Recently, Electroencephalogram (EEG) shows potential in the diagnosis of Alzheimer's disease and other dementia. We aim to investigate whether EEG and selected cognitive biomarkers can classify mild cognitive impairment (MCI), dementia and healthy subjects using support vector machine classifier in Indian cohort.

Eight EEG biomarkers, power spectral density, skewness, kurtosis, spectral skewness, spectral kurtosis, spectral crest factor, spectral entropy (SE), fractal dimension (FD) were analyzed from 44 subjects in four conditions; eye-open, eye-close, finger tapping test (FTT) and continuous performance test (CPT). FFT and CPT are used to measure motor speed and sustained attention as these cognitive biomarkers are free from the educational barrier.

We achieved very good accuracy for each event from 73.4% to 89.8% for three binary classes. We investigated that FTT (84% accuracy), CPT (88% accuracy) were the most efficient events to diagnose MCI from dementia. MCI from control successfully diagnosed with 89.8% accuracy in FTT, 73.4% accuracy in CPT and 84.1% accuracy in eye open resting state. Even though cognitive biomarkers were also adequately diagnosed MCI from other groups.

Our classifier findings are consistent with the utmost evidence. Yet, our results are promising and especially newfangled in the case of FTT and CPT from the prior studies. We developed an experimental protocol and proposed a novel technique to classify MCI with efficient biomarkers.