Electrophysiological evidence for primary semantic memory functional organization deficits in schizophrenia

Abstract

N400, an event-related brain potential (ERP) waveform elicited by meaningful stimuli, is normally reduced by stimulus repetition (N400 repetition priming), and relatedness between the eliciting stimulus and preceding ones (relatedness priming). Schizophrenia patients' N400 relatedness priming deficits suggest impairment in using meaningful prime stimuli to facilitate processing of related concepts in semantic memory. To examine whether this deficiency arises from difficulty activating the prime concept per se, as indexed by reduced N400 repetition priming; or from impaired functional connections among concepts in semantic memory, as reflected by reduced relatedness priming but normal repetition priming; we recorded ERPs from 16 schizophrenia patients and 16 controls who viewed prime words each followed at 300- or 750-ms stimulus-onset asynchrony (SOA) by an unrelated, related or repeated target word, or a nonword, in a lexical-decision task. In both groups, N400s were largest (most negative) for unrelated, intermediate for related, and smallest for repeated targets. Schizophrenia patients exhibited subnormal N400 relatedness priming at the 300-ms SOA, but normal repetition priming at both SOAs, suggesting that their impairment in using prime words to activate related concepts results from abnormal functional connections among concepts within semantic memory, rather than inability to activate the prime concept itself.

Figure 1

Grand average ERPs to target words at the 300-ms prime-target SOA, at all electrode sites, for: (a) schizophrenia patients and (b) NCPs. Negative amplitudes are plotted upward.

Figure 1

Grand average ERPs to target words at the 300-ms prime-target SOA, at all electrode sites, for: (a) schizophrenia patients and (b) NCPs. Negative amplitudes are plotted upward.

Figure 2

Grand average ERPs to target words at the 750-ms prime-target SOA, at all electrode sites, for: (a) schizophrenia patients and (b) NCPs. Negative amplitudes are plotted upward.

Figure 2

Grand average ERPs to target words at the 750-ms prime-target SOA, at all electrode sites, for: (a) schizophrenia patients and (b) NCPs. Negative amplitudes are plotted upward.

Figure 3

Grand average relatedness priming effects (formed by subtracting ERPs for related targets from ERPs for unrelated targets), at midline electrode sites, for schizophrenia patients and NCPs, for: (a) the 300-ms SOA and (b) the 750-ms SOA. Negative amplitudes are plotted upward.

Figure 3

Grand average relatedness priming effects (formed by subtracting ERPs for related targets from ERPs for unrelated targets), at midline electrode sites, for schizophrenia patients and NCPs, for: (a) the 300-ms SOA and (b) the 750-ms SOA. Negative amplitudes are plotted upward.

Figure 4

Grand average repetition priming effects (formed by subtracting ERPs for repeated targets from ERPs for unrelated targets), at midline electrode sites, for schizophrenia patients and NCPs, for: (a) the 300-ms SOA and (b) the 750-ms SOA. Negative amplitudes are plotted upward.

Figure 4

Grand average repetition priming effects (formed by subtracting ERPs for repeated targets from ERPs for unrelated targets), at midline electrode sites, for schizophrenia patients and NCPs, for: (a) the 300-ms SOA and (b) the 750-ms SOA. Negative amplitudes are plotted upward.

Figure 5

Amplitudes of N400 relatedness and repetition priming effects (averaged across nine medial posterior sites: Cz, P7, P3, Pz, P4, P8, PO3, PO4 and Oz) for patients and NCPs. Error bars indicate standard deviations.