On the matching of top-down knowledge with sensory input in the perception of ambiguous speech.

BACKGROUND: How does the brain repair obliterated speech and cope with acoustically ambivalent situations? A widely discussed possibility is to use top-down information for solving the ambiguity problem. In the case of speech, this may lead to a match of bottom-up sensory input with lexical expectations resulting in resonant states which are reflected in the induced gamma-band activity (GBA). METHODS: In the present EEG study, we compared the subject's pre-attentive GBA responses to obliterated speech segments presented after a series of correct words. The words were a minimal pair in German and differed with respect to the degree of specificity of segmental phonological information. RESULTS: The induced GBA was larger when the expected lexical information was phonologically fully specified compared to the underspecified condition. Thus, the degree of specificity of phonological information in the mental lexicon correlates with the intensity of the matching process of bottom-up sensory input with lexical information. CONCLUSIONS: These results together with those of a behavioural control experiment support the notion of multi-level mechanisms involved in the repair of deficient speech. The delineated alignment of pre-existing knowledge with sensory input is in accordance with recent ideas about the role of internal forward models in speech perception.

Top-down knowledge supports the retrieval of lexical information from degraded speech.

How is it that the human brain is capable of making sense from speech under many acoustically compromised conditions? The support through top-down knowledge is inevitable but can we identify brain measures of this matching process between degraded auditory input and possible meaning? To answer these questions, the present study investigated the modulation of the induced gamma-band activity (GBA) in the auditory domain in response to degraded speech. During an EEG experiment subjects first listened to digitally degraded unintelligible speech signals (derived from German nouns). In an exposure sequence, half of the nouns were presented in a non-degraded intelligible format and memorized, while in the crucial test sequence subjects listened to all degraded speech signals again and were asked to identify the words. The induced GBA (40-Hz range) showed an increase at left temporal electrode sites around 350 ms only for words correctly identified in the test sequence. No differences in induced GBA were evident in the baseline sequence; neither did the evoked brain potentials yield any comparable effect. We conclude that the observed enhancement in induced gamma-band activity reflects a matching process of top-down lexical memory traces with degraded sensory input to form a comprehendible speech percept. The findings are highly corroborant to analogous studies in the visual system. They lend further plausibility to a left-lateralized fronto-temporal network enabling lexically guided speech perception, and they demonstrate the complementary role of time-sensitive brain analyses in discerning the functional neuroanatomy of speech.

Recovery from aphasia as a function of language therapy in an early bilingual patient demonstrated by fMRI.

Knowledge about the recovery of language functions in bilingual aphasic patients who suffer from left-hemispheric stroke is scarce. Here, we present the case of an early bilingual patient (German/French) with chronic aphasia. Functional magnetic resonance imaging (fMRI) was used to investigate neural correlates of language performance during an overt picture naming task in German and French (a) 32 months after stroke to assess differential recovery of both languages as a function of the preceding language therapy that was provided exclusively in German and (b) after additional short-term intensive (German) language training. At the first investigation behavioral performance confirmed selective recovery of German naming ability which was associated with increased functional brain activation compared to the French naming condition. Changes in behavioral performance and brain activation pattern as disclosed by fMRI after an additional experimental treatment were confined to the trained (German) language and indicate bilateral neuroplastic reorganization. No generalization to the untrained (French) language was observed. The present case results demonstrate use and/or training-dependent differential recovery of expressive language functions and an enhanced pattern of brain activation as a function of the rehabilitation efforts that were focussed exclusively on the patient's German language abilities.

Gender differences in functional hemispheric asymmetry during processing of vowels as reflected by the human brain magnetic response.

A number of findings indicate gender differences in language-related functional hemispheric brain asymmetry. To test if such gender-specific laterality is already present at the level of vowel-processing, the auditory evoked magnetic field was recorded in healthy right-handed male and female participants in response to the German synthetic vowels [a], [e] and [i]. Female participants exhibited stronger N100m responses than male participants over the left hemisphere. This observation was highly reliable across repeated experimental sessions. The present lateralization shows that previous findings suggesting a stronger left-hemispheric dominance for verbal material in males than in females can not be generalized to basic speech elements. Furthermore, the present results support the importance of controlling for gender ratio in studies of phonetic processing.

Grapheme monitoring in picture naming: an electrophysiological study of language production.

Electrocortical correlates of language production were examined in two picture naming tasks that involved grapheme monitoring. In both tasks subjects (N=12) had to detect target letters in picture names, the target letter being positioned either at the beginning or at the end of the picture name. Between tasks, the target letter was shown either before (target-picture, TP) or after (picture-target, PT) the presentation of the object pictures. In both tasks, subjects responded faster, whenever the target letter appeared at the beginning of the picture name than at its end. The EEG, recorded from 64 electrodes, was analyzed in the signal and in the source space (using the Minimum Norm estimate). Differences in the event-related potential (ERP) following the second stimulus became evident earlier in the PT (at 320 ms) than the TP (456 ms) task. This onset of diverging ERPs was called the "point of divergence" (POD). The ERP following the POD was characterized by a positive deflection in the "begin" condition in both tasks. In the "end" condition, the sources of brain activity were focused over the left hemisphere in the TP, while a bilateral distribution characterized the PT task. Performance and electrocortical indices support the hypothesis of serial "left-to-right" processing of a representation of the picture name. The left-hemispheric activity focus in the TP task is assumed to indicate the encoding of the picture name, while frontal symmetrical activity in the PT task might indicate the involvement of working memory processes.

Altered hemispheric asymmetry of auditory magnetic fields to tones and syllables in schizophrenia.

BACKGROUND: A growing body of literature suggests that schizophrenic patients often do not show the normal brain hemispheric asymmetry. We have found this for simple tones presented to the right ear in a previous study. In this study we extended this investigation to left ear stimulation and verbal stimuli. METHODS: With a whole-head neuromagnetometer, contra- and ipsilateral auditory-evoked magnetic fields in response to tones (1000 Hz) and to the syllables ("ba") delivered to the left and right ears in separate runs were compared between schizophrenic patients (n = 17) and healthy control subjects (n = 15). RESULTS: In response to tones, all control subjects showed the expected asymmetry (contralateral predominance) of the auditory-evoked magnetic N100m (dipole moment). In the patient sample asymmetry was reversed following tones presented to the left ear in 47% and following tones to the right ear in 24%. In response to syllables, the asymmetry was similar between groups. In patients compared with control subjects the N100m was located more anterior without asymmetry between hemispheres. CONCLUSIONS: Results suggest that deviation from the normal functional lateralization in schizophrenia appears in a proportion of patients at a basic stage of auditory processing, but may be compensated for at higher levels such as the processing of syllables.

Auditory temporal processing deficit in dyslexia is associated with enhanced sensitivity in the visual modality.

Developmental dyslexia has been associated with a deficit in temporal processing, but it is controversial whether the postulated deficit is pansensory or limited to the auditory modality. We present psychophysical assessment data of auditory and visual temporal processing abilities in children with dyslexia. While none of the dyslexic children displayed temporal processing abnormalities in the visual sensory modality, dyslexics with poor auditory temporal scores reached high-level visual performance. Our results do not confirm the hypothesis of a general temporal processing deficit for dyslexia but suggest that limitations in auditory temporal processing might be compensated for by a well-functioning visual sensory modality.

Atypical organisation of the auditory cortex in dyslexia as revealed by MEG.

Neuroanatomical and -radiological studies have converged to suggest an atypical organisation in the temporal bank of the left-hemispheric Sylvian fissure for dyslexia. Against the background of this finding, we applied high temporal resolution magnetoencephalography (MEG) to investigate functional aspects of the left-hemispheric auditory cortex in 11 right-handed dyslexic children (aged 8-13 years) and nine matched normal subjects (aged 8-14 years). Event-related field components during a passive oddball paradigm with pure tones and consonant-vowel syllables were evaluated. The first major peak of the auditory evoked response, the M80, showed identical topographical distributions in both groups. In contrast, the generating brain structures of the later M210 component were located more anterior to the earlier response in children with dyslexia only. Control children exhibited the expected activation of more posterior source locations of the component that appeared later in the processing stream. Since the group difference in the relative location of the M210 source seemed to be independent of stimulus category, it is concluded that dyslexics and normally literate children differ as to the organisation of their left-hemispheric auditory cortex.

Electroencephalographic activity over temporal brain areas during phonological encoding in picture naming.

OBJECTIVES: The present study examined electroencephalographic (EEG) correlates of phonological encoding during picture naming with special emphasis on hemispheric asymmetries of these EEG correlates. We also examined whether a small set of stimuli was sufficient to study the phonological encoding, and to what extent the complexity of the produced message affects the EEG responses. METHODS: Event-related electrical brain activity during the covert and overt production of names and nominal phrases derived from 16 variants from 4 different pictures was compared with that during passive viewing of the same pictures. RESULTS: Topographical and source analyses of the differential EEG activity (naming versus passive viewing) indicated that the N1 and P2 components of the visual evoked potential resulted from the same brain areas, but were activated stronger during naming as compared to passive viewing. In contrast, the differential EEG activity from 275 to 400 ms suggested the involvement of additional brain areas during naming with more pronounced left- than right-hemispheric activation in middle and posterior temporal regions for both overt and covert naming. CONCLUSIONS: Results suggest the involvement of Wernicke's area in the phonological encoding of a message during speech production, which can even be obtained with a small set of pictures and a large number of repetitions.

Magnetic brain activity evoked and induced by visually presented words and nonverbal stimuli.

Evoked and induced magnetic brain activity measured over the left hemisphere were tested for their specificity to language-related processing. Induced activity refers to oscillatory alterations time locked but not phase locked to the stimulus. Words, false font stimuli, and two types of nonverbal patterns were presented visually while subjects performed a nonlinguistic visual feature detection task. The comparison of evoked and induced brain activity around 200 ms after stimulus onset revealed differential sensitivity to the stimuli. The M180 component of the evoked magnetic field was larger at the processing of words and false font stimuli compared with nonverbal stimuli. The induced magnetic brain activity in the 60-Hz band at a compatible latency range was correlated with the familiarity of the visual Gestalt. Sensitivity to language-specific information processing can be concluded if a parameter differentiates the word condition from the nonlexical conditions. Such a difference was observed at sensors located over the frontal-temporal scalp regions for induced but not evoked magnetic brain activity. Thus, evoked and induced magnetic brain activity revealed a differential sensitivity to elements of cognitive processing during the given task.

Slow cortical DC-potential responses to sweet and bitter tastes in humans.

Processing of hedonic stimulus quality is assumed to be accompanied by a tuning of cortical arousal and excitability. In this pilot study in 11 healthy humans scalp-recorded DC potentials were assessed during application of a sweet (sucrose) and bitter (quinine hydrochloride) taste, i.e., primary reinforcers of positive and negative quality. Muscular, ocular, and skin potential activity were controlled. Application of sucrose induced a widespread positive DC-potential shift with an amplitude of 40-50 microV and persisting for more than 120-s post-stimulus onset. Following administration of quinine hydrochloride, this positive shift was reduced, most distinctly between 48- and 88-s post-stimulus onset. The reduction appeared to be most consistent at anterior midline recording sites (Fz, Cz). It is assumed that the higher DC-potential positivity during sweetness than during bitterness points to a differential tuning of cortical excitability by a widespread decrease in depolarization of apical dendrites.

Mobile phones modulate response patterns of human brain activity.

Mobile phones emit a pulsed high-frequency electromagnetic field (PEMF) which may penetrate the scalp and the skull. Increasingly, there is an interest in the interaction of this pulsed microwave radiation with the human brain. Our investigations show that these electromagnetic fields alter distinct aspects of the brain's electrical response to acoustic stimuli. More precisely, our results demonstrate that aspects of the induced but not the evoked brain activity during PEMF exposure can be different from those not influenced by PEMF radiation. This effect appears in higher frequency bands when subjects process task-relevant target stimuli but was not present for irrelevant standard stimuli. As the induced brain activity in higher frequency bands has been proposed to be a correlate of coherent high-frequency neuronal activity, PEMF exposure may provide means to systematically alter the pattern fluctuations in neural mass activity.

Monitoring brain activity of human subjects during delayed matching to sample tasks comparing verbal and pictorial stimuli with modal and cross-modal presentation: an event related potential study employing a source reconstruction method.

The time course of the event related potentials evoked within a delayed matching to sample task employing verbal and pictorial stimuli was analyzed with a source reconstruction method (minimum norm method). During signal stimulus presentation pictorial stimuli evoked more activity than verbal stimuli. Activity was particularly prominent in left frontal areas for the match of verbal-verbal stimulus pairs and over right posterior regions for the match of verbal-pictorial stimuli. Anticipation of the to-be-matched stimulus produced more pronounced activity for pictorial stimuli and generally stronger left and frontal activity. Results are discussed referring to a biological model of language processing.

Methohexital-induced changes in spectral power of neuromagnetic signals: beta augmentation is smaller over the hemisphere containing the epileptogenic focus.

Previous research has suggested that methohexital, a short-term barbiturate, alters activity in the primary epileptogenic area. It can be assumed that drug-induced activation of the epileptogenic focus provides a rapid and safe method to obtain a sufficient amount of information relevant for the lateralization and localisation of the primary epileptogenic area. This study shows that methohexital changes spectral power in the beta band derived from magnetoencephalographic (MEG) signals over the hemisphere ipsilateral to the primary epileptogenic area. This effect was demonstrated for 10/13 of the investigated patients suffering from unilateral temporal lobe epilepsy (TLE). The side and location of the primary epileptogenic area of these patients (5 left TLE, 8 right TEL) was determined invasively during presurgical evaluation. During a 1-2 minute interval after intravenous bolus injection of 100 mg methohexital a clear lateralization effect in the beta band was observed, which differed marginally between fronto-central, fronto-temporal and temporo-parietal brain regions. In addition, bilateral spectral power changes were obtained in the theta, alpha and gamma bands that differed between brain regions. Analyses of simultaneously recorded scalp electroencephalographic (EEG) data revealed effects consistent with those of the MEG analysis. The reduced enhancement of beta band spectral power of MEG recordings provides a potential application for the non-invasive lateralization of the primary epileptogenic area.

Differential outcomes from magneto- and electroencephalography for the analysis of human cognition.

Theoretical considerations show that magnetoencephalography (MEG) and electroencephalography (EEG) provide different information about ongoing human brain activity. The paper presents simultaneously measured MEG and EEG data showing that these measures may lead to different conclusions about cognitive models under investigation. This was demonstrated for amplitude results of the P300/N400 complex in a study of the secondary processing of lexical and non-verbal information in visual stimuli. As both methods provide different information about ongoing brain activity, their combined analysis is valuable. This seems particularly true for studies of higher order cognitive processing.

Binaural fusion and the representation of virtual pitch in the human auditory cortex.

The auditory system derives the pitch of complex tones from the tone's harmonics. Research in psychoacoustics predicted that binaural fusion was an important feature of pitch processing. Based on neuromagnetic human data, the first neurophysiological confirmation of binaural fusion in hearing is presented. The centre of activation within the cortical tonotopic map corresponds to the location of the perceived pitch and not to the locations that are activated when the single frequency constituents are presented. This is also true when the different harmonics of a complex tone are presented dichotically. We conclude that the pitch processor includes binaural fusion to determine the particular pitch location which is activated in the auditory cortex.

Oscillatory neuromagnetic activity induced by language and non-language stimuli.

Event-related oscillatory brain activity during language perception differs from activity occurring during the processing of comparable non-language stimuli. This fact became apparent in the observation of changes in the normalized spectral power of magnetoencephalographic (MEG) signals during the subject's processing of these stimuli. MEG was recorded over the left and right hemispheres of 12 right-handed subjects. During the experimental session, bisyllablic content words and physically similar non-language stimuli were presented with equal probability in a randomized order in either the visual or auditory modality. Approximately 15% of these stimuli were marked and the subject's task was to detect these marked stimuli. As a major characteristic of language vs. non-language processing, we obtained an enhancement of the normalized spectral power around 240 ins in the 60-65-Hz band over the left hemisphere for the language condition and over the right hemisphere for the non-language condition, independent of the modality of stimulus presentation. Starting at approximately the same latency but in lower-frequency bands (15-45-Hz), an extended (250-600 ms) reduction of normalized spectral power was observed. This reduction, although it generally confirmed previous results, differed in the no hemisphere-specific reduction was found for the processing of words. A domain-specific enhancement of normalized spectral power was also evident around 800-1200 ms in the 15-30-Hz band. In the auditory condition, this enhancement of the normalized spectral power was larger after the presentation of language stimuli whereas in the visual condition a larger enhancement of the normalized spectral power was obtained after presentation of non-language stimuli. As this latter effect appears relatively late after the stimulus onset and differs in expression for both modalities of stimulus presentation, a simple relationship between language perception and oscillatory brain dynamics can be excluded for this enhancement. In contrast, the left hemispheric enhancement of the normalized spectral power present around 240 ms in the 60-65-Hz band seems to reflect oscillatory pattern specific to the processing of words.

The auditory evoked "off" response: sources and comparison with the "on" and the "sustained" responses.

OBJECTIVE: It is well known that tone bursts elicit a prominent N1/P2 complex in the auditory evoked potential (the on-response), but less is known about a morphologically similar complex (the off-response) that can be recorded under suitable stimulus conditions. The interaction between the two responses indicated that the responses were not physiologically independent. The present experiment employed neuromagnetic methods to determine the cortical sources of N1 and P2 on- and off-responses and their relation to other events observed in the auditory evoked field. DESIGN: Five female and five male subjects with no history of otologic or neurological disorders and with normal audiological status participated in this study. Tone bursts of 2 sec duration (10 msec rise and decay time, cosine function), carrier frequency of 1 kHz, and intensity of 60 dB nHL (normative hearing level) were presented 512 times to the subject's right ear (contralateral to the investigated hemisphere) with an interstimulus interval randomized between 5 and 7 sec. RESULTS: The present study is unique in that several components of the complex auditory evoked response (P1, N1on, P2on, sustained-field, N1off, P2off) were recorded and localized in the same subjects and in the same experiment. The source coordinates obtained for N1 and P2 on- and off-responses indicated that the two responses are generated by overlapping cortical regions. Sources for the P2 components were situated anterior and medial to sources for the N1 components and were indistinguishable from sources for the auditory sustained-field. An early P1on event preceded the N1on (but not the N1off) response and was spatially indistinguishable from the N1on. The equivalent source strength was greater for N1on and P2on sources compared with N1off and P2off sources. CONCLUSIONS: The recoding process signaled by on-and off-responses may be a dynamic form of plasticity in the auditory cortex with a time constant on the order of hundreds of milliseconds, corresponding to the duration of sustained-responses released by acoustic changes and to the duration of the acoustic foreperiod that is necessary before on-and off-responses to acoustic changes can be observed.

The neurotopography of vowels as mirrored by evoked magnetic field measurements.

The auditory evoked neuromagnetic field elicited by synthetic specimens of the vowels [a], [ae], [u], and [i] was recorded over the left and the right hemisphere of 11 subjects. The N100m and the SF deflection of the recorded signal was submitted to equivalent current source analysis using the model of a single dipole in a spherical volume conductor. Vowel processing was hypothesized to occur in a multistage process rendering a sequence of representations of the acoustic input. Vowel representations were considered to differ among each other in the features they make salient, thus, in the kind of dissimilarity relationship they establish, and, by implication, in terms of the vowel space defined by the respective set of dissimilarities. It was investigated whether a mapping exists between at least one of a number of hypothetical vowel spaces and the cortical response space spanned by the spatial distribution of vowel evoked equivalent current dipoles. Although the spatial configuration of vowel evoked sources proved to be highly variable across subjects, the ordering of distances between N100m and SF equivalent current dipole locations turned out to correspond to the ordering of distances between the corners of a vowel trapezium. There were some, albeit weak, indications of hemispheric differences in vowel processing. The results suggest that the spatial distribution of the equivalent current dipole sources of both the N100m and the SF deflection of the neuromagnetic field elicited by vowels reflect a processing stage transitional between auditory and phonetic representation.

High-frequency cortical responses reflect lexical processing: an MEG study.

Meaningful words and matched pseudowords, such as moon vs. noom, are of equal perceptual complexity, but invoke different cognitive processes. To investigate high-frequency cortical responses to these stimuli, biomagnetic signals were recorded simultaneously over both hemispheres of right-handed individuals listening to words and pseudowords. Consistent with earlier EEG studies, evoked spectral responses recorded from the left hemisphere revealed depression of spectral power in the low gamma band (around 30 Hz) after pseudowords but not after words. Similar differences between stimulus categories were present in the beta range. These results indicate that distinct patterns of high-frequency cortical responses correspond to the different cognitive processes invoked by words and pseudowords. It is hypothesized that differential high-frequency cortical responses signal the activation or activation failure of distributed Hebbian cell assemblies representing words and other elements of cognitive processing.