Test–retest reliability of N400 event-related brain potential measures in a word-pair semantic priming paradigm in patients with schizophrenia

The N400 event-related brain potential (ERP), a negative voltage deflection occurring approximately 400ms after onset of any meaningful stimulus, is reduced in amplitude when the stimulus is preceded by related context. Previous work has found this N400 semantic priming effect to be decreased in schizophrenia, suggesting impairment in using meaningful context to activate related concepts in semantic memory. Thus, N400 amplitude may be a useful biomarker of abnormal semantic processing and its response to treatment in schizophrenia. To help assess the validity of N400 amplitude as a longitudinal measure in schizophrenia, we evaluated its test-retest reliability. ERPs were recorded in sixteen schizophrenia patients who viewed prime words, each followed at 300- or 750-ms stimulus-onset asynchrony (SOA) by a target that was either a related or unrelated word, or nonword. Participants' task was to indicate whether or not the target was a real word. They were retested on the same procedure one week later. Test-retest reliability was assessed by calculating Pearson's r and intraclass correlation coefficients (ICCs) across timepoints for N400 amplitudes for related and unrelated targets, at each SOA. Consistent with previous results, there were no significant differences between patients' N400 amplitudes for related and unrelated targets, at any SOA/timepoint combination. Pearson's r and ICCs for N400 amplitudes at Fz across timepoints were significant for both target types at each SOA (ranges: r 0.52-0.64, ICC 0.52-0.63; all p<.04). The results suggest potential utility of N400 amplitude as a longitudinal neurophysiological biomarker of semantic processing abnormalities in schizophrenia.