




Description
BCC is my Brain Code Camp project with ADNI data. It asks whether we can predict which people with mild cognitive impairment (MCI) progress to dementia (an ADNI clinical diagnosis) within 30 months (a 24-month target plus 6 months of visit tolerance), using only tests that Thai hospitals already have: age, sex, years of education, MMSE, MoCA, and hippocampal volume with intracranial volume from a T1 MRI. Out of 63 input sets, a 4-variable set (age, education, MoCA, and hippocampal volume with ICV) reached an AUROC of 0.815. For clinics without MRI, a paper integer score from MoCA, MMSE, sex and years of education reached about 0.769. The integer score is a points table: a doctor adds up the points on paper, with no computer, and then reads the risk of dementia (an ADNI clinical diagnosis) within that same 30-month window from a second table. I built it alone. BCC is research. It is not a diagnostic product, and it is not validated for Thai patients.


