Fuzzy Clustering to Asses BALI and LIBRA factors for Estimation of DTI measures

Diffusion magnetic resonance imaging (dMRI) is a popular technique for diagnosing dementia through finding a number of measures with diffusion tensor imaging (DTI). However, this technique is too expensive to be widely used to scan populations. The primary objective of this research is to identify factors/indices which are both (i) rather inexpensive to find, and (ii) usable to estimate DTI measures and eventually to diagnose dementia. This will the basis for a low-cost diagnostic solution. Such factors are selected amongst lifestyle for brain health (LIBRA) and brain atrophy and lesion index (BALI) factors. These factors are pertinent to dementia and relatively inexpensive to find. However, BALI and LIBRA are comprised of 49 factors altogether, and development of a diagnostic algorithm with 49 inputs is infeasible. Therefore, it is necessary to pick the most impactful factors to be used in diagnosis algorithm development. Fuzzy subtractive clustering was employed for this purpose. This research shows that the grey matter lesions and subcortical dilated perivascular spaces (GM-SV) and periventricular white matter lesions (PV) from BALI and age, level of education, job status, antidepressant drugs, diabetes control drugs, obesity (BMI) and dementia preventive diet from LIBRA are the most influential factors to identify DTI measures.