Recent estimates indicate that the Alzheimer’s disorder may rank third, just behind heartdisease and cancer, as a cause of death for older people, affecting more than 50 million patientsaround the world. A significant way for identifying and analyzing Alzheimer activity in humansis by using Electroencephalogram (EEG) signal. In this work, EEG signals will classified asnormal (healthy) signals or as Alzheimer signal using an automated system using BBNN andPSO. This work focuses on the analysis of EEG signal based on BBNN and PSO. Once the EEGsignals are analyzed, the features will be extracted using Discrete Wavelet Transform (DWT)and from the selected features, a BBNN is trained for and on the basis of training samples, testsamples are classified accordingly. The parameters of BBNN are optimized using PSOalgorithm. Based on this, the class of the input signal is predicted using BBNN. Finally, theperformance parameters such as classification accuracy, specificity, sensitivity of the automaticclassification system for the EEG signals proposed, will be measured and performance of thesystem will be evaluated.The proposed method is automatic and hence it is not subjective and thereby eliminates the needfor the visual inspection based method which is subjective. Moreover, the proposed methodoffers better performance than the existing visual inspection based method of EEG signalclassification. An EEG signal is analyzed and fed to a classifier. The input signal received by theclassifier, uses it for classifying on the basis of input signals received during the training phase.Proposed system uses a novel classification method using Block Based Neural Network and theparameters are optimized using Particle Swarm Optimization (PSO). The proposed system hasfive stages Signal Acquisition, Preprocessing, Feature Extraction, Feature Selection andClassification. The proposed architecture is shown in Figure 1. It shows how each of the phasesis related with its predecessor phases.