Cognitive tasks augment gamma EEG power

Share Embed

Descrição do Produto

Clinical Neurophysiology 115 (2004) 1802–1809

Cognitive tasks augment gamma EEG power S.P. Fitzgibbona,b, K.J. Popec, L. Mackenziea, C.R. Clarkb, J.O. Willoughbya,* a

Centre for Neuroscience and Department of Medicine (Neurology), Flinders University, P.O. Box 2100 Adelaide, SA, Australia b Cognitive Neuroscience Laboratory, School of Psychology, Flinders University, P.O. Box 2100 Adelaide, SA, Australia c School of Informatics & Engineering, Flinders University, P.O. Box 2100 Adelaide, SA, Australia Accepted 3 March 2004

Abstract Objective: Gamma EEG oscillations are low amplitude rhythms in the 30 – 100 Hz range that correlate with cognitive task execution. They are usually reported using time-locked averaging of EEG during repetitive tasks. We tested the hypothesis that continuous gamma EEG would be measurable during mental tasks. Methods: We investigated sustained human gamma EEG oscillations induced by 8 cognitive tasks (Visual Checkerboard, Expectancy, Reading, Subtraction, Music, Expectancy, Word learning, Word recall, and a Video Segment) in 20 subjects using standard digital EEG recording and power spectral analysis. Results: All of the cognitive tasks augmented gamma power relative to a control condition (eyes open watching a blank computer screen). This enhancement was statistically significant at more than one scalp site for all tasks except checkerboard. The Expectancy, Learning, Reading and Subtraction tasks expressed the most impressive gamma response, up to 5 fold above the control condition and there was some task-related specificity of the distribution of increased gamma power, especially in posterior cortex with visual tasks. Conclusions: Widespread gamma activation of cortical EEG can easily be demonstrated during mental activity. Significance: These results establish the feasibility of measuring high frequency EEG rhythms with trans-cranial recordings, demonstrate that sustained gamma EEG activity correlates with mentation, and provides evidence consistent with the temporal binding model. q 2004 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved. Keywords: Spectral analysis; Theta; Alpha; Beta

1. Introduction Gamma EEG oscillations (low amplitude rhythms in the 30 –100 Hz range) became a topic of intense interest in humans (Aoki et al., 1999; Joliot et al., 1994; Keil et al., 1999; Revonsuo et al., 1997; Sauve et al., 1998; TallonBaudry et al., 1998) after it was established in animal models that synchronous recurrent discharge bursts within the gamma range are involved in perception and cognition and are correlated with cognitive task execution (Engel and Singer, 2001). In several animal studies, gamma EEG derived from cortex using small electrodes has been closely correlated with both local multi-unit and single-unit discharges. Synchronous gamma discharges have been identified in the visual, auditory, somatosensory, olfactory, motor and * Corresponding author. E-mail address: [email protected] (J.O. Willoughby).

memory modalities in a wide range of animal species (Engel and Singer, 2001). Within these areas, there were synchronous bursts within groups of neurons in different cortical columns (Gray et al., 1990), spatially distributed within the same hemisphere (Engel et al., 1991a; Frien et al., 1994), between inter-hemispheric sites (Engel et al., 1991b) and in sub-cortical structures (Alonso et al., 1996). Other animal experiments have demonstrated the functional significance of synchronous gamma oscillations by showing that synchroneity of gamma discharges correlates with cognitive function (Fries et al., 1997; Murthy and Fetz, 1996; Roelfsema et al., 1997) and that disrupted synchroneity of gamma discharges correlates with diminished cognitive function (Roelfsema et al., 1994; Stopfer et al., 1997). In humans, trans-cranial EEG integrates measures summed electrical fields which are volume conducted from a large population of neurons within an extensive region around a recording site. Such measurement is unlikely to detect EEG rhythms derived from small cell

1388-2457/$30.00 q 2004 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.clinph.2004.03.009

S.P. Fitzgibbon et al. / Clinical Neurophysiology 115 (2004) 1802–1809

assemblies and it would be even less likely to permit identification of synchronous rhythms from small cell assemblies in different cortical areas. However, the presence of gamma oscillations in trans-cranial EEG indicates significant synchronicity in large populations of subjacent neurons and, therefore, that there are relatively large local cell assemblies exhibiting rhythmic bursting, possibly synchronous with cell groups elsewhere. Synchronous gamma activity between widely distributed cell groups, sometimes referred to as ‘binding’ gamma, is thought to integrate (bind) information processed in distributed neurons and/or neural circuits and/or cortical areas into a coherent cognitive process/percept (Engel and Singer, 2001). Thus synchronous cortical gamma EEG provides a measure of binding activity and has been observed for short durations, for example in auditory discrimination (Joliot et al., 1994), somatosensory discrimination (Sauve et al., 1998), stereoscopic fusion (Revonsuo et al., 1997), the formation of percepts (Keil et al., 1999), working memory (Tallon-Baudry et al., 1998), and sensory-motor processing (Aoki et al., 1999). In many studies such as these, the involvement of gamma has been demonstrated using repetitive tasks, timelocked averaging, and short post-stimulus time windows. We determined if sustained gamma EEG oscillations, induced by a variety of complex mental tasks in human subjects, could be measured trans-cranially and if the strength and spatial pattern of enhancement would be task-dependant.

2. Method 2.1. Subjects Twenty adults (8 male and 12 female) free of psychiatric or neurological disorder participated in the study, approved by the Flinders Clinical Research Ethics Committee, and all subjects gave informed and written consent. They were recruited as age, gender and education-matched controls for patients as part of a larger clinical study (published elsewhere (Willoughby et al., 2003)). 2.2. EEG Sixty-four channel EEG was recorded continuously (linked-ear reference, 512 samples per second, 16-bit analog to digital conversion, 107 Hz low-pass filter) using a commercial EEG acquisition system (Compumedics, Victoria, Australia). A 64-channel electrode cap with tin electrodes provided uniform scalp coverage. Electrode impedances were kept below 5 kV. 2.3. Cognitive tasks EEG was recorded whilst participants performed the following 8 tasks, chosen to activate mental activity in a variety of circumstances, and a Control procedure:


2.3.1. Visual checkerboard The participant was instructed to fixate for 20 s on a red dot located in the centre of an alternating, rectangular blackand-white checkerboard pattern for 20 s (check size 1.5 £ 2 cm, alternation rate 8 Hz). 2.3.2. Story reading The participants were instructed to read silently for a period of 28 s from the beginning of page 46 of ’Politically Correct Bed-Time Stories’ by James Finn Garner (Souvenir Press, London, 1994). The book contained only text in size 12 point font. 2.3.3. Subtraction task The participants were instructed to serially subtract 7 from 1000 and, in a practice session, the participants were intermittently interrupted to check their accuracy and to ensure compliance with the task. The recorded Subtraction period was not interrupted. A small number of subjects who were unable to serially subtract 7 from 1000 were given a simpler serial Subtraction task. 2.3.4. Music The participants listened to a segment of Pachelbel’s Cannon in D for 28 s. 2.3.5. Expectancy The participant was presented with a series of 11 expectancy trials, each involving stimulus pairs. Each pair consisted of a visual direction cue presented at eye height in the centre of the computer screen followed 4 s later by a lateralised visual target. Each cue consisted of an arrow pointing equally probably to the left or right. The target was a cross, located on the side indicated by the preceding arrow. The period between each direction cue and the subsequent target was 4 s. Participants were required to respond by a button press as soon as a target was presented. The expectancy period recorded was the interval between each directional cue and the subsequent target. 2.3.6. Learning The participants were instructed to memorise a set of 10 words that were presented simultaneously on a computer monitor for 20 s. The words were medium frequency (Kucera and Francis, 1967) concrete nouns of 4 – 7 letters in length. In order to induce intentional learning, participants were advised that they would be tested on them subsequently. 2.3.7. Recall Three lists of 10 words were presented sequentially for 10 s each, with an inter-list interval of 2 s. The first list contained the same words as the Learning list but in different order. The other two lists contained some of the words in the first list. Following the presentation of all 3 lists the participant was asked to identify which list contained


S.P. Fitzgibbon et al. / Clinical Neurophysiology 115 (2004) 1802–1809

the words that they had previously learnt. EEG was recorded during the presentation of the first list. 2.3.8. Strictly ballroom The participant was required to watch a 28 s segment of the film ‘Strictly Ballroom’ (with permission of Andrew Pike, Ronin Films, Canberra, Australia). The segment selected contained a complex mix of intense colour, movement, speech, drama and music. Subjects also completed a Control task requiring them to relax with eyes open and to look at a blank computer screen. During the experiment, participants were seated 1.5 m in front of a computer monitor on which the various task materials were presented. Participants were instructed on the requirements of each task immediately beforehand. Because the subject group were controls for a study of patients with different forms of epilepsy and the numbers of patients were not known, it was not feasible in advance to undertake a Latin square design in administering the tasks. In addition, there were two tasks that required a fixed order, namely Learn and Recall. The tasks were therefore administered in fixed order: Checkerboard, Reading, Subtraction, Music, Expectancy, Learn, Recall, Strictly Ballroom, Control and there was a half to 1 min rest between each task. 2.4. EEG analysis All uncontaminated EEG for each task (usually around 20 s) was epoched into consecutive 1 s blocks. Each epoch was transformed from the temporal domain to the frequency domain using fast-Fourier transform (1 Hz resolution, 512 point block-size, Hanning window, 1 – 100 Hz). While artifact due to eye blinks was present, eye blinks frequency and expression in the power spectra could not be detected because of their brief duration and infrequent occurrence (around once every 10 s) even during the Strictly Ballroom task. Muscle activity, if present, occurred in isolated regions and did not prevent the use of recordings from other leads. When muscle activity affected electrodes in a widespread distribution, all leads were edited from the analysis. The frequency epochs were averaged within each task for each subject to yield an averaged power spectrum and then divided by the power spectrum for the Control task for that subject. This provided a relative spectral response profile for each subject for each task. The data was banded into theta (3 –7 Hz), alpha (8 –12 Hz), beta (13 –29 Hz) and gamma (30 –100 Hz) frequency bands. For visualisation purposes the relative spectra were averaged across subjects within each task to provide group mean spectra relative to Control for each task. To ensure data were not contaminated by 50 Hz mains frequency and 60 Hz computer monitor refresh rate, power values in the 50 ^ 1, 60 ^ 1, 99 and 100 Hz values were omitted from analysis.

2.5. Statistical analyses The Kolmogorov – Smirnov test was used to assess normality of power estimates, which was achieved when the power estimates were log-transformed. Significant differences in spectral power between experimental tasks and the Control task were calculated for each power band using a single factorial analysis of variance for each electrode. Significances were corrected form multiple comparisons using the Modified Bonferonni procedure.

3. Results Relative to the Control condition there were widespread increases in EEG power during all tasks except Checkerboard. We illustrate in Fig. 1 the average relative powerresponse to Learning at each electrode as a montage of power spectra with 1 Hz resolution. 3.1. Gamma power All of the tasks other than Checkerboard exhibited significantly increased gamma band power ðP , 0:05Þ relative to the Control condition (Fig. 2). These increases were evident in all cases at more than one scalp site and their spatial distribution was task-specific. The most striking enhancement of gamma power was expressed in the Expectancy, Learning, Reading, and Subtraction tasks. All 4 of these tasks induced widespread 2 – 5 fold increases in gamma power relative to Control at a large number of posterior and central scalp sites (Fig. 2). The Recall, Music and Strictly Ballroom tasks also induced increases in gamma power, but of lower order and at fewer electrodes. The spatial distribution of enhanced gamma power for Recall was similar to that for Expectancy and Learning. 3.2. Theta, alpha and beta power There were no significant increases in alpha band power relative to the Control condition for any of the experimental tasks. The distribution of beta band power for each experimental task relative to the Control condition was similar to the gamma power distribution, however, the relative enhancement was weaker than for gamma and tended to occur at frequencies close to the gamma frequency band. Increased beta power therefore very likely reflects an increase in power of an equivalent phenomenon to gamma activity. The only significant increases in apparent beta power were observed during the Reading and Subtraction tasks ðP , 0:05Þ and these were localised to a small subset of the scalp sites that were significant for the gamma band. There were 2 –5 fold increases in theta power at bilateral and midline frontal scalp sites for all tasks except

S.P. Fitzgibbon et al. / Clinical Neurophysiology 115 (2004) 1802–1809


Fig. 1. Montage display of individual (a) and mean (with SEM) of 20 subjects (b) EEG power increases during Reading relative to the Control state (ordinate scale: from 1 to 5 fold) between 0 and 100 Hz (abscissa) recorded over the scalp. Decreases relative to the Control state (values below 1) are not shown.


S.P. Fitzgibbon et al. / Clinical Neurophysiology 115 (2004) 1802–1809

Fig. 2. Topographic maps of group mean gamma power for each Experimental task relative to the Control condition. The maps were scaled from 1 (black) to 5 fold (white) increases in gamma power relative to the Control condition. Each greyscale increment represents a 0.125 fold increase. A relative power of 1 indicates no difference to the Control condition. The transparent red overlay marks scalp sites at which the increase in gamma power relative to Control was significant (P , 0.05).

Checkerboard. The Subtraction task produced the most striking augmentation of theta power with a large number of central, temporal and frontal sites reaching significance relative to the Control condition (Fig. 3). In addition, there were 1 – 3 fold increases in theta power in the Read, Recall, Strictly Ballroom and Subtraction tasks at bilateral and midline occipital sites.

4. Discussion The key finding is that mental activity can easily be demonstrated to augment gamma activity using transcranial recordings. We recorded ongoing EEG during continued mental activity without time-locked averaging of repeated tasks, revealing increased gamma during all

Fig. 3. Topographic maps of mean theta power for each Experimental task relative to the Control condition. The maps were scaled from 1 (black) to 5 fold (white) increases in theta power relative to the Control condition. Each greyscale increment represents a 0.125 fold increase. A relative power of 1 indicates no difference to the Control condition. The transparent red overlay indicates scalp sites at which the increase in theta power relative to Control was significant ðP , 0:05Þ:

S.P. Fitzgibbon et al. / Clinical Neurophysiology 115 (2004) 1802–1809

tasks other than Checkerboard. Increases were widespread but quite variable from individual to individual, as indicated by the maximal mean increases not always reaching statistical significance. Increases were most striking for tasks of Expectancy, Learning, Reading, and Subtraction in which significant, 2 –5 fold increases in gamma power were obtained at widespread posterior and central scalp sites. In addition, the distributions of augmented gamma EEG demonstrated some specificity for individual tasks. The tasks constitute very complex stimuli but we used a ‘blunt’ approach to inducing gamma activity that many would have hypothesised we would not find. Clearly, given our findings, the methodology is open to be used in almost unlimited kinds of studies that will address the detail of ‘how and where gamma correlates with thinking’, both in health and in disease. We have interpreted the findings generally and with an eye to avoiding over-interpretation. The Expectancy task was similar to tasks used to elicit the contingent negative variation (CNV). The CNV is a steady slow negative shift in the EEG observed in the period prior to the presentation of an expected stimulus. Therefore the CNV is thought to reflect a state of anticipation by the brain for the expected stimulus (Walter, 1968). It is this state of anticipation that is likely to be inducing the enhanced gamma power response observed in the Expectancy task. The Expectancy task is a visual and motor paradigm so that visual and motor areas are obvious candidates for contributing to the anticipatory state. The distribution of the gamma power response at posterior and central sites is consistent with the implied role of the visual and motor systems. The Learning task is an intentional episodic memoryencoding task. Cabeza and Nyberg (2000) conducted a largescale review of the many functional neuro-imaging studies that have investigated intentional episodic memory encoding. They concluded that the key cortical areas associated with episodic memory encoding are the prefrontal, cerebellar and medial-temporal brain regions. We did not record from medial temporal cortex nor from cerebellum. While there was some pre-frontal gamma augmentation, this was not a consistent finding and did not reach statistical significance. However the comparisons made in most encoding studies contrast a condition involving encoding with a very similar condition involving less encoding. In this study, we contrasted a visual language-encoding task with a nonlanguage visual control condition, and as such the observed responses may be language related gamma power enhancement at posterior sites related to reading as opposed to episodic encoding. This task, like others we used, is very complex and we think it would be difficult to reach consensus on what might be an appropriate control task. This experiment also emphasises an obvious limitation of surface EEG recordings: they do not provide information about brain regions remote from the scalp. The distribution of gamma power in the reading task relative to the Control condition is consistent with


neuro-imaging studies that have consistently demonstrated temporal, parietal and occipital involvement in written word recognition and comprehension (Cabeza and Nyberg, 2000). Language studies using event-related potentials (ERP) have demonstrated that early ERP components first appear at occipital sites followed closely by the expression of the subsequent components at occipital-temporal sites (Vitacco et al., 2002). We report gamma power for Word Reading relative to Control that is sustained for the duration of the task and as such we cannot delineate the timing at which various cortical regions are recruited, however our distribution of gamma power is supportive of their involvement. The Subtraction task exhibited the greatest increase in gamma power relative to the Control condition and it was significant at more scalp sites than any of the other conditions. This may be related to the inherent difficulty of the task as task complexity correlates with gamma EEG power (Simos et al., 2002). It is also the only task to exhibit extensive frontal augmentation of theta power, a correlate of the augmented attention requisite for this task. It would be useful in future work to obtain measures of difficulty from the subjects for each of the tasks. The distribution of the significant gamma power response at occipital scalp sites is curious given that subtraction was not a visual task. As the primary visual cortex has been reported to be involved in visual imagery (Cabeza and Nyberg, 2000), occipital gamma enhancement may point to the use of visual imagery during serial subtraction. Mental tasks led to augmented theta activity in central frontal leads and, in addition to frontal sites, Subtraction was associated with a marked increase in theta power in temporal sites. In this study, Subtraction was the most powerful in enhancing gamma power and, intuitively, it would be expected to be the most mentally challenging of our tasks. Frontal theta correlates with mental tasks requiring attention as originally demonstrated with arithmetic and reasoning (Ishihara and Yoshii, 1972). Recently, the probable source of this activity has been demonstrated by magneto-encephalography to be medial prefrontal cortex (Ishii et al., 1999). Theta activity is generally associated with cognition and memory (Klimesch, 1999). Further, intensified theta activity has previously been reported in humans during recall and other tasks, with some correlation with task effort (Gundel and Wilson, 1992; Grunwald et al., 2001; Schober et al., 1995). Theta activation is generated in hippocampus and related structures in response to alerting stimuli in rabbit, rat and other species (Blessing, 1997). Thus enhanced theta activity may be reflective of mental arousal, analogous to the findings in the hippocampus in animals. Its temporal prominence may also be partially supportive of a hippocampal to temporal process, consistent with the possibility of arousal-induced hippocampal theta generation in humans. There was a small, but significant, augmentation of theta power during Reading, Recall and Strictly Ballroom, an observation that is difficult to account


S.P. Fitzgibbon et al. / Clinical Neurophysiology 115 (2004) 1802–1809

for in terms of occipital cortical involvement in mental processing given the absence of this finding in the Learn task, which was a visual task. The Recall, Music and Strictly Ballroom tasks expressed significant increases in gamma power relative to Control in only a small number of sites, whilst Checkerboard expressed no significant gamma power enhancement relative to Control. These results may be related to lack of complexity in the task. With the exception of Recall, these tasks were passive, as they did not require the subject to actively engage in the task. This is in contrast to the Expectancy, Learning, Reading and Subtraction tasks all of which required active involvement by the participant and therefore, we propose, induced far more impressive gamma power responses. While the Recall task was not passive, it was still simple as the participants were presented with lists from which they had to select the list they had previously learnt. They were not required to spontaneously recall the learnt words un-cued. Although gamma enhancement with Music was left-sided, a subset of subjects undertook a similar task, viz listening to Mozart, in which the gamma augmentation was bi-temporal (unpublished). We are therefore reluctant to interpret the apparent lateralization with Pachelbel as a robust finding at this stage. There were different distributions between the areas of augmented gamma EEG determined by significance versus areas determined by maximal amplitude. Maximal mean power was sometimes markedly influenced by powerful gamma responses in a few subjects with little or no enhancement at the same area in others. This observation points to striking individual variability in the brain areas recruited into mental activity, possibly correlating with the widely differing strategies individuals use in solving mental problems. The areas most consistently activated were parietal and central, an observation that fits with the known general involvement of fronto-parietal networks during working memory processes in humans (Cabeza and Nyberg, 2000). How do the findings from this study relate to the temporal binding model of cortical processing? The model proposes that increased gamma power, as presented here, reflects large-scale integration of many coactive cell assemblies synchronously discharging in recurrent bursts at different periodicities within the gamma band. The discharges would serve the purpose of binding assemblies, both local and distant, so that the information processed in both could be integrated into a coherent whole. While we have not attempted to define different regions with synchronous (same-phase) gamma oscillations, enhanced synchroneity of neuronal bursting locally is a prerequisite for enhancement of gamma EEG activity, and we observed 2 – 5 fold increases in mean gamma power, strongly supportive of gamma involvement in cerebral processing and consistent with the temporal binding model. In the Reading task, for example, the gamma response reflects binding in

the various visual perception and language comprehension assemblies across the occipital, parietal and temporal areas to form the story. In the temporal binding model, task complexity demands utilisation of more cognitive resources. This requires more binding and subsequently results in increased gamma power. From this viewpoint, Subtraction and Reading would be the most complex of the tasks we administered, intuitively something that seems likely. Using spectral analysis to examine EEG correlates of mental processing permits measurement of oscillatory phenomena only. While evidence that oscillatory activity is an important aspect of cortical processing and reflects synchrony and binding, it is not established that oscillatory activity is essential for all mental processing nor for mediating all synchronous (bound) neuronal activity. For example, Newsome et al. (1990) demonstrate that firing rates of neurons correlate with the perception of motion (control condition 10– 45 spikes per 2 s versus activated condition 45– 90 spikes per 2 s). Konig et al. (1995) have also observed that synchronous neural discharges may be achieved over distances less than 2 mm with or without oscillating firing patterns. Our methodology does not permit measurement of any aspect of such non-rhythmic activity if it occurs during the mental tasks we used. In conclusion, all of the cognitive tasks we administered clearly increased gamma power relative to a Control condition. This enhancement was significant at more than one scalp site for all tasks with the exception of Checkerboard. The Expectancy, Learning, Reading and Subtraction tasks were the most complex tasks and expressed the most impressive gamma power response. Finally, different tasks led to different distributions of gamma EEG power increase. These results establish the easy feasibility of examining sustained high frequency EEG activity without time-locked averaging and demonstrate some of the EEG correlates of mentation. It provides evidence of the involvement of gamma EEG rhythms in these processes and demonstrates some specificity in the distribution of gamma EEG activation. The findings are consistent with the temporal binding model. This method also offers a simple means of defining the distribution of gamma over the cerebral convexity correlating with thought processes in individuals, as well as in health and in disease.

Acknowledgements Funded by National Health and Medical Research Council.

References Alonso JM, Usrey WM, Reid RC. Precisely correlated firing in cells of the lateral geniculate nucleus. Nature 1996;383:815–9.

S.P. Fitzgibbon et al. / Clinical Neurophysiology 115 (2004) 1802–1809 Aoki F, Fetz EE, Shupe L, Lettich E, Ojemann GA. Increased gamma-range activity in human sensorimotor cortex during performance of visuomotor tasks. Clin Neurophysiol 1999;110:524– 37. Blessing WW. The lower brainstem and bodily homeostasis. New York: Oxford University Press; 1997. Cabeza R, Nyberg L. Imaging cognition II: an empirical review of 275 PET and fMRI studies. J Cognit Neurosci 2000;12:1–47. Engel AK, Singer W. Temporal binding and the neural correlates of sensory awareness. Trends Cognit Sci 2001;5:16–25. Engel AK, Kreiter AK, Konig P, Singer W. Synchronization of oscillatory neuronal responses between striate and extrastriate visual cortical areas of the cat. Proc Natl Acad Sci USA 1991a;88:6048– 52. Engel AK, Konig P, Kreiter AK, Singer W. Interhemispheric synchronization of oscillatory neuronal responses in cat visual cortex. Science 1991b;252:1177–9. Frien A, Eckhorn R, Bauer R, Woelbern T, Kehr H. Stimulus-specific fast oscillations at zero phase between visual areas V1 and V2 of awake monkey. NeuroReport 1994;5:2273 –7. Fries P, Roelfsema PR, Engel AK, Konig P, Singer W. Synchronization of oscillatory responses in visual cortex correlates with perception in interocular rivalry. Proc Natl Acad Sci USA 1997;94:12699–704. Gray CM, Engel AK, Konig P, Singer W. Stimulus-dependent neuronal oscillations in cat visual cortex: receptive field properties and feature dependence. Eur J Neurosci 1990;2:607–19. Grunwald M, Weiss T, Krause W, Beyer L, Rost R, Gutberlet I, Gertz HJ. Theta power in the EEG of humans during ongoing processing in a haptic object recognition task. Brain Res Cogn Brain Res 2001;11: 33–7. Gundel A, Wilson GF. Topographical changes in the ongoing EEG related to the difficulty of mental tasks. Brain Topogr 1992;5:17–25. Ishihara T. Yoshii N Multivariate analytic study of EEG and mental activity in juvenile delinquents. Electroencephalogr Clin Neurophysiol 1972; 33:71–80. Ishii R, Shinosaki K, Ukai S, Inouye T, Ishihara T, Yoshimine T, Hirabuki N, Asada H, Kihara T, Robinson SE, Takeda M. Medial prefrontal cortex generates frontal midline theta rhythm. NeuroReport 1999;10: 675–9. Joliot M, Ribary U. Llinas R Human oscillatory brain activity near 40 Hz coexists with cognitive temporal binding. Proc Natl Acad Sci USA 1994;91:11748–51. Keil A, Muller MM, Ray WJ, Gruber T, Elbert T. Human gamma band activity and perception of a gestalt. J Neurosci 1999;19: 7152–61. Klimesch WEEG. alpha and theta oscillations reflect cognitive and memory performance: a review and analysis. Brain Res Brain Res Rev. 1999;29: 169–95.


Konig P, Engel AK. Singer W Relation between oscillatory activity and long-range synchronization in cat visual cortex. Proc Natl Acad Sci USA 1995;92:290–4. Kucera H, Francis NW. Computational analysis of present-day American English. Providence, RI: Brown University Press; 1967. Murthy VN, Fetz EE. Synchronization of neurons during local field potential oscillations in sensorimotor cortex of awake monkeys. J Neurophysiol 1996;76:3968–82. Newsome WT, Britten KH, Salzman CD, Movshon JA. Neuronal mechanisms of motion perception. Cold Spring Harb Symp Quant Biol 1990;55:697– 705. Revonsuo A, Wilenius-Emet M, Kuusela J. Lehto M The neural generation of a unified illusion in human vision. NeuroReport 1997;8:3867–70. Roelfsema PR, Konig P, Engel AK, Sireteanu R. Singer W Reduced synchronization in the visual cortex of cats with strabismic amblyopia. Eur J Neurosci 1994;6:1645–55. Roelfsema PR, Engel AK, Konig P, Singer W. Visuomotor integration is associated with zero time-lag synchronization among cortical areas. Nature 1997;385:157 –61. Sauve K, Wang G, Rolli M, Jagow R, Kronberg E, Ribary U. Llinas R Human gamma-brain activity covaries with cognitive temporal binding of somatosensory stimuli in sighted and blind subjects. Soc Neurosci Abstr 444.419 1998;24:1128. Schober F, Schellenberg R, Dimpfel W. Reflection of mental exercise in the dynamic quantitative topographical EEG. Neuropsychobiology 1995; 31:98–112. Simos PG, Papanikolaou E, Sakkalis E, Micheloyannis S. Modulation of gamma-band spectral power by cognitive task complexity. Brain Topogr 2002;14:191–6. Stopfer M, Bhagavan S, Smith BH, Laurent G. Impaired odour discrimination on desynchronization of odour-encoding neural assemblies. Nature 1997;390:70 –4. Tallon-Baudry C, Bertrand O, Peronnet F. Pernier J Induced gamma-band activity during the delay of a visual short-term memory task in humans. J Neurosci 1998;18:4244–54. Vitacco D, Brandeis D, Pascual-Marqui R. Martin E Correspondence of event-related potential tomography and functional magnetic resonance imaging during language processing. Human Brain Mapp 2002;17: 4–12. Walter WG. The contingent negative variation: an electro-cortical sign of sensori-motor reflex association in man. Progress in Brain Research 1968;22:364–77. Willoughby JO, Fitzgibbon SP, Pope KJ, Mackenzie L, Medvedev AV, Clark CR, Davey MP, Wilcox RA. Persistent abnormality detected in the non-ictal electroencephalogram in primary generalised epilepsy. J Neurol, Neurosurg Psychiatry 2003;74:51–5.

Lihat lebih banyak...


Copyright © 2017 DADOSPDF Inc.