Brain morphology normative modelling platform for abnormality and centile estimation: Brain MoNoCle

Abstract

Normative models of brain structure estimate the effects of covariates such as age and sex using large samples of healthy controls. These models can then be applied to, for example, smaller clinical cohorts to distinguish disease effects from other covariates. However, these advanced statistical modelling approaches can be difficult to access, and processing large healthy cohorts is computationally demanding. Thus, accessible platforms with pre-trained normative models are needed. We present such a platform for brain morphology analysis as an open-source web applicationhttps://cnnplab.shinyapps.io/BrainMoNoCle/, with six key features: (i) user-friendly web interface, (ii) individual and group outputs, (iii) multi-site analysis, (iv) regional and whole-brain analysis, (v) integration with existing tools, and (vi) featuring multiple morphology metrics. Using a diverse sample of 3,276 healthy controls across 21 sites, we pre-trained normative models on various metrics. We validated the models with a small sample of individuals with bipolar disorder, showing outputs that aligned closely with existing literature only after applying our normative modelling. Using a cohort of people with temporal lobe epilepsy, we showed that individual-level abnormalities were in line with seizure lateralisation. Finally, with the ability to investigate multiple morphology measures in the same framework, we found that biological covariates are better explained in specific morphology measures, and for applications, only some measures are sensitive to the disease process. Our platform offers a comprehensive framework to analyse brain morphology in clinical and research settings. Validations confirm the superiority of normative models and the advantage of investigating a range of brain morphology metrics together.

Keywords: bipolar disorder; brain structure; morphology; normative model; structural abnormality; temporal lobe epilepsy.

Fig. 1.

Demographics of the data underlying the normative model. Age distributions and proportion of female participants are shown for each study. Some studies contained multi-site data (e.g. MEGUK), but these are not shown separately here.

Fig. 2.

Alterations in cortical thickness associated with bipolar disorder derived from a case–control studyvs.normative modelling. Group-level abnormalities in cortical thickness in n = 56 people with bipolar disorder for (A) a small, matched control group (n = 26) and (B) the normative reference population (n = 3,276).

Fig. 3.

Group-level output for mesial temporal lobe epilepsy cohort after normative modelling. Group-level summary of abnormalities in cortical thickness for left mTLE (n = 74, A) and right mTLE (n = 59, B), showing Cohen’s d effect size for each cortical region.

Fig. 4.

Individual-level z-scores after normative modelling for mTLE cohort. Difference in hemisphere-level z-score between left and right hemisphere is shown in controls and right/left TLE subgroups for three example morphological measures. Individual subjects are shown as single data points, distributions of subjects are displayed as violin plots.

Fig. 5.

Variance explained by normative model in each morphometric. (A & B) Harmonised normative data (grey dots) and predicted model centiles of mean cortical thickness and K across the lifespan (n = 3,276). (C) Model fit statistics$R^2$for each metric and hemisphere. CT, CV, and SA are structural metrics estimated using FreeSurfer; T, At, Ae, K, I, and S are structural metrics estimated using the Cortical Folding toolbox. CT = cortical thickness, CV = cortical volume, SA = surface area, T = average thickness, At = total pial surface area, Ae = exposed surface area.