Event-related brain potentials to grammatical errors and semantic anomalies

Event-related potential asymmetries during the reading of sentences

This report is an overview of the lateral distribution of event-related brain potentials (ERPs) recorded during silent reading in 7 different experiments. Both single word and cross-sentence averages revealed the presence of several ERP asymmetries. The P1, P2 and a negativity between 300 and 500 msec were found to be larger over the right than the left hemisphere. It was argued that this asymmetric negativity was due primarily to the contribution of the N400 elicited by all content words. The degree of N400 asymmetry was unaffected either by the rate of sentence presentation or the ratio of congruent to incongruent sentences but was quite sensitive to family history of left-handedness. In contrast, the P1 and P2 asymmetries were uninfluenced by lexical class or familial sinistrality.

Sentence verification and event-related brain potentials

The purpose of this study was to examine whether the decision process affected scalp recorded event-related brain potentials (ERPs). ERPs were recorded while 10 subjects evaluated the veracity of four types of Japanese sentences; true-affirmative (TA), true-negative (TN), false-affirmative (FA), and false-negative (FN). Each sentence was presented visually, one word at a time in the following order: Subject (S); object (O); and verb (V). This order of words, which is normal in Japanese grammar, allowed the ERP waveforms associated with semantic mismatch between the S and O occurring in the middle of the sentence to be separated from those elicited by the decision concerning the sentence's truth or falsity occurring at the end of the sentence. Semantic mismatch was associated with a negative component (N310) following the O-word. In addition, P3s were elicited by sentence final words. Affirmative sentences elicited larger P3s than did negative sentences, and true sentences elicited earlier P3s than did false sentences. P3s were smaller in amplitude when they followed a mismatch between the S and O words. The result suggested that by using Japanese sentences, it may be possible to examine sentence-level rather than word-level processes.

Visual and auditory event-related potentials elicited by linguistic and non-linguistic incongruities

To test the hypothesis that violation of overlearned relations between stimuli in a sequence is critical for eliciting the event-related potential (ERP) N400, we recorded scalp ERPs in sequences where interstimulus relations were either overlearned (sentences or famous songs) or not (geometric figures or scale notes): 25% of the sequences ended with an incongruous stimulus. N400 occurred after incongruous words in sentences only. It seems to index the further processing necessary to specify the nature of the mismatch detected between attended and expected stimuli, when complex information is involved. Earlier components also varied as a function of experimental factors.

Electrocortical signs of word categorization in saccade-related brain potentials and visual evoked potentials

Visual evoked potentials (VEPs) elicited by foveal presentation of words were compared to brain potentials evoked by the same words in a condition where subjects had to make a saccadic eye movement in order to perceive the words (saccade-related brain potentials, SRPs). Subjects had to categorize the words responding with a button press to stimuli belonging to the target (infrequent, P = 0.2) category. The VEP and SRP waveforms showed divergences in the early (up to 250 ms) components, but a marked similarity between the late components. Principal Component Analysis also revealed the same relationship between the two types of brain responses. Peak latency of the late SRP components measured from saccade offset showed an apparent processing advantage over the corresponding late components of VEPs. The N3 component, indexing semantic processing of visual patterns, peaked between 310 and 375 ms in the SRPs, while in the VEPs it appeared between 410 and 470 ms. The P4 component, associated with final stimulus evaluation, showed a similar latency benefit in favour of SRPs (420-500 ms vs 530-590 ms in VEPs). The mean reaction time was 74 ms shorter in the eye movement condition (measured from saccade offset) than in the VEP condition (703 vs 777 ms). The question of what kind of processes may contribute to the differences in mean RTs and to the latencies of the late components between the two conditions are discussed. We suggest that the late components (P3, N3 and P4) of the VEP and the SRP, respectively, index identical brain processes.

Event-related potentials, lexical decision and semantic priming

ERPs were recorded during a lexical decision task in order to investigate electrophysiological concomitants of semantic priming. The stimuli were 240 words and 240 nonwords presented one per trial at a fixed intertrial interval. Subjects were required to classify each stimulus as a word or nonword by pressing one of two response buttons. ERPs were recorded from 14 scalp locations, the right suborbital ridge, and the left earlobe, all referred to a balanced non-cephalic reference. RT and error data confirmed that semantic priming occurred under the conditions employed: primed words (those preceded by a semantically related word) were identified as words faster and more accurately than were unprimed words (those preceded by semantically unrelated words or nonwords). ERPs for all stimulus types were characterized by a large positivity peaking between 550 and 650 msec, preceded by a negative-going deflection peaking at approximately 400 msec. ERPs for primed and unprimed words were shown to differ significantly, diverging 200-250 msec following stimulus onset, reaching a maximum near the peak of the negative-going deflection at 400 msec. These differences were observed at locations over both hemispheres and were maximal in the centroparietal region. Although P300 latency differences between primed and unprimed words were also obtained, the priming effect on ERPs at shorter latencies could not be explained solely by P300 latency effects. Possible relationships between these ERP concomitants of semantic priming and P300, N200, and N400 were discussed.

The effects of physical and semantic incongruities on auditory event-related potentials

Twenty subjects listened to a series of simple sentences, spoken by a male voice, in which the last word was on occasions either made semantically incongruous or was unexpectedly spoken by a female voice. Averages of ERPs to the last words revealed that a consistent late negative component (N456) was associated with semantic incongruity and a late positive component (P416) with physical (voice) incongruity. The results were consistent with those in the visual modality by Kutas and Hillyard (1980a,b) and are interpreted in terms of the facilitatory and inhibitory effects of contextual priming on the processing of the words concerned. In a subsidiary experiment 6 subjects were required to repeat the last words of the same set of sentences as rapidly as possible. Verbal response latency increased by 62 msec to physically incongruous words and by 185 msec to semantically incongruous words.

Brain potentials during reading reflect word expectancy and semantic association

The neuroelectric activity of the human brain that accompanies linguistic processing can be studied through recordings of event-related potentials (e.r.p. components) from the scalp. The e.r.ps triggered by verbal stimuli have been related to several different aspects of language processing. For example, the N400 component, peaking around 400 ms post-stimulus, appears to be a sensitive indicator of the semantic relationship between a word and the context in which it occurs. Words that complete sentences in a nonsensical fashion elicit much larger N400 waves than do semantically appropriate words or non-semantic irregularities in a text. In the present study, e.r.ps were recorded in response to words that completed meaningful sentences. The amplitude of the N400 component of the e.r.p. was found to be an inverse function of the subject's expectancy for the terminal word as measured by its 'Cloze probability'. In addition, unexpected words that were semantically related to highly expected words elicited lower N400 amplitudes. These findings suggest N400 may reflect processes of semantic priming or activation.

Brain potentials during sentence verification: late negativity and long-term memory strength

Subjects decided whether self-referential statements were true or false. Event-related potentials (ERPs) associated with final words creating false statements displayed a late negativity (N340) relative to ERPs for true completions. The size of this difference between true and false statements was greater for highly familiar statements (e.g. "My name is Ira") than for less familiar ones (e.g. "I go to bed late") even after all the statements had been practised a number of times. The late negativity appears to be associated with a discrepancy between presented and remembered information, and its magnitude reflects the long-term familiarity or strength of the remembered information.