Meta-analysis of P300 waveform in panic disorder.

The P300 waveform has been inconsistently linked to the maladaptive information-processing characterized in panic disorder (PD). The purpose of this study was to synthesize previous event-related potential (ERP) findings and determine whether patients with PD have significant abnormalities in the P300 wave compared to controls. We performed a systematic literature search for studies published between 1980 and 2013 that reported P300 measurements in patients with PD and controls. Effect size estimates were computed using the restricted maximum likelihood model. We identified 14 ERP studies that analyzed P300 amplitude (461 PD and 355 controls), and 11 ERP studies that analyzed latency (320 PD and 282 controls). Patients with PD had reduced P300 amplitudes compared to controls, but this difference was non significant at midline electrodes (n = 14, ES -0.16; z = -1.55, p = 0.122). However, P300 amplitude was significantly reduced when analyzing the Pz electrode independently (n = 7, ES -0.48, z = -3.92, p < 0.001). No significant differences between cases and controls in P300 latency were observed at the midline electrodes (n = 11, ES 0.11, z = 0.64, p = 0.524). This meta-analysis included non-peer reviewed literature and ERP stimuli with varying levels of emotional salience, which may have introduced bias into the analysis. There is no robust evidence that P300 latency alterations are present in patients with PD; however, there are indications of reduced amplitude at Pz relative to controls. Reductions in amplitude may be associated with reduced neural resources allocated to contextual updating, selective attention, and neural inhibition mechanisms.