Brain potentials during sentence verification: automatic aspects of comprehension

College students learned a set of facts relating fictitious people and their occupations (e.g. 'Matthew is a lawyer'). Event-related brain potentials (ERPs) were recorded while they subsequently viewed a series of such statements presented in segments (e.g. 'Matthew/is a/dentist'). ERPs to occupations completing statements falsely were significantly more negative than those to true statements in an interval 200-420 msec poststimulus (peak N320), whether subjects were required to make a decision about each statement or passively view the presented segments (Experiments 1 and 2). A later ERP positivity was observed during 'response' trials that was of longer latency for false than true completions; but this positive component was greatly attenuated during 'no-response' trials. The enhanced N320 for false completions was not affected by requiring subjects on some trials to respond incorrectly (Experiment 3). It is concluded that attending to a presented word results in an automatic analysis of its meaning in the context of a preceding verbal input, and that ERPs can indicate the nature of the output of that analysis.