Abnormal subcortical nuclei shapes in patients with type 2 diabetes mellitus.

OBJECTIVES: Type 2 diabetes mellitus (T2DM) increases the risk of brain atrophy and dementia. We aimed to elucidate deep grey matter (GM) structural abnormalities and their relationships with T2DM cognitive deficits by combining region of interest (ROI)-based volumetry, voxel-based morphometry (VBM) and shape analysis.

METHODS: We recruited 23 T2DM patients and 24 age-matched healthy controls to undergo T1-weighted structural MRI scanning. Images were analysed using the three aforementioned methods to obtain deep GM structural shapes and volumes. Biochemical and cognitive assessments were made and were correlated with the resulting metrics.

RESULTS: Shape analysis revealed that T2DM is associated with focal atrophy in the bilateral caudate head and dorso-medial part of the thalamus. ROI-based volumetry only detected thalamic volume reduction in T2DM when compared to the controls. No significant between-group differences were found by VBM. Furthermore, a worse performance of cognitive processing speed correlated with more severe GM atrophy in the bilateral dorso-medial part of the thalamus. Also, the GM volume in the bilateral dorso-medial part of the thalamus changed negatively with HbA1c .

CONCLUSIONS: Shape analysis is sensitive in identifying T2DM deep GM structural abnormalities and their relationships with cognitive impairments, which may greatly assist in clarifying the neural substrate of T2DM cognitive dysfunction.

KEY POINTS: • Type 2 diabetes mellitus is accompanied with brain atrophy and cognitive dysfunction • Deep grey matter structures are essential for multiple cognitive processes • Shape analysis revealed local atrophy in the dorso-medial thalamus and caudatum in patients • Dorso-medial thalamic atrophy correlated to cognitive processing speed slowing and high HbA1c. • Shape analysis has advantages in unraveling neural substrates of diabetic cognitive deficits.