TL;DR
This study reveals that the temporal and spatial scales of brain activity, as measured by fMRI, generally decrease with age across the human lifespan, with some exceptions in prefrontal regions. These changes are closely linked to cortical myelination, suggesting a coordinated developmental and aging process.
What Problem Does This Paper Address?
While we know that cognitive abilities change significantly throughout life, the underlying functional properties of cortical networks and how they evolve remain less understood. Researchers are interested in how the brain's activity patterns change from childhood to old age. Temporal and spatial scales of resting-state fMRI signals offer a way to characterize these patterns, reflecting the speed and extent of neural communication. However, a comprehensive understanding of how these scales change across the entire lifespan, from childhood to old age, and how they relate to other biological factors like myelination, has been lacking.
Methods
The researchers used resting-state functional magnetic resonance imaging (fMRI) data to analyze brain activity.
- Temporal Scale: This was estimated using the "lag-1 autocorrelation" of the fMRI signal. Essentially, they looked at how correlated the signal was with itself in consecutive time steps. A higher correlation suggests a slower, more sustained signal (longer timescale).
- Spatial Scale: This was determined by measuring how quickly functional connectivity (how brain regions work together) decays with distance across the cortical surface. A slower decay indicates a larger spatial scale of interaction.
- Cortical Myelination: This was estimated using the ratio of T1-weighted to T2-weighted MRI signal intensities, a common proxy for myelin content.
- Statistical Analysis: They employed linear regression and linear mixed-effects models to examine age-related changes and interactions between age and brain networks. Locally weighted scatterplot smoothing (LOWESS) was used for visualizing age-related trends.
Datasets
The study utilized data from the Human Connectome Project (HCP), specifically:
- HCP-Developmental (HCP-D): Youth and adolescents (ages 5-22).
- HCP-Young Adult (HCP-YA): Young adults (ages 22-37).
- HCP-Aging (HCP-A): Middle-aged and elderly adults (ages 37-100).
This combined dataset provided a broad age range (5-100 years) with a total of 2,352 subjects.
Key Insights and Ideas
- Lifespan Trajectory: The most striking finding is the general decrease in both temporal and spatial scales of brain activity as people age, suggesting a more localized and faster communication pattern in older brains.
- Prefrontal Exceptions: The prefrontal cortex shows a different pattern, with scales increasing during adolescence before decreasing, hinting at a prolonged maturation process in these higher-order cognitive areas.
- Myelination Link: The strong correlation between age-related changes in temporal/spatial scales and cortical myelination is a crucial insight, suggesting that the development and aging of brain networks are tightly coupled with the maturation and changes in white matter.
- Network Specificity: Not all brain regions or networks change uniformly. The study highlights differences across various functional networks, underscoring the complexity of brain development and aging.
Key Results
- General Decrease: For most cortical regions, both temporal and spatial scales decreased with age.
- Prefrontal Non-monotonicity: Some prefrontal regions showed non-monotonic trajectories, peaking around adolescence.
- Myelination Increase: Cortical myelination generally increased throughout the lifespan.
- Correlations:
- Age-related changes in temporal and spatial scales were correlated with cortical myelination in adulthood.
- Spatial scale was significantly correlated with cortical myelination across all age groups.
- Age coefficients for temporal and spatial scales were significantly correlated, suggesting coordinated age-related changes.
Limitations and Future Work
- Dataset Discontinuities: The study acknowledges discontinuities at age boundaries between the different HCP datasets, likely due to variations in scanning protocols and preprocessing methods. They chose not to harmonize these to preserve original units.
- Hemodynamic Response: The authors note that age-related changes in hemodynamic response functions could contribute to the observed fMRI signal changes, but the exact contribution to timescales is not fully understood.
- Future Research: More research is needed to precisely disentangle the neural and neurovascular factors influencing temporal and spatial changes in BOLD activity. Further investigation into how these scales predict specific cognitive functions is also warranted.
Summary updated