Application of magnetic resonance imaging to the study of age related changes and gender differences in the human brain

Abstract

Magnetic Resonance Imaging (MRI) is an invaluable tool used in medical research and clinical practice. T1 weighted (T1w) and T2 weighted (T2w) magnetic resonance images are used to detect pathological changes in brain tissue such as demyelination and iron accumulation. Voxel based statistical analysis (VBSA) is a technique used to identify brain regions that show evidence of abnormality in disease populations or individual patients. While VBSA is used extensively in the analysis of functional magnetic resonance imaging and voxel based morphometry images, its use in T1w and T2w image datasets has been much more limited. In this work, age related changes and gender differences in T1w and T2w images are examined in a dataset of 152 subjects taken from a publicly available dataset of healthy volunteers, demonstrating the utility of VBSA when applied to these mage types. To investigate the most suitable intensity normalisation method, VBSA results were obtained with three different global intensity normalisation algorithms, the first algorithm was based on the whole brain mean, the second method was the voxel based iterative sensitivity (VBIS) method of (Abbott et al. 2009) and the third was based on the location of the intensity histogram peak. The latter two algorithms showed good agreement. The most striking and statistically strongest result obtained is a strong negative correlation between T2w signal and age in the putamen, caudate head, para-hippocampal gyrus and amygdala, consistent with iron accumulation in these regions and most likely due to a higher rate of micro-bleeds occurring in these brain regions. Other results include decreased T1w signal and increased T2w signal in the thalamus with advancing age. Results for Gender differences include a significantly higher T2w signal in the superior longitudinal fasciculus and internal capsule for females and a significantly higher T2w intensity in the medulla for males. T1w intensity was significantly higher in males in parts of the midbrain. These results demonstrate that VBSA of T1w and T2w images is an effective tool for establishing physiologically meaningful localised patterns of change in cross-sectional neuro-image data-sets.