Research on the treatment of Alzheimer’s disease based on the analysis of EEG signals can help to cope with the aging of the population. In this paper, the EEG signals of patients were collected as research data by combining channel sampling methods. For the existence of interference artifacts, downsampling method combined with bandpass filter, and independent component analysis are used for signal preprocessing. A time series channel complex network based on deep optimization algorithm is constructed to explore the nonlinear features of EEG signals, and the t-test reveals the pathogenesis of different EEG signal frequency bands. EEG signal acquisition of Alzheimer’s disease patients, the characterization of time-level changes in sub-brain regions, and cognitive therapy control-related practices are carried out to verify the validity of the method in this paper. The results showed that the brain regions with more obvious changes in music stimulation were located in the parietal lobe and temporal lobe for the three indexes of arrangement entropy, sample entropy, and Lz complexity in mild to moderate patients (p<0.05). The regions with more pronounced changes in severe patients were located only in the temporal lobe. The relative power of the frequency band EEG signals of the patients in the experimental group after treatment were significantly higher than those of the control group and before treatment (p<0.05).
