电生理图蕴涵机体生理过程的周期性信息和时间先后的电位变化相关性信息，对于诊断心脏或大脑活动的异常具有重要价值。本研究对ECG/EEG信号进行极大重叠离散小波变换（MODWT）后，提取方差、相关系数、四分位数间距、Shannon熵这4个特征，进行特征选择，进而建立判别模型，对心血管疾病或脑神经系统疾病进行鉴别诊断。本课题组基于PhysioBank标准数据库的心肌梗死、心律失常、睡眠障碍的数据，分别探索本方法的可行性。进一步拟建立不同疾病的ECG/EEG信号及相关生理指标的实例数据库，综合更多的信息进行更准确的疾病判别。本研究结合电信号特征信息和生理指标进行定量分析，为心血管疾病或脑神经系统疾病的自动诊断和疾病鉴别诊断标准（如心肌梗死、癫痫等）的完善提供依据。

Electrophysiological diagram is of great value in diagnosing cardiac or cerebral abnormalities,which contains the periodic information of physiological processes and time-dependent information reflecting the electrical conduction of myocardial cells and ions flow.We apply Maximal Overlap Discrete Wavelet Transform (MODWT) to ECG/EEG signal, and extract the four features including variance, correlation coefficient, interquartile range and Shannon entropy.Then feature selection and discriminant model are carried out to differentiate between the signal patterns of healthy subjects and patients with specific heart or brain disorders. Based on myocardial infarction, arrhythmia and sleep disorders datasets from PhysioBank database, we explore the feasibility of this method. Furthermore, we will establish a database of ECG/EEG signals and corresponding physical symptoms so that more information can be used to discriminate diseases accurately. The approach developed will be a feasible classification strategy to diagnose abnormal ECG/EEG automatically and we intend to provide the perfection of diagnostic guidelines.