A cortical network for semantics: (de)constructing the N400

Key Points The N400 response is one of the best-studied ERP components associated with language processing. Manipulations such as semantic priming ('coffee' → 'tea') and contextual fit ('I like my coffee with cream and sugar/socks') modulate the amplitude of the N400 response to words; this is referred to as the 'N400 effect'. Two contrasting hypotheses address the functional interpretation of the N400 effect. The effect may reflect differences in the amount of effort needed to semantically integrate the incoming word into the previous sentential context; alternatively, it may reflect facilitated lexical access due to pre-activation by supportive contexts. Previous imaging and patient data implicate the posterior middle temporal cortex in storing and activating information associated with lexical representations. Some evidence suggests that anterior temporal and inferior parietal cortices are involved in constructing new inputs and/or integrating them into temporary representations of the larger syntactic and semantic structure, and that the inferior frontal cortex subserves control processes of retrieval and selection of lexical representations that make the input available to integrative processes. The functional anatomic model outlined leads to predictions about the locus of the N400 effect: if the effect reflects integration difficulty, it should localize to inferior frontal, anterior temporal or inferior parietal regions; if the effect reflects facilitated lexical access, it should localize to the posterior middle temporal cortex. Functional MRI studies show semantic priming effects in the inferior frontal cortex only when there is a long interval between primes and targets. Priming effects are observed in the posterior middle temporal gyrus across all prime–target intervals, as is the N400 effect. Magnetoencephalography and event-related optical signal (EROS) studies of sentence context show semantic-priming and semantic-anomaly effects in the posterior temporal cortex at the same latency as the N400 effect. The data from across these techniques provide strong evidence that the posterior middle temporal cortex is involved in generating the N400 effect, which supports the claim that the N400 effect is at least to some degree due to facilitated lexical access, and argues against an account in which the effect is purely integrative. Event-related potentials (ERPs) have been widely used to investigate language processing in the brain. Lau and colleagues