatergic synapse and provide some mechanistic insights. Loss or reduction in the levels of the gap junction protein Gjd2b decreased the frequency of glutamatergic miniature excitatory postsynaptic currents (mEPSCs) in cerebellar Purkinje neurons (PNs) in larval zebrafish. Ultrastructural analysis in the molecular layer showed decreased synapse density. Further, mEPSCs had faster kinetics and larger amplitudes in mutant PNs, consistent with their stunted dendritic arbors. Time-lapse microscopy in wild type and mutant PNs reveals that Gjd2b puncta promote the elongation of branches and that CaMKII may be a critical mediator of this process. These results demonstrate that Gjd2b-mediated gap junctions regulate glutamatergic synapse formation and dendritic elaboration in PNs.