Objective To investigate the changes in electroencephalogram (EEG) signals during seizures of juvenile myoclonic epilepsy (JME) using wavelet entropy and explore the dynamic mechanism of JME, and to provide a better indicator for monitoring seizures.Methods The seizure EEG signals of 10 patients with JME were decomposed with Gaussian continuous wavelet transform. On this basis, we analyzed the wavelet power spectrum and calculated the integral wavelet entropy (En) and individual-scale wavelet entropy [En(a)]. According to the visual EEG analysis, the wavelet entropy was categorized into inter-ictal, pre-ictal and ictal phases. We observed the dynamic changes of wavelet entropy during the entire period of seizure of JME and compared the change pattern with normal controls.Results The integral wavelet entropy in ictal phase was significantly lower than those in inter-ictal phase and normal controls (P<0.017), especially in the frontal region. The 11th scale wavelet entropy of JME in ictal phase was significantly higher than those in inter-ictal phase and normal controls (P<0.017), while the opposite was observed for the 9th scale wavelet entropy (P<0.017).Conclusions The lower integral wavelet entropy in ictal phase of JME suggests simple and regular activity and reduced richness and complexity of the whole brain as compared with inter-ictal phase and normal controls. The increase in the 11th scale wavelet entropy and the decrease in the 9th scale wavelet entropy in ictal phase indicate increased frequency of brain electrical activity in the 11th scale and decreased frequency in other scales. The integral wavelet entropy of different leads in ictal phase decreased to various extents, and the greater decrease in the frontal region suggests the frontal cortex as a source of discharges in JME.