Abstracts of the Clinical EEG & Neuroscience Journal, Volume 32
JULY 2001, Volume 32
Real Time Human Brain Function: Observations and Inferences from Single Trial Analysis of Magnetoencephalographic Signals
Andreas A. Ioannides
This paper brings together results obtained by applying Magnetic Field Tomography (MFT) to the analysis of Magnetoencephalography (MEG) data over the last decade. It emphasizes the most recent developments where the availability of helmet-like MEG probes with well over 100 sensing coils provides a full coverage of the head. The paper shows that it is possible to extract tomographic information from single trial, millisecond-by-millisecond MEG signal, and demonstrates two ways that this capability can be exploited.
First, the single trial reconstructions are used to obtain robust statistical measures of changes of activity over small latency windows. Second, the interaction between areas is studied by computing the mutual information between short, time-lagged sections of the single trial time-courses. The usefulness of the computationally demanding approach is demonstrated by analyzing experiments using two widely used protocols, one for face and affect recognition and the other for Contingent Magnetic Variation (CMV).
The results show foci of significant changes of activity, which are consistent with what is reported in the literature and provide a deeper understanding of their significance. Some new, but not all that unexpected, findings also emerge from the analysis.
Human Sleep EEG Analysis Using the Correlation Dimension
Toshio Kobayashi, Shigeki Madokoro, Yuji Wada, Kiwamu Misaki and Hiroki Nakagawa
Sleep electroencephalograms (EEG) were analyzed by non-linear analysis. Polysomnography (PSG) of nine healthy male subjects was analyzed and the correlation dimension (D2) was calculated. The D2 characterizes the dynamics of the sleep EEG, estimates the degrees of freedom, and describes the complexity of the signal. The mean D2 decreased from the awake stage to stages 1, 2, 3 and 4 and increased during rapid eye movement (REM) sleep. The D2 during each REM sleep stage were high and those during each slow wave sleep stage were low, respectively, for each sleep cycle. The mean D2 of the sleep EEG in the second half of the night was significantly higher than those in the first half of the night. Significant changes were also observed during sleep stage 2, but were not seen during REM sleep and sleep stages 3 and 4. The D2 may be a useful method in the analysis of the entire sleep EEG.
Ultrafast Frequencies: A New Challenge for Electroencephalography with Remarks on Ultraslow Frequencies
E. Niedermeyer and David L. Sherman
EEG frequencies are not limited to the usual 0.5-70/sec (or 0.3-100/sec) range. In recent years, ultrafast activities between 100 and 1000/sec have been the topic of various studies with regard to physiological and paroxysmal conditions.
Personal work on ultrafast frequencies in deep structures (elicited with pentylenetetrazol in rats) is mentioned in passing and will be the object of a special study. Other work focusing on the sensorimotor cortex and thalamocortical connections has proved to be seminal for ultrafast EEG research in conjunction with evoked responses (N20 response, SSEP) and experimental neurophysiological studies of afferent volleys, including those causing paroxysmal cortical responses. The well-known decremental seizures with initially flat tracings require clarification with ultrafast recordings. In the physiological-neurocognitive domain, Pfurtscheller’s event-related desynchronization might also benefit from the use of ultrafast recording.
A plea for additional ultraslow recording (DC recording) is also being made, since paroxysmal flattening (electrodecrement) may be associated with an ultraslow negative baseline deflection. The combination of ultrafast (facilitated by digital technique) and ultraslow (technically difficult in patients, easier in experimental animals) would finally denote the frequency-wise complete EEG.
Decomposition of Event-Related Brain Potentials into Multiple Functional Components Using Wavelet Transform
Tamer Demiralp and Ahmet Ademoglu
Event related brain potential (ERP) waveforms consist of several components extending in time, frequency and topographical space. Therefore, an efficient processing of data which involves the time, frequency and space features of the signal, may facilitate understanding the plausible connections among the functions, the anatomical structures and neurophysiological mechanisms of the brain. Wavelet transform (WT) is a powerful signal processing tool for extracting the ERP components occurring at different time and frequency spots. A technical explanation of WT in ERP processing and its four distinct applications are presented here.
The first two applications aim to identify and localize the functional oddball ERP components in terms of certain wavelet coefficients in delta, theta and alpha bands in a topographical recording. The third application performs a similar characterization that involves a three stimulus paradigm. The fourth application is a single sweep ERP processing to detect the P300 in single trials. The last case is an extension of ERP component identification by combining the WT with a source localization technique.
The aim is to localize the time-frequency components in three dimensional brain structure instead of the scalp surface. The time-frequency analysis using WT helps isolate and describe sequential and/or overlapping functional processes during ERP generation, and provides a possibility for studying these cognitive processes and following their dynamics in single trials during an experimental session.
Wavelet Transform of the EEG Reveals Differences in Low and High Gamma Responses to Elementary Visual Stimuli
Ivan Bodis-Wollner, Jiang Davis, Areti Tzelepi and Tassos Bezerianos
Multiunit electrophysiological studies indicate that oscillatory activity is common in the awake mammalian central nervous system. Synchronous 20-80 Hz oscillations, so called gamma rhythms, have been proposed as a possible fundamental physiological mechanism of binding neuronal activity underlying visual object recognition. The purpose of this study was to determine whether or not gamma band oscillatory activity in the human brain is modulated by attributes of elementary visual stimulation. The experiment was performed on 7 normal subjects. Sinusoidal gratings were presented over a range of spatial frequencies. Evoked potentials were recorded over 5 surface electrodes placed in a horizontal occipital chain across the back of the head. Discrete wavelet transform was performed on the first 200 msec following stimulus onset on the average data of 256 sweeps. Power was analyzed with ANOVA across conditions.
In our previous studies we have separated a “low” (14-28 Hz) and “high” (28-55 Hz) gamma band* The current results indicate that both gamma bands to full-field stimulation have the highest power at the midline (inion) electrode to a spatial frequency of 5.5 cpd, which is the peak spatial frequency from foveal psychophysical data. However, the spatial frequency bandwidth is considerably narrower in the HG than in the LG band. Occipital spatial frequency tuning of the massed high gamma response is narrower than the tuning of individual cortical neurons. The bandwidth difference between low and high gamma band suggests that different frequency gamma range oscillations may represent not only different functional properties of visual processing, but may also reflect underlying differences in excitatory and postsynaptic inhibitory circuits shaping the contrast sensitivity of the human observer. Our study emphasizes the importance of elementary visual filter properties for gamma responses and the need to subdivide gamma frequency ranges according to functional properties.
Stimulus- and Frequency-Specific Oscillatory Mass Responses to Visual Stimulation in Man
Fabrizio De Carli, Livio Narici, Paola Canovaro, Simone Carozzo, Emanuela Agazzi and Walter G. Sannita
Oscillatory mass responses centered at about 20-35 Hz or 100-120 Hz occur (after contrast or luminance visual stimulation, respectively) in the retina and cortex of animals and man and are recorded by electrical or magnetic methods. These oscillatory events reflect stimulus-related uni/multicellular oscillations of the firing rate/membrane potential and result from synchronization of neuronal assemblies selectively responding to the stimulus characteristics. Methodological problems in the study of these events derive from the contiguity in frequency between the ERG or VEP and the oscillatory responses and from the need to reliably define oscillatory events in time and frequency. Two methods (time-frequency analysis by matching pursuit and locking index) have been implemented to approach this issue. Theory and application are reviewed.
The Processing of Stereoscopic Information in Human Visual Cortex: Psychophysical and Electrophysiological Evidence
Three-dimensional depth perception relies in part on the binocular fusion of horizontally disparate stimuli presented to the left and right eye. The mammalian visual system offers a unique possibility to study electrophysiologically cortical neuronal mechanisms: since the input of the two eyes remains separated up to the level of the visual cortex, evoked potential components that are generated exclusively by cortical structures may be explored when dynamic random-dot stereograms (dRDS) are presented.
In a series of independent studies, we determined the scalp topography of dRDS evoked brain activity in different groups of healthy subjects, and we found consistent results. Major differences between stereoscopic and contrast evoked brain activity are seen in the strength of the potential fields as well as in their topography. Our findings suggest that there are fewer neurons in the human visual cortex that are responsive to horizontal disparity, and that higher visual areas like V2 are more engaged with stereoscopic processing than the primary visual cortex. On the other hand, component latencies of evoked brain activity show no effect signifying that the binocular information flow to the visual cortex has a similar time course for both the processing of contrast information and of dRDS stimuli.
We could also verify that healthy subjects can learn to perceive 3D structure contained in dRDS. Changes in perceptual ability as measured with psychophysical tests are paralleled by systematic alterations in the topography of stereoscopically evoked potential fields. Stereoscopic VEP recordings may also be of clinical use: in patients with selectively disturbed depth perception but normal visual acuity there is a high correlation between clinical symptoms, perceptual deficiency, and altered VEP amplitudes and latencies.
Auditory Evoked Responses to Similar Words with Phonemic Difference: Comparison Between Children with Good and Poor Reading Scores
Ignacio Valencia, Gloria B. McAnulty, Deborah P. Waber and Frank H. Duffy
Our previous study demonstrated a physiologic deficit in two-tone discrimination in poor readers.1 This was specific to the left parietal area suggesting that poor readers handled rapid tones differently. The current paper extends this finding in the same population, demonstrating that poor readers also have difficulty with phonemic discrimination.
Long latency auditory evoked potentials (AEP) were formed using a phonemic discrimination task in a group of children with reading disabilities and controls. Measuring peak-to-peak amplitude of the waveforms, we found reduced N1-P2 amplitude in the Poor Reader group.
Using the t-statistic significance probability map (SPM) technique, we also found a group difference, maximal over the mid-parietal area, from 584 msec to 626 msec after the stimulus onset. This difference was due to a lower amplitude on the Poor Reader group. We hypothesized that this late difference constitutes a P3 response and that the Poor Reader group generated smaller P3 waves.
These auditory evoked response (AER) data support a discrimination deficit for close phonemes in the Poor Reader group as they had smaller N1-P2 absolute amplitude and developed smaller P3 waves.
Based on these data we should be able to differentiate between Good and Poor readers based on long latency potentials created from phonemic stimuli.
Auditory Evoked Response Data Reduction by PCA: Development of Variables Sensitive to Reading Disability
Frank H. Duffy, Ignacio Valencia, Gloria B. McAnulty and Deborah P. Waber
Long latency auditory evoked responses (AER) were formed on 232 healthy normal and learning impaired subjects to tone pairs of 50 msec inter-stimulus interval (TALAER) and also to the words “tight” and “tyke” (TTAER). Both evoked potential (EP) types have been used to demonstrate differences between good readers (WIAT Basic Reading score > 115, N=42) and poor readers (Reading score < 85, N=42). A largely automated, hands off approach was used to reduce artifact contamination, to develop canonical measures for discriminating good from poor readers, and to predict reading scores across the entire population including intermediate (average) readers. Eye and muscle artifact were diminished by multiple regression. Substantial EP data reduction was enabled by an unrestricted use of Principal Components Analysis (PCA). For each EP type, 40 factors encompassed 70-80% of initial variance, a meaningful data reduction of about 90:1. Factor interpretation was enhanced by mapping of the factor loadings. By discriminant analysis, resulting factors predicted reading group membership with over 80% jackknifed and also split –half replication accuracy. By multiple regression, they produced a canonical variate correlating significantly (p<0.001) with the Basic Reading score (r=0.39). The TTAER factors were more useful than the TALAER factors. The relevance of rapid auditory processing and phonemic discrimination measurements to dyslexia is discussed.
The case of a 66-year-old patient is reported in view of the rarity of his condition: a case of subacute encephalopathy with seizures in alcoholics (SESA syndrome), described first in 1981 by Niedermeyer, et al.1 Wernicke-type aphasia, epileptic seizures (generalized tonic-clonic) and PLEDs EEG pattern dominated the neurological picture, in addition to hepatomegaly and rhabdomyolysis. This condition differs from all other known CNS complications in chronic alcoholism and is withdrawal-independent. It is prognostically favorable as far as the syndrome as such is concerned.
Report of a case of SESA syndrome: a rare CNS complication of chronic alcoholism, known since 19811 and characterized by epileptic seizures, multiple and reversible neurological deficits, as well as PLEDs in the EEG. The MRI showed enhanced occipital signals in the T2-weighted sequence, which resolved together with the clinical findings.
Somatosensory evoked potentials (SSEP) to electrical stimulation of the median nerve by using cephalic and noncephalic references were studied to detect the generator sources of short latency evoked potentials in 29 patients with cerebral, brainstem, spinal and peripheral nerve lesions.
Patients were divided into six groups according to the localization of their lesions:
Group 1: cortical and subcortical lesions, group 2: basal ganglion lesions, group 3: pons and mesencephalon lesions, group 4: diffuse cerebral lesions, group 5: cervical cord lesions, group 6: brachial plexus lesions. Potentials were recorded using cephalic and noncephalic references after median nerve stimulation.
Evidence obtained from patients suggested the following origins for these short latency SSEPs: P9 may arise in brachial plexus, P11 in dorsal basal ganglions or dorsal column, P13 and P14 in the nucleus cuneatus and lemniscal pathways, N16 in subthalamic structures and most likely mid and lower pons, N18 from the thalamus and thalamocortical tract, and N20 from primary somatosensory cortex.
The interactions between two different nerves occur by occlusion or inhibition when two nerves share the synaptic connections. In our previous study, we have demonstrated that posterior tibial nerve and peroneal nerve sensory inputs interact with each other, i.e., preceding stimulus to one nerve suppresses the somatosensory evoked potential (SEP) of the other nerve when two stimuli are delivered in close sequence. The course of suppression follows two phases; the first one occurring at short interstimulus intervals (ISIs) of the two nerves less than 10 msec, and the second one being at around 30 msec ISI after partial recovery following the first suppression phase. In that study, we have postulated that the second phase suppression was equivalent for the movement induced “gating” mechanism.
In this study, the interactions of mixed nerve (posterior tibial) and sensory nerve (sural), and also sensory (sural) and sensory (saphenous) nerves were examined. We found that the mixed nerve (posterior tibial) exerted similar dual phases of suppression (as was seen in posterior tibial – peroneal nerve study) on to the sural nerve SEP, but the reverse was not true. Also the sensory and sensory nerve interactions were not mutually equal; the sural nerve stimulation caused two phases suppression but the reverse condition did not show significant suppression. The above findings suggest (1) interference input from the sensory nerve to the mixed nerve is much weaker than the reverse condition, and (2) sensory and sensory nerves interactions occur but two nerves’ interference inputs are not necessarily equal and one could dominant the other.
The acute effect of traumatizing events on the human brain has long been studied. The major obstacles to this endeavor have been the severity and the delay from time of exposure to the traumatizing events. To avoid these issues, EEG and clinical examinations of 18 healthy, drug-free subjects were performed 2 weeks after an earthquake occurred in Turkey in 1999. It was found, for the first time to our knowledge, that EEG parameters can predict startle response in acute stress reaction correctly in 100% of the cases. EEG measures of the reactivity to eyes opening were especially important in this regard.