TL;DR
This study reveals that the neural pathways linking psychological stress to brain health are remarkably similar in both healthy individuals and those with multiple sclerosis (MS). While the stress response itself doesn't differ, the impact of stress on brain health, as measured by "brain-predicted age differences" (brain-PAD), is amplified in MS, suggesting disease-specific vulnerabilities play a crucial role.
What Problem Does This Paper Address?
We know that psychological stress can negatively impact brain health, both in general and in individuals with neurological disorders like multiple sclerosis (MS). However, the specific neural mechanisms that mediate this link and whether these pathways are conserved across health and disease have remained unclear. This paper aims to bridge this gap by investigating whether the brain's response to stress and its subsequent impact on brain health are comparable in healthy individuals and MS patients. Understanding these shared pathways could pave the way for universal stress-reduction strategies that benefit everyone, regardless of their health status.
Methods
The researchers employed a multi-faceted approach combining functional Magnetic Resonance Imaging (fMRI) with advanced machine learning techniques.
- Arterial Spin Labeling (ASL) fMRI: This technique was used to measure regional cerebral blood flow (rCBF), which is a proxy for neural activity. ASL was chosen over the more common BOLD fMRI due to its robustness to slow signal fluctuations, which are characteristic of stress responses. Participants underwent fMRI during both a resting state and a stressful mental arithmetic task.
- Brain-Predicted Age Differences (brain-PAD): This is a novel structural brain health biomarker derived from machine learning. A convolutional neural network (CNN), pre-trained on a large dataset of healthy brains, was used to predict brain age from individual MRI scans. The difference between the predicted brain age and the individual's chronological age (brain-PAD) serves as an indicator of brain health – a higher brain-PAD suggests poorer brain health.
- Functional Connectivity (FC): The study analyzed the functional connectivity between stress-responsive brain regions identified during the fMRI stress task. This was done by correlating the rCBF time series of these regions.
- Statistical Modeling: Robust linear models were used to investigate the associations between stress responsivity, functional connectivity, and brain-PAD, while controlling for various nuisance factors (e.g., age, sex, disease duration). Permutation testing was employed for statistical inference.
Datasets
The study utilized data from two previously conducted projects at Charité – Universitätsmedizin Berlin, Germany.
- Participants: The combined dataset included 42 healthy persons (HPs) and 56 persons with multiple sclerosis (PwMS).
- MRI Data: Anatomical and functional MRI scans were acquired using a 3 Tesla whole-body tomograph. ASL fMRI was used for functional scans.
- Brain-Age Prediction Model: A pre-trained CNN from Bashyam et al. (2020) was used, which was trained on anatomical MRI scans from 11,729 healthy individuals.
Key Insights and Ideas
- Shared Stress Pathways: The most striking finding is that the brain regions activated by stress and the functional connections between them are largely similar in both healthy individuals and MS patients. This suggests a fundamental, conserved neural mechanism for processing stress.
- Amplified Impact in MS: Despite similar stress responses, MS patients exhibited significantly higher brain-PAD, indicating poorer brain health. This highlights that while the initial stress response might be comparable, the downstream consequences on brain health are exacerbated in MS.
- Functional Connectivity as a Bridge: The study found that stress-stage functional connectivity between specific regions (anterior insula and occipital areas) was associated with brain-PAD in both groups. This suggests that these connections act as a crucial link between stress and brain health, and that this link is more impactful in MS.
- Brain-Age as a Generic Biomarker: The brain-age prediction approach proved to be a sensitive measure of brain health, capable of detecting variations in both healthy individuals and those with MS. The study also found that regional brain aging had comparable neurobiological substrates in both groups, further supporting the idea of a generic stress-brain health pathway.
Key Results
- Brain-Age Prediction: The machine learning model accurately predicted chronological age in healthy individuals, explaining 81% of the variance. MS patients showed significantly higher brain-PAD (average of 10.08 years) compared to healthy individuals (average of 3.95 years).
- Regional Stress Responsivity: A wide set of stress-responsive regions was identified in both groups, with no significant differences in the number or location of these regions between HPs and PwMS.
- Functional Connectivity and Brain-PAD: Stress-stage functional connectivity between the right anterior insula and left occipital pole was positively associated with brain-PAD in HPs. A similar association was found between the left anterior insula and the left occipital fusiform gyrus in PwMS. Resting-stage FC showed no such associations.
- Brain Aging and Grey Matter: A strong negative association was found between regional brain-age scores and grey matter fraction in the parietal cortex in both groups, suggesting that over-aging in this region is linked to grey matter loss.
- Future Brain Health Prediction: In MS patients, higher stress-stage functional connectivity between the right middle frontal cortex and cerebellar vermal lobules was associated with future increases in brain-PAD, suggesting this connectivity has prognostic value for future brain health.
Limitations and Future Work
- Single Neurological Disorder: The study focused solely on MS patients, which limits the generalizability of the findings to other neurological disorders. Future research should include patients with conditions like dementia to broaden the scope.
- Sample Size: While considered large for MS fMRI studies, the authors acknowledge that the sample size was not predetermined by a formal sample size calculation.
- Nuisance Factors: Although the study attempted to control for numerous nuisance factors, the multifactorial nature of MS means that subtle influences might remain.
- Future Directions: The authors suggest that stress modification techniques, which have shown promise in reducing brain-PAD in healthy individuals and lesions in MS patients, could be a valuable intervention for promoting brain health across the board. Further research could explore the efficacy of these techniques in different patient populations.
Summary updated