OBJECTIVES: Sensory gating dysfunctions in patients with schizophrenia and bipolar disorder have been investigated through two similar methods ; P50 suppression and prepulse inhibition paradigms. However, recent studies have demonstrated that the two measures are not correlated but rather constitute as distinct neural processes. Recent studies adopting spectral frequency analysis suggest that P50 suppression reflects the interaction between gamma and other frequency bands. The aim of the present study is to investigate which frequency component shows more significant interaction with gamma band. METHODS: A total of 108 mood disorder patients and 36 normal subjects were included in the study. The P50 responses to conditioning and test stimuli with an intra-pair interval of 500 msec were measured in the study population. According to P50 ratio (amplitude to the test stimulus/amplitude to the conditioning stimulus), the subjects with P50 ratio less than 0.2 were defined as suppressed group (SG) ; non-suppressed group (NSG) consisted of P50 ratio more than 0.8. Thirty-five and 25 subjects were included in SG and NSG, respectively. Point-to-point correlation coefficients (PPCCs) of both groups were calculated between two time-windows : the first window (S1) was defined as the time-window of one hundred millisecond after the conditioning auditory stimulus and the second window (S2) was defined as the time-window of 100 msec after the test auditory stimulus. Spectral frequency analysis was performed to investigate which frequency band results in the difference of PPCC between SG and NSG. RESULTS: Significant reduction of PPCC between S1 and S2 was observed in the SG (Pearson's r = 0.24), compared to PPCC of the NSG (r = 0.58, p < 0.05). In spectral frequency analysis, gamma band showed "phase-reset" and similar responses after the two auditory stimuli in suppressed and non-suppressed group. However in the case of alpha band, comparison showed significantly low PPCC in SG (r = -0.14) compared to NSG (r = 0.36, p < 0.05). This may be reflecting "phase-out" of alpha band against gamma band at approximately 50 msecs after the test stimulus in the SG. CONCLUSIONS: Our study suggests that normal P50 suppression is caused by phase-out of alpha band against gamma band after the second auditory stimulus. Thus it is demonstrated that normal sensory gating process is constituted with attenuated alpha power, superimposed on consistent gamma response. Implications of preserved gamma and decreased alpha band in sensory gating function are discussed.