Abstract
Background Puberty depicts a period of profound and multifactorial changes ranging from social to biological factors. While brain development in youths has been studied mostly from an age perspective, recent evidence suggests that pubertal measures may be more sensitive to study adolescent neurodevelopment, however, studies on pubertal timing in relation to brain development are still scarce.
Methods We investigated if pre- vs. post-menarche status can be classified using machine learning on cortical and subcortical structural magnetic resonance imaging (MRI) data from strictly age-matched adolescent females from the Adolescent Brain Cognitive Development (ABCD) cohort. For comparison of the identified menarche-related patterns to age-related patterns of neurodevelopment, we trained a brain age prediction model on data from the Philadelphia Neurodevelopmental Cohort and applied it to the same ABCD data, yielding differences between predicted and chronological age referred to as brain age gaps. We tested the sensitivity of both these frameworks to measures of pubertal maturation, specifically age at menarche and puberty status.
Results The machine learning model achieved moderate but statistically significant accuracy in the menarche classification task, yielding for each subject a class probability ranging from 0 (pre-) to 1 (post-menarche). Comparison to brain age predictions revealed shared and distinct patterns of neurodevelopment captured by both approaches. Continuous menarche class probabilities were positively associated with brain age gaps, but only the menarche class probabilities – not the brain age gaps – were associated with age at menarche.
Conclusions This study demonstrates the use of a machine learning model to classify menarche status from structural MRI data while accounting for age-related neurodevelopment. Given its sensitivity towards measures of puberty timing, our work suggests that menarche class probabilities may be developed toward an objective brain-based marker of pubertal development.
Highlights
✓ We classified pre- vs. post-menarche status in adolescent females from structural brain imaging data
✓ We compared class probabilities to brain-age predictions to disentangle puberty- vs. age-related patterns of brain development
✓ The derived continuous brain-based menarche class probabilities captured shared but also unique variations of adolescent neurodevelopment, and were associated with pubertal timing and status
Plain English Summary Puberty is a period of substantial changes in the life of youths, and these include profound brain changes. Most studies have investigated age related changes in brain development, recent work however suggests that looking at brain development through the lens of pubertal development can provide additional insights beyond age effects. We here analyzed brain imaging data from a group of same-aged adolescent girls from the Adolescent Brain Cognitive Development study. Our goal was to investigate if we could determine from brain images whether a girl had started her menstrual period (menarche) or not, and we used machine learning to classify between them. This machine learning model does not just return a “yes/no” decision, but also returns a number between 0 and 1 indicating a probability to be pre- (0) or post- (1) menarche. To rule out, that our approach only maps age-related development, we selected a strictly age-matched sample of girls and compared our classification model to a brain age model trained on independent individuals. Our model classified between pre- and post-menarche with moderate accuracy. The obtained class probability was partly related to age-related brain development, but only the probability was significantly associated with pubertal timing (age at menarche). In summary, our study uses a machine learning model to estimate whether a girl has reached menarche based on her brain structure. This approach offers new insights into the connection between puberty and brain development and might serve as an objective way to assess pubertal timing from imaging data.
Competing Interest Statement
The authors have declared no competing interest.
Funding Statement
TK received support by the Fortuene Program (2660-0-0), Faculty of Medicine, University of Tuebingen, and the Research Council of Norway (#323961). TK is a member of the Machine Learning Cluster of Excellence, EXC number 2064/1, Project number 39072764. Brain imaging data was processed on the BMBF-funded de.NBI Cloud as part of the German Network for Bioinformatics Infrastructure (031A537B, 031A533A, 031A538A, 031A533B, 031A535A, 031A537C, 031A534A, 031A532B).
Author Declarations
I confirm all relevant ethical guidelines have been followed, and any necessary IRB and/or ethics committee approvals have been obtained.
Yes
The details of the IRB/oversight body that provided approval or exemption for the research described are given below:
PNC study procedures were approved by institutional review boards of the University of Pennsylvania and the Children's Hospital of Philadelphia. All participants or their caregiver provided written informed consent. ABCD procedures were approved by either the local site Institutional Review Board or by local Institutional Review Board reliance agreements with the central Institutional Review Board at the University of California. All participants and their parents provided written informed consent.
I confirm that all necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived, and that any patient/participant/sample identifiers included were not known to anyone (e.g., hospital staff, patients or participants themselves) outside the research group so cannot be used to identify individuals.
Yes
I understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance).
Yes
I have followed all appropriate research reporting guidelines, such as any relevant EQUATOR Network research reporting checklist(s) and other pertinent material, if applicable.
Yes
Data Availability
Data used in this study has been shared with the authors under the data usage agreements of the respective cohorts (i.e., ABCD and PNC). To use these resources, interested scholars can obtain access by signing their own respective data usage agreements with the officials. Code used to analyze the data will be made available on github upon acceptance of the manuscript.