Abstract: Objective To investigate the characteristic patterns of neural oscillation of working memory (WM) dysfunction in temporal lobe epilepsy (TLE) with depression. Methods TLE patients were divided into two groups: TLE with depression (n=10) and TLE without depression (n=10), according to the Beck depression inventory Ⅱ (BDI-Ⅱ) and the Hamilton’s Depression Scale (HAMD)-17 score. The matched healthy controls were also included in this study (n=10). The 34-channel EEGs were recorded in three groups of patients while they performed a visual WM task. The reaction time and accuracy were analyzed in there groups. For EEGs analysis, noises were removed from the original EEGs. Short-time Fourier transform (STFT) was used to obtain the time-frequency distribution and the principal frequency band and time period for determined WM. The topographical maps in the principal frequency band and time period were presented, respectively. The power was then compared quantitatively between three groups. Results Compared with controls, the WM task accuracy was decreased in TLE with depression group and TLE without depression group (P＜0.05). The reaction time was decreased in turn in TLE with depression group, TLE without depression group and control group (P＜0.05). In the working memory delay phase, the percent of theta power in the total energy percentage of 34 channel EEGs was significantly lower in TLE with depression group and TLE without depression group than that in control group (P＜0.01). The EEGs spectrum showed that theta power in frontal lobe (especially in Fz channel) was decreased in TLE with depression group compared with that of controls and TLE without depression group (P＜0.05). Conclusion Both TLE patients with and without depression have WM dysfunction. Moreover, the TLE patients with depression are more injured. The absence of theta oscillation in the frontal lobe is one of the potential neural mechanisms for WM dysfunction in TLE patients with depression.

Key words: epilepsy, temporal lobe, depression, memory disorders, electroencephalography, signal processing, computer-assisted, theta oscillation