Assessment of senile dementia of alzheimer type using artificial neural networks

We constructed a model for assessment of senile dementia of Alzheimer type (SDAT) from electroencephalogram (EEG) using artificial neural networks (ANN). EEG data of the SDAT patients and the non SDAT patients were collected. At first, power spectrum of EEG was extracted by the fast Fourier transform (FFT). The power spectrum was separated into 9 regions, corresponding to the characterized waves, and relative power values were calculated from them. The regions with 4.0-6.0, 6.0-8.0 and 8.0-13.0 Hz were assigned as the frequency bands of θ₁, θ₂, and α waves, respectively. The average power value in all electrode positions of the head was also calculated. The severity of SDAT was assessed by Hasegawa's dementia rating scale (HDS-R). The relative power values and the average power value were input into each ANN model for the distinction of SDAT patients from non SDAT patients and estimation of HDS-R score of the patients. Using the acquired ANN model, SDAT patients are distinguished from non SDAT patients. The average error of ANN model for HDS-R score was 2.64 points out of 30. In conclusion these models are the useful tools in order to distinguish SDAT patients from non SDAT patients and quantify the severity of SDAT from EEG.