The clinical picture of salicylate-poisoning is characterized by neurological symptoms reaching from mental slowing, confusion, hallucinations to convulsions, We describe a patient, who showed EEG abnormalities in the form of mild diffuse slowing with additional intermittent slow activity on repeated admissions, These abnormalities rapidly disappeared together with the clinical symptoms of mental slowing, ataxia, dysarthria and confusion. Eventually, a chronic salicylate intoxication was revealed. It is suggesed that EEG abnormalities serve as an additional diagnostic clue in this potentially fatal condition.

The investigation was aimed at the question whether or not the N400-component of the event-related brain potential and reaction time index the same processes in semantic priming experiments. To this end, word lists comprised of exemplars of ten common taxonomic categories were constructed and structured such that words that came from the same category as their predecessor and words that came from a different category were equally probable. Furthermore the sequence of the preceding items was pseudo-randomized yielding equal numbers of each of sixteen possible sequences of four words. Subjects had to judge the category-membership of each word as compared to the preceding word (same/different category, speeded button press). Event-related potentials were obtained in twelve subjects from several scalp sites and averaged according to the sequences of preceding words. Reaction times varied systematically as a function of the preceding words whereas for the N400 amplitude only a difference for the "category same" (amplitude low) and "category different" (amplitude high) events was found. No effect of the sequence of preceding events was found. It is concluded that N400 and reaction time index are different processes in semantic priming experiments. Furthermore the constant N400 amplitude irrespective of the sequence of preceding events strongly supports the view that N400 is different from N2.

A number of recent investigations emphasized the sensitivity of event-related potentials (ERP) in visuospatial selective attention tasks. If a spatial location is cued by a central (symbolic) stimulus such as a left or right-pointing arrow an amplitude enhancement for the P1, N1 and N2 components of the ERP to target stimuli appearing on the cued side relative to target stimuli appearing on the uncued side has been described. In the current study peripheral cues were used with a long (1000 ms) cue target interval. Three questions were addressed: 1. How do the effects using peripheral cues compare to those described for central cues? 2. Can similar cue-validity effects be observed for auditory and visual cues? 3. Can electrophysiological correlates of attentional orienting be revealed by examining the ERPs in the cue-target interval? In the condition with visual cues in addition to a marked reaction time effect an enhancement of the posterior N1 and N2 was observed for the validly cued targets. No effect for the P1 component was found. This pattern suggests similarities of attentional orienting after central and peripheral cues. In contrast, the auditory cues did not produce any behavioral effects, while a frontal ERP-effect (N2-modulation) was found as a function of cue-validity. The psychological significance of this effect remains unclear. In the cue-target interval no differential effects of cue-location were seen in the auditory condition. In the visual condition a pronounced negativity with an onset of about 500 ms was seen contralaterally to the cue at the parietal locations. This effect is discussed in light of the proposed role of the parietal lobes for attentional orienting.

In this overview recent developments of the analysis of language perception by means of event-related potentials (ERP) are highlighted. Some of the more popular designs, including the sentence reading and lexical decision paradigms are described and their respective benefits and shortcomings are pointed out. The questions that can be addressed with ERPs are exemplified in two cases: 1) The current trend towards the analysis of syntactic processes has yielded several ERP components that appear to be sensitive to the syntactic representation of a sentence. This research has revealed - aside from late positive effects - a frontal negativity in response to syntactic errors. This finding has been interpreted as being supportive of modular models of language processing. 2) The degree to which certain subroutines of language processing are shared with other cognitive domains (e.g. memory) can be addressed by comparing ERP findings from repetition and recognition paradigms using word and pictorial material with other language experiments. Some of the current lines of interpretation are reiterated with special emphasis on the N400 being a marker of a context-integration process. New possibilities of the analysis of ERPs such as fine grain topographical analysis and dipole localization methods are illustrated. Applications of language related ERPs in patients are outlined. Data from brain-injured patients show a delayed and reduced N400 component in a sentence reading task and an absent N400 component in a lexical decision task.

There is a hotly debated question as to whether the meaning of a word can be assessed in the lexicon before conscious identification of this word is possible. Some authors have reported shorter lexical decision times for words, which had been preceded by primes, that had been masked below threshold. This effect was taken as evidence for an automatic activation of the lexicon by the unidentified prime. We report an ERP study using the N400 potential and lexical decision time as an index for associative activation of the lexicon. In a condition with supra-threshold primes the typical acceleration of lexical decision time and reduction of N400 amplitude was found for semantically related target words. In the sub-threshold condition, however, no evidence was found for priming in RT and ERP data. This pilot data, as well as previous RT studies, question the findings gathered on sub-threshold priming. No dissociation between RT and ERP parameters was found in the present study.