Large-scale cortical networks estimated from scalp EEG signals during performance of goal-directed motor tasks.

The evaluation of the topological properties of brain networks is an emergent research topic, since the estimated cerebral connectivity patterns often have relatively large size and complex structure. Since a graph is a mathematical representation of a network, the use of a theoretical graph approach would describe concisely the topological features of the functional brain connectivity network estimated using neuroimaging techniques. In the present study, we analyze the changes in brain synchronization networks using high-resolution EEG signals obtained during performance of a complex goal-directed visuomotor task. Our results show that the cortical network is more stable when subjects reach the goal than when they fail by hitting an obstacle. These findings suggest the presence of a possible cerebral "marker" for motor actions that result in successful reaching of a target.

Estimation of individual evoked potential components using iterative independent component analysis.

Independent component analysis (ICA) has been successfully employed in the study of single-trial evoked potentials (EPs). In this paper, we present an iterative temporal ICA methodology that processes multielectrode single-trial EPs, one channel at a time, in contrast to most existing methodologies which are spatial and analyze EPs from all recording channels simultaneously. The proposed algorithm aims at enhancing individual components in an EP waveform in each single trial, and relies on a dynamic template to guide EP estimation. To quantify the performance of this method, we carried out extensive analyses with artificial EPs, using different models for EP generation, including the phase-resetting and the classical additive-signal models, and several signal-to-noise ratios and EP component latency jitters. Furthermore, to validate the technique, we employed actual recordings of the auditory N100 component obtained from normal subjects. Our results with artificial data show that the proposed procedure can provide significantly better estimates of the embedded EP signals compared to plain averaging, while with actual EP recordings, the procedure can consistently enhance individual components in single trials, in all subjects, which in turn results in enhanced average EPs. This procedure is well suited for fast analysis of very large multielectrode recordings in parallel architectures, as individual channels can be processed simultaneously on different processors. We conclude that this method can be used to study the spatiotemporal evolution of specific EP components and may have a significant impact as a clinical tool in the analysis of single-trial EPs.

Comparison between ICA and wavelet-based denoising of single-trial evoked potentials.

Analysis of evoked potentials (EPs) on a single-trial basis allows the study of the dynamical characteristics of brain activity. However, single-trial responses are buried into the more prominent ongoing electroencephalographic (EEG) activity, and thus advanced procedures are needed to obtain the activity only of the cortical generators that are activated by the experimental task under study. We compare the effectiveness of two methods at removing extraneous activity from single-trial EPs, namely, a recently-proposed iterative procedure based on independent component analysis (ICA) and wavelet denoising, using simulated data and actual EP recordings from normal subjects, more specifically the auditory N100-P200 complex.

Early diagnosis of skin cancer based on segmentation and measurement of vascularization and pigmentation in Nevoscope images.

This work presents techniques developed for automated image segmentation and classification of skin lesions as malignant or benign based on the ground truth. For each skin lesion two images are obtained, one in each of two different modalities of epiluminescence microscopy (ELM): side-transillumination which highlights the subsurface vasculature and surface pigmentation, and cross polarization, which only highlights the details of skin surface pigmentation. The automated procedure consists of three steps: i) Segmentation of images, using three segmentation methods; ii) Selection of the most accurate segmentation results based on a weighted scoring technique; and iii) classification of the lesion as malignant or benign by verifying the presence of a ring of hypervascularity around the lesion in the side transillumination images. The segmentation results were validated against manual segmentation by an expert and the malignancy results were validated against the result from pathology.

Pallidotomy: a comparison of responders and nonresponders.

OBJECTIVE: We reviewed a prospective series of 32 unilateral, large-volume, microelectrode-guided posteroventral pallidotomies to determine the differences between responsive and nonresponsive patients. METHODS: Our patients underwent extensive pre- and postoperative evaluations. One year postoperatively, we correlated the outcomes of 25 patients with their histories, physical findings, neuropsychological assessments, and lesion characteristics to further understand the indications, limitations, and pitfalls of unilateral pallidotomy. Our group judged responsiveness by comparing the preoperative total Unified Parkinson's Disease Rating Scale off-state scores with those obtained 1 year postoperatively. A score indicating greater than 20% improvement at 1-year follow-up was rated a good outcome; improvement of greater than 40% was rated an excellent outcome. RESULTS: Although most patients sustained long-term benefits, some demonstrated little or no improvement. Patient and lesion factors influenced outcome. Younger age (<60 yr), tremor, unilateral predominance, L-dopa responsiveness, motor fluctuations with dyskinesia, and good lesion placement predicted a good response to unilateral pallidotomy. Advanced age (>70 yr), absence of tremor, increased duration of disease, reduced responsiveness to L-dopa, frontal behavioral changes, prominent apraxic phenomena, and improper lesion placement predicted a poor response. CONCLUSION: Unilateral, large-volume pallidotomy with precise lesion control provides long-lasting benefits for carefully selected patients.

Lateralization of activity associated with language function using magnetoencephalography: a reliability study.

This study was conducted to investigate the reliability of magnetoencephalography in lateralizing and localizing brain activity associated with receptive language function. Sixteen right-handed adults with no history of neurologic disorder engaged in a continuous recognition memory task for visually presented words in two separate sessions. The magnetic flux normal to the scalp surface was measured with a whole-head neuromagnetometer during task performance. Using the total number of acceptable activity sources as an index, overall activation was greater in the left compared with the right hemisphere for all 16 subjects in both sessions. Sources of activity were consistently found in the temporoparietal areas of the left hemisphere in all subjects. Moreover, clusters of activity sources in this region either overlapped spatially or were found in close proximity across sessions. Medial and basal temporal lobe activity was also observed in most subjects during at least one session, and tended to be lateralized to left hemisphere. These results suggest that magnetoencephalography is a promising tool for determination of cerebral dominance for language and localization of temporal lobe language areas.

Identification of reliable spike templates in multi-unit extracellular recordings using fuzzy clustering.

A method for extracting single-unit spike trains from extracellular recordings containing the activity of several simultaneously active cells is presented. The technique is particularly effective when spikes overlap temporally. It is capable of identifying the exact number of neurons contributing to a recording and of creating reliable spike templates. The procedure is based on fuzzy clustering and its performance is controlled by minimizing a cluster-validity index which optimizes the compactness and separation of the identified clusters. Application examples with synthetic spike trains generated from real spikes and segments of background noise show the advantage of the fuzzy method over conventional template-creation approaches in a wide range of signal-to-noise ratios.

Temporal course of regional brain activation associated with phonological decoding.

Spatiotemporal maps of brain activity were obtained from nine normal volunteers engaged in a pseudoword rhyme-matching task using Magnetic Source Imaging. Following bilateral occipital activity there was a clear temporal progression of activity in the left hemisphere beginning with basal temporal areas followed closely by activity in the angular gyrus and inferior frontal lobe. Medial temporal and posterior superior temporal lobe became active last. There was a strong interhemispheric asymmetry in activity favoring the left hemisphere in all areas. A control, pattern-matching task, with no verbal demands, exhibited a different spatiotemporal pattern of activity. These results agree anatomically with previous studies in other imaging modalities. In addition, the relative timing of activity in basal temporal areas of the left hemisphere supports hypotheses regarding a role for this area in early processing of visually presented verbal material.

Language dominance determined by magnetic source imaging: a comparison with the Wada procedure.

OBJECTIVE: To evaluate the validity of data derived from magnetic source imaging (MSI) regarding cerebral dominance for language in patients with intractable seizure disorder. METHOD: The authors performed functional imaging of the receptive language cortex using a whole-head neuromagnetometer in 26 consecutive epilepsy patients who also underwent the intracarotid amobarbital (Wada) procedure. During MSI recordings, patients engaged in a word recognition task. This task was shown previously to activate language areas in normal adults as well as in patients who undergo intraoperative language mapping, allowing confirmation of MSI findings. Language laterality indices were formed for both the Wada and the MSI procedures. In addition, clinical judgments regarding cerebral dominance for language were made using the two methods by independent raters. RESULTS: Cluster analysis indicated excellent agreement between the quantitative MSI and Wada indices. Rater judgments showed almost complete agreement as well. CONCLUSION: MSI is a promising method for determining cerebral dominance for language.

Magnetoencephalographic mapping of the language-specific cortex.

OBJECT: In this paper the authors introduce a novel use of magnetoencephalography (MEG) for noninvasive mapping of language-specific cortex in individual patients and in healthy volunteers. METHODS: The authors describe a series of six experiments in which normative MEG data were collected and the reliability, validity, and topographical accuracy of the data were assessed in patients who had also undergone the Wada procedure or language mapping through intraoperative cortical stimulation. CONCLUSIONS: Findings include: 1) receptive language-specific areas can be reliably activated by simple language tasks and this activation can be readily recorded in short MEG sessions; 2) MEG-derived maps of each individual are reliable because they remain stable over time and are independent of whether auditory or visual stimuli are used to activate the brain; and 3) these maps are also valid because they concur with results of the Wada procedure in assessing hemispheric dominance for language and with the results of cortical stimulation in identifying the precise topography of receptive language regions within the dominant hemisphere. Although the MEG mapping technique should be further refined, it has been shown to be efficacious by correctly identifying the language-dominant hemisphere and specific language-related regions within this hemisphere. Further development of the technique may render it a valuable adjunct for routine presurgical planning in many patients who harbor tumors or have epilepsy.

Atypical temporal lobe language representation: MEG and intraoperative stimulation mapping correlation.

Functional brain imaging techniques hold many promises as the methods of choice for identifying areas involved in the execution of language functions. The success of any of these techniques in fulfilling this goal depends upon their ability to produce maps of activated areas that overlap with those obtained through standard invasive procedures such as electrocortical stimulation. This need is particularly acute in cases where active areas are found outside of traditionally defined language areas. In the present report we present two patients who underwent mapping of receptive language areas preoperatively through magnetoencephalography (MEG) and intraoperatively through electrocortical stimulation. Language areas identified by both methods were located in temporoparietal regions as well as in less traditional regions (anterior portion of the superior temporal gyrus and basal temporal cortex). Importantly there was a perfect overlap between the two sets of maps. This clearly demonstrates the validity of MEG-derived maps for identifying cortical areas critically involved in receptive language functions.

Lateralization of cerebral activation in auditory verbal and non-verbal memory tasks using magnetoencephalography.

The magnetic flux normal to the scalp surface was measured with a whole-head neuromagnetometer while right-handed subjects (N = 15) were engaged in either an auditory word- or a tone-recognition task. Sources of the recorded magnetic fields were modeled as equivalent current dipoles at 4 ms intervals and the number of sources in the later portion of the magnetic response was used as an index of the degree of brain activation. Significantly more sources were found in the left as compared to the right hemisphere in the word but not the tone task on a group basis. On an individual basis, 13/15 subjects had more sources in the left as compared to the right hemisphere during the word task, while in the tone task 3/10 subjects showed this pattern. Sources of activity were found in the left superior and middle temporal gyri in all subjects with available MRI scans. Sources were also found in the supramarginal gyrus and in medial temporal areas, including the hippocampus, in the majority of cases. MEG appears to be a promising tool for detecting activity in cerebral areas specialized for language and memory function.

MEG correlates of categorical perception of a voice onset time continuum in humans.

Event-related magnetic fields (ERFs) were recorded from the left hemisphere in nine normal volunteers in response to four consonant-vowel (CV) syllables varying in voice-onset time (VOT). CVs with VOT values of 0 and +20 ms were perceived as /ga/ and those with VOT values of +40 and +60 ms as /ka/. Results showed: (1) a displacement of the N1m peak equivalent current dipole toward more medial locations; and (2) an abrupt reduction in peak magnetic flux strength, as VOT values increased from +20 to +40 ms. No systematic differences were noted between the 0 and +20 ms stimuli or between the +40 and +60 ms CVs. The findings are in agreement with the results of multiunit invasive recordings in non-human primates regarding the spatial and temporal pattern of neuronal population responses in the human auditory cortex which could serve as neural cues for the perception of voicing contrasts.

Functional hemispheric asymmetry assessment in a visual language task using MEG.

We used magnetoencephalography (MEG) to assess the degree of hemispheric activation in eleven normal, right-handed subjects with no history of neurological disorder or learning disability during performance of a word- and a face-recognition tasks. Neuromagnetic activity was recorded using a whole-head system, and the sources of the recorded magnetic fields were modeled as single equivalent current dipoles. Early (<200 msec) cerebral activation, defined by the number of dipoles identified by the data-fitting algorithm, was localized in the occipital cortex during both tasks, as expected. During the language task, the extent of the later (>200 msec) cerebral activation was approximately double in the left hemisphere in almost all subjects, involving temporal and temporoparietal areas. In contrast, during the face-recognition task, the corresponding activation was mostly symmetrical across the two occipital lobes, also involving the posterior-inferior aspect of the right temporal lobe. Our results suggest that the MEG is a suitable method of assessing noninvasively hemispheric specialization for language.

Assessment of functional cerebral laterality for language using magnetoencephalography.

The purpose of the study was to examine the feasibility of using magnetoencephalography (MEG), a noninvasive functional brain imaging technique, to assess cerebral laterality for language. The magnetic flux normal to the scalp surface was measured with a whole-head neuromagnetometer while subjects (n = 16) were engaged in a word-matching and a tone-matching task. The effect of hemisphere and task on the number of satisfactory equivalent current dipole (ECD) solutions obtained during the late portion of the responses to the word and tone stimuli was examined. An interhemispheric ECD laterality index was also computed. Satisfactory ECD solutions were localized in perisylvian cortices during both tasks. A greater number of ECDs was found in the left hemisphere in 14 (87%) of 16 of the subjects in the word-matching task, a proportion that approaches the reported incidence of left-hemisphere dominance among right-handers. A similar proportion of subjects also showed a clear asymmetry in the number of ECDs favoring the left hemisphere in the language task as compared to the nonlanguage task. These findings suggest that MEG is a promising tool for laterality assessment. Magnetoencephalography-based functional asymmetry data are currently being compared against invasive presurgical procedures (i.e., intracarotid amobarbital procedure).

MEG correlates of categorical-like temporal cue perception in humans.

The purpose of this study was to extend previous findings of changes in magnetic flux amplitude associated with neuronal currents in the human auditory cortex in response to two-tone stimuli selected from a tone-onset time (TOT) continuum. The results replicated previous findings by indicating a close correspondence between N1m field strength functions and individual perceptual identification curves. In a second experiment event-related fields were recorded in response to two-tone stimuli in which the two sinusoid components always started simultaneously, yet they showed the same energy envelope as the TOT stimuli (amplitude modulated (AM) tokens). The results showed that N1m modulation as a function of TOT could not be accounted for by an effect of AM. The findings suggest that discontinuities in N1m field amplitude functions were the result of interactions between two populations of active neurons each displaying distinct average sound frequency preferences.

Magnetic fields elicited by a tone onset time continuum in humans.

Event-related Fields (ERFs) were recorded in eleven normal volunteers in response to four two-tone stimuli selected from a tone onset time (TOT) continuum. The latter has been used extensively in the past as an analog of voice onset-time, an important cue for making voicing distinctions (e.g., ¿pa¿ versus ¿ba¿). TOT values ranged from 0 to 60 ms in 20-ms steps. Three components of the ERF (P50m, N1m, and P2m) in the hemisphere contralateral to the stimulated ear were analyzed. The most significant finding was an abrupt reduction in the peak field amplitude of N1m (bilaterally) as TOT values increased from +20 to +40 ms. No systematic differences were noted between the 0 and +20 ms stimuli or between the +40 and +60 ms tokens. On the other hand, peak component latencies rose steadily with increasing TOT values. The discontinuity in the change of peak RMS as a function of TOT is in close agreement with behavioral evidence of categorical perception of similar non-speech as well as speech stimuli.

Multiple bilaterally asymmetric cortical sources account for the auditory N1m component.

The hypothesis that the N1, the major negative component of the cortical evoked response to auditory stimuli, originates from the primary auditory cortex has been supported by several studies. In a previous study we showed that, when monaural stimulation with pure tones is used, the distribution of the N1 peak over the scalp could be accounted for by successive activation of adjacent sources on the floor of the Sylvian fissure. In an attempt to establish the generality of the phenomenon, in this study we investigated further the generation of the N1 component using a variety of auditory stimuli, including pure tones, complex sounds (musical notes), and words, as well as binaural stimulus presentation. Additionally, we used a new recording system which allows recording of the distribution of the magnetic flux over the entire head simultaneously, thus eliminating the need for multiple recording sessions and the related problems of habituation and of changes in attention level. We found that a series of single dipolar sources could account for the entire duration of the N1m component. The location of the sources fell within the primary auditory cortex and, during the evolution of the component, they followed a posterior-anterior, medial-lateral, superior-inferior trajectory, bilaterally, along the superior surface of the temporal lobes. Additionally, the distribution of N1 sources on the two hemispheres showed a marked asymmetry, with the right hemisphere sources covering a larger area. The established consistency of successive source excitation across subjects, studies, types of stimuli, and recording systems, as well as the newly demonstrated hemispheric asymmetry of source extent, suggest the presence of a reliable phenomenon indicative of the functional organization of the auditory cortex.

Identification of language-specific brain activity using magnetoencephalography.

The purpose of the present investigation was to explore the ability of magnetoencephalography (MEG) to identify brain areas involved in language comprehension. Event-related magnetic fields (ERFs) were recorded from 7 right-handed adults with no history of neurological disorder or learning disability as they engaged in an auditory and a visual word-recognition task. A face-recognition task served as control. During the later portion of the ERFs, activity sources from both language tasks tended to overlap in temporal and temporo-parietal cortices. There was a clear preponderance of such sources in the left compared to the right hemisphere in all participants. These findings demonstrate that MEG is a promising tool for identifying brain regions involved in the analysis of linguistic stimuli, in addition to the initial encoding of stimulus features.

Relative timing of neuronal activity in distinct temporal lobe areas during a recognition memory task for words.

Nine neurologically normal volunteers engaged in a recognition memory task for visually presented words while the magnetic flux normal to the scalp surface was measured with a whole-head neuromagnetometer. The relative timing of activity occurring during the task in different brain areas was determined. Occipital lobe areas became active first, followed by basal temporal areas. Temporoparietal and medial temporal areas became active next, during later portions of the recorded epoch, and appeared to activate in concert. Except for the occipital lobe, activity was predominantly in the left hemisphere. This pattern was consistent across participants. Results suggest that MEG, in capturing temporal as well as spatial aspects of brain activity, has a promising role as an adjunct to other functional imaging modalities.