An exploratory data science project investigating structural MRI biomarkers of aging using the OASIS-1 dataset.
This project utilizes Python to analyze the relationship between chronological age and normalized whole-brain volume (nWBV). Beyond simple linear trends, this analysis explores gender-based dimorphism and identifies "Super-Ager" individuals who exhibit high neurological resilience despite advanced age.
The analysis confirmed a strong negative correlation between age and brain volume, with an R-squared value of 0.78. This indicates that approximately 78% of the variance in brain volume is explained by chronological aging.
Stratified analysis revealed a divergence in aging patterns between biological sexes:
- Males (r = -0.93): Exhibited a highly linear and aggressive decline in brain volume.
- Females (r = -0.85): Showed slightly more variance, suggesting potential differences in neuroprotective factors.
Using a relative resilience threshold, the pipeline identified 8 individuals (Age 70+) whose brain volumes remained in the top 25th percentile of their cohort.
- Clinical Validation: Super-Agers maintained a higher mean cognitive score (MMSE = 28.0) compared to their age-matched peers (MMSE = 27.0), linking structural resilience to functional cognitive health.
- Language: Python 3.12
- Data Science:
Pandas,NumPy,Scikit-Learn - Neuroimaging Tools:
Nilearn(OASIS-1 dataset) - Visualization:
Matplotlib,Seaborn
analysis.py: Main processing and visualization pipeline.results/: Contains high-resolution plots of regression, gender comparisons, and heatmaps.README.md: Project documentation and summary of findings.