* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download Neurofeedback Treatment of Epilepsy
Survey
Document related concepts
Cognitive neuroscience wikipedia , lookup
Neuroplasticity wikipedia , lookup
Time perception wikipedia , lookup
Persistent vegetative state wikipedia , lookup
Neuropsychopharmacology wikipedia , lookup
Neural oscillation wikipedia , lookup
Biology of depression wikipedia , lookup
Neural correlates of consciousness wikipedia , lookup
Brain–computer interface wikipedia , lookup
Dual consciousness wikipedia , lookup
Magnetoencephalography wikipedia , lookup
Electroencephalography wikipedia , lookup
Metastability in the brain wikipedia , lookup
Transcript
Neurofeedback Treatment of Epilepsy by Jonathan E. Walker, MD, Neurotherapy Center of Dallas Clinical Professor of Neurology, Southwestern Medical School (Corresponding Author) and Gerald Kozlowski, Ph.D., Neurotherapy Center of Dallas, Professor of Physiology, Southwestern Medical School Mailing Address (for both authors): Neurotherapy Center of Dallas 12870 Hillcrest Rd., Suite 201 Dallas, TX 75230 Phone: 972-991-1153 Fax: 972-991-1346 Email: [email protected] Web site: neurotherapydallas.com Neither author has a relationship with a commercial company that has a direct financial interest in the subject matter or materials discussed in this article or with a company having a competing product. 1 SUMMARY By using neurofeedback, it is possible to train the brain to de-emphasize those rhythms which lead to generation and propagation of seizure, and to emphasize those rhythms that make seizures less likely to occur. With recent improvements in quantitative EEG measurement and improved neurofeedback protocols, it is often possible now to eliminate seizures or reduce the amount of medication required to control them. The use of neurofeedback to treat human epilepsy dates back to the early 1970’s.1 Subsequent studies (without QEEG guidance) showed that seizure frequency could be reduced in the majority of patients who were trained to decrease slow frequencies in the 3-8 Hz range and reward frequencies in the 9-18 Hz range. However, complete abolition of seizures was rare. With the development of sophisticated QEEG date bases it has become possible to more precisely characterize power and coherence abnormalities associated with drugresistant epilepsy. If there is focal excessive power in a frequency band, it may be downtrained. If there is a focal deficiency in power, it may be uptrained. Similarly, significantly decreased coherence between brain areas may be uptrained, and significantly increased coherences may be downtrained. This approach has been found to decrease or abolish seizures in all patients thus trained. Many even become drug-free. In this chapter, we discuss the history of neurofeedback for epilepsy, followed by discussions of the relevant neurophysiology of epilepsy. A model of how neurofeedback might raise the seizure threshold in then presented. We then summarize our results between 2002-2003, using our approach. Finally, we discuss potentially profitable future directions for making this training even more effective. 2 Neurofeedback is noninvasive, safe, and relatively inexpensive in comparison to surgical therapy of drug-resistant epilepsy. Commonly, drugs may be reduced or eliminated, and side effects consistently reduced. Improvements in comorbidities, such as ADD, learning disabilities, depression, anxiety, and memory loss also occur. It should be an integral part of the treatment of these patients. We have also been able to use neurofeedback in patients who have intolerable side effects from drugs to protect them from seizures. We have also been able to stop antiepileptic therapy of epileptic women who wish to become pregnant, as it poses no risk to the fetus. 3 Neurofeedback for Epilepsy I. A Brief History The value of neurofeedback for human epilepsy was first established by Stevman and his colleagues in 1974.1 12-16 Hz activity (“SMR”) at 20-25 μV was rewarded over central and frontal areas in 4 medically refractory patients, 3 sessions/week for 6-18 months. All 4 patients exhibited a reduction in seizure frequency during “SMR” training over 0.5-1.5 years. A rebound in seizure frequency occurred after cessation of training in 3 patients. Finley et al2 found that reduction of “SMR” was associated with reduced seizure frequency and normalization of the EEG in a severely epileptic patient. In an important paper, Lubar and colleagues3 carried out a double-blind study of neurofeedback for uncontrolled epilepsy. They found that reward to inhibit 3-8 Hz activity, along with reward to increase 11-19 or 12-15 Hz activity was associated with a decrease in seizure frequency. If increase in 3-8 Hz activity was rewarded there was an increase in seizure activity. Wyler et al4 rewarded desynchronizing action (increase in fast activity, decrease in slow activity) which was effective in reducing seizure frequency in several medically intractable patients. Ayers5 reported successful treatment of absence epilepsy by training patients to inhibit 4-7 Hz activity and reward 15-18 Hz activity, using bipolar training at T3/C3 and T4/C4. Ten patients became and remained seizure-free for 10 years. In another study, non-contingent training had no effect on seizures, but seizure reduction was observed with 9-14 Hz or 14-26 Hz training (Kuhlman6). Later, Lantz and Sterman7 showed that contingent training was superior to non-contingent training in 24 patients, and the effect persisted at least 6 months. Reward of 6-9 Hz activity increased seizure activity, while reward of 12-15 Hz activity decreased seizure activity (using the same 4 patients, with an A-B-A design).8 Sterman reported that he trained at CZ because that was where the 12-15 Hz activity was best seen. Sterman reviewed all of the literature up to 2000, and found that every single study of neurofeedback for epilepsy reported positive results.1 In his meta-analysis, 82% of patients demonstrated > 30% reduction in seizures, with an average greater than 50% reduction. Approximately 5% remained seizure-free for up to 1 year. There has been no systematic study to determine which precise frequencies should be trained to maximize the reduction in seizures. Several other studies, not reviewed by Sterman in 2000, also support the effectiveness of neurofeedback in reducing seizure frequency in poorly controlled epilepsy.9-12 In recent years, in our clinic, we have tried a different approach to try to normalize the EEG of epileptics by using QEEG to guide our neurofeedback training. The general approach is to determine the most significant abnormalities and train those areas to normal. Power training alone resulted in significant seizure reductions, but rarely resulted in cessation of seizures. Based on our successful use of coherence training in closed head injury, 13 we began using protocols to normalize coherence in medically refractory epileptics. The combination of power and coherence training to normalize the EEG has proven dramatically effective in several cases.14 Recent advances in QEEG, notably reliable databases which include single-hz bins15,16 and broadly based coherence determinations, have made more precise determinations of what to treat possible. Several studies have looked at QEEG abnormalities in epileptic patients.17,18,19 Traditional broad-band spectral analyses generally find focal increases in relative power of theta, with decreased values in the alpha and delta bands when compared with normative data. Diaz et al 20 found a positive correlation between the magnitude of the 5 quantitative abnormalities and the amount of paroxysmal activity. There was a slowing of mean frequency of alpha in the epileptic group. They found that focal epilepsy may be associated with wide-spread changes in the frequency spectra. We have also seen widespread abnormalities in coherence and phase, often remote from the spike focus. We could find no published studies of coherence abnormalities in epileptic patients. Another important issue relates to the transition from interictal to ictal EEG.21 We think an important goal of EEG biofeedback is to prevent that transition, if the transitional discharge leading to seizures can be inhibited by the training. This proved to be the key to controlling episodic status epilepticus in one of our patients, whose seizures invariably began with rhythmic theta activity in one or the other temporal lobe.14 Long term monitoring may be required to determine those rhythms responsible for the transition from interictal to ictal activity. III. Possible Mechanisms of Reducing Seizures in Epilepsy A. Neurophysiology of Epilepsy The human brain must be both stable and flexible to operate efficiently.22 By changing the amounts of excitation and stimulation in simulated neural networks, one can induce synchronous bursting behavior of large neural networks (the hallmark of epilepsy).23 In an animal model of epilepsy (low-magnesium), Nyikos et al24 have shown that recurrent seizure-like events in brain slices are initially characterized by high amplitude electrical triplets (“paroxysmal spikes”4) with a single rhythm that starts at 200 Hz and continuously declines to below 1 Hz at termination of the seizure-like event. High frequency bursts have also been described in a high-potassium in vitro model.25 These rapid oscillations (100-300 Hz) have also been observed in vitro.26 They are not 6 usually detected in conventional EEG recordings, because of muscle artifact and remoteness from surface electrodes. Medvedeu27 has pointed out that intense gamma activity (40-60 Hz) often precedes epileptic discharges in patients and in some animal models. He analyzed power, coherence, and phase in the Kainic acid model in the rat hippocampus and neocortex. At onset of discharges, highly coherent, intense gamma rhythms were followed by a slow rhythm of epileptiform spikes and sharp waves. During the spike activity and immediately afterwards, the gamma power and coherence were significantly decreased. He hypothesizes that epileptiform spike activity may result from extreme activation of the “anti-binding” mechanism controlling temporal binding at high frequencies, i.e., the “epileptiform4” discharge develops as a protective mechanism suppressing fast activity. Fisher et al28 found an increase in 40 Hz activity just before epileptiform discharges in patients with complex partial seizures. Medvedev suggests that post-ictal depression of the EEG may represent a selective decrease in coherence at high frequencies, rather than a non-specific suppression of all frequencies. It may be that fast activity is desynchronized and suppressed by spike activity. Previously, Engel29 had suggested that interictal spikes may inhibit seizure activity in Kindling models of epilepsy. Nuwer17 noted a decrease in beta frequencies over well-defined epileptic spike foci. Another way to view the pathophysiology of epilepsy is to consider seizures as an example of a disturbance in excitatory and inhibitory input to the seizure focus.30,31 The simplest models are consistent with the generation of seizures when there is a disruption of the normal balance of inhibitory/excitatory input, such that there is an excess of excitatory input (e.g., glutamergic) connections or a deficiency in inhibitory connections 7 (e.g., GABA ergic). Drugs that increase glutamergic activity typically produce seizures, while drugs that enhance GABA-ergic activity typically are anti-epileptic. It is not clear that human epilepsy is produced in this way, however. Slices of brain taken from seizure foci in children undergoing surgical treatment for epilepsy, appeared to be qualitatively normal in terms of synaptic transmission and local neurological circuits.32 Often, ictal recordings show that seizure onset is remote from the spike focus.33 Stevens has suggested that the interictal spikes must have survival value to have persisted through evolutionary history.34 Spikes (vertex waves) are normal when confined to certain axial nuclei of the brain during sleep (K-complexes) or during photic or sexual excitation and may be compensatory after deafferentation or hypoxia. In epilepsy surgery, it is usually assumed that the structural and biochemical pathology is confined to the area in and around the spike “focus”, based on the fact that seizures often stop when the focus is removed. However, there is evidence from animal models and humans that, in many cases, the behavioral abnormalities responsible for the epileptic activity are remote from the focus. When a freezing lesion is made in the right posterior cortex of the rat, the spike focus is likely to be in the anterior cortex on the same of the other side. Slowing is seen over the site of the biochemical lesion. Application of tetanus toxin to one hemisphere is associated with seizure development in the contralateral hemisphere.35 Lieb et al 36 looked at propagation of ictal discharge in patients with complex partial seizures of mesial temporal origin, using depth electrodes in mesial temporal, lateral temporal and frontal lobes. During seizure onset there was a marked increase in intrahemispheric coherence, indicating that areas remote from the focus participate in the initiation of focal seizures. Steriade et al 37 have reviewed a body of evidence which 8 supports the importance of thalamocortical synchronization in the genesis of epileptic rhythms. Bartolomei et al 38 implanted temporal lobe depth electrode in patients with drug resistant epilepsy and describe four different seizure types. One type begins in the medial cortex with phasic discharges spreading to neocortex and high coherence between medial temporal cortex and neocortex. A second group had seizures starting in medial cortex and spreading to lateral cortex with a fast low-voltage discharge (FLVD) which spread rapidly to neocortex and high coherence between medial and lateral structures. A third group exhibits FLVD starting in lateral neocortex and spreading rapidly to hippocampus and amygdala, with high coherence between these structures. These studies suggest that a technique, such as neurofeedback, which can normalize coherence, could likely reduce propagation of seizures from the site of onset. Interhemispheric connections probably play a less important role then intrahemispheric connections, even when the seizures spread to the opposite hemisphere.39 Chavez et al 40 have recently described decreased synchrony at 10-25 Hz in the seizure onset zone 30 minutes prior to seizures. It may be that transient synchrony changes are being prevented by neurofeedback training of coherence (see below). III. Thoughts on how neurofeedback might lower seizure threshold in epileptic patients. There have been few efforts at developing a model of how neurofeedback works. Lubar has proposed an excellent model for how neurofeedback enhances attentional capabilities.41 This is easily adapted for explaining how neurofeedback might work for epilepsy. (Fig. 1) A more detailed model, which could be adapted to include neurotransmitter influences, may be found in the Hughes and John paper.42 For our purposes, it is sufficient to say that out goal is to produce optimal coherence and phase 9 synchronization, to raise seizure threshold and reduce the likelihood of having a seizure. Lubar has pointed out that this training produces very long-lasting and perhaps permanent, learning. Thus, one literally learns not to have seizures. Obviously, one would not want to diffusely decrease gamma or delta/theta activity, but rather train the epileptic foci or connections relevant to the seizure, since delta, theta and gamma are important for normal sleep and learning. IV. Results of QEEG – guided power and coherence training at our center. Our approach is to “train away” abnormalities of power and coherence in an attempt to thereby decrease and hopefully abolish seizure likelihood. We generally treat the most statistically abnormal power or coherence abnormality first, then the next most severe abnormality, then the next most severe, with 5-10 sessions of each protocol. Details of our methodology may be found in our paper on closed head injury. Table II indicates details for each patient, including the spike focus (if one was found), training protocols, and degree of success. These results appear to be superior to previously published results in neurofeedback for partial complex seizures, in that all patients became seizure free and many were able to stop their anticonvulsant treatment. A detailed abstract on one patient was presented as an abstract at the ISNR conference in 2003. 14 10 Table I. Epilepsy Patients Class I – Seizure-free off medications x 3 mos. Class II – Seizure-free on medications x 3 mos. Class III – Occasional seizure (< 1/mo.) x 3 mos. Class IV – > 1 sz/mo. (failure) (none seen) Patient #1 – M Age 62 #2 – F 43 #3 – F 19 #4 – M 31 #5 – F #6 – F 5 Pre-NF Type seizure Sz freq PCS, 2° gen 3/wk (F8) Simple partial (P3) PCS (brief) 5/day (average) PCS (no 3x/week focus) PCS, status epilepticus (no focus) 6/day 4x/month PCS (No focus) 1/day PCS – Multifocal spikes 1/week On meds Dilantin, 300 mg/d Neurontin, 3600 mg/d Dilantin, 300 mg/d Lamictal, 50 mg bid Tapetol, 400mg bid Protocols 1) Downtrain 1-8 Hz F4 2) Uptrain 27-30 Hz F4 3) Downtrain coh α T3/F3 # sessions 5 5 10 Class II Class I Uptrain coh Θ F3/O1 Downtrain coh Θ FP1/F3 Downtrain coh β C4/P4 Downtrain coh α FP1/F3 Uptrain coh β O1/O2 5 5 5 5 10 Uptrain coh α O1/O2 Uptrain coh Θ T5/T6 Downtrain coh Θ/↑ SMR C4 Downtrain coh Θ/↑ 15-18 Hz Cz Downtrain coh Θ/ ↑ 1518 Hz C3 Downtrain coh SMR C4/P2 Tegretol, Uptrain coh Θ O1/O2 400 mg Uptrain coh Θ F7/F8 bid Downtrain coh Θ/ ↑ 15Neurotin, 18 Hz C3 300 mg tid Downtrain coh 22-30 Hz C3 Downtrain coh 4-7 Hz FZ 9 10 5 10 10 10 Lamictal 25mg tid Result Downtrain 4-8/↑15-18 F8 Downtrain 4-8/↑15-18 P4 Uptrain coh ∆ P3/O1 Downtrain coh Θ FP1/T3 Class II Class II 5 5 5 10 2 Class III No seizures but “drop attack” 1 x/ week 10 6 9 3 T=28 11 Class II Patient #7 – F #8 – M #9 – M Age Type seizure 6 PCS – Right occipital spike focus 19 PCS with 2° generalization sharp waves – O1 FP1 Biparietal 35 #10 – M 9 Pre-NF Sz freq 5/day 5/month PCS with 2° gen No spike foci 1/week PCS with 2° gen No spike foci 1/week On meds 5/day Primidone 125mg Lamictal 350mg/d Topamax 200mg/d Tegretol, 300mg tid Protocols Downtrain 27/↑ 15-18 Hz O1 Downtrain 27/↑ 15-18 Hz O2 Downtrain 1-7 Hz O2 Uptrain coh ∆ F3/C4 Uptrain coh Θ F7/CZ Uptrain coh α FP1/F3 Downtrain coh β C4/T6 Downtrain 2-7 Hz/↑ 1518 Hz T5 Downtrain 1-6 Hz/↑ 1518 Hz FP1 Uptrain coh ∆ F7/FP2 Uptrain coh β F3/T5 Downtrain 1-4 Hz FP1 Uptrain coh ∆ C3/C7 Uptrain coh α F8/O2 Downtrain coh Θ F8/O2 Uptrain coh α F3/FZ Downtrain coh Θ FP1 Downtrain coh Θ T6 Uptrain α coh F3/O1 Uptrain α coh F4/O2 Uptrain α coh T3/T5 Uptrain β coh T3/T5 Uptrain SMR CZ Depakote, Downtrain Θ coh F1/F3 500mg bid Downtrain Θ coh F2/F4 Downtrain α coh F1/F3 Downtrain Θ power F8 # sessions 10 7 T=17 Result Class I 7 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 T=82 10 10 7 10 4 T=41 8 5 5 5 10 T=25 PCS = partial complex seizures 2° gen = secondary generalization 12 Class II Class II Class II IV. Possible Future Directions Recent work suggests that feedback of slow D.C. potentials could be quite effective in reducing seizure frequency. Future magnetoencephalography machines will be able to record simultaneously from both hemispheres to better localize deep epileptogenic zones and determine where training is most likely to be effective.43 LORETA neurofeedback may be used to train some deep foci.44 EEG-linked fMRI has been used to image single interictal discharges and could be used to image sites of seizure onset. Then LORETA training might be used to train at the focus rather than at the surface. Modern spectral analysis techniques allow high resolution in the highter beta range (> 18 Hz).45 There have been few studies of higher beta in epilepsy or neurofeedback, and it appears that higher beta may be very important in both fields.46 The availability of implantable subdural electrode arrays has made systematic studies of electrocorticographic coherence possible.47 These could enable more precise coherence training, which might spare the need for brain resection. Non-linear dynamic approaches to studying epilepsy have recently come to the fore (e.g., chaos theory).48 Newer feedback systems that use non-linear approaches have been developed.49 These may prove very valuable in remediating seizures. Finally, down training of gamma activity over the spike focus may be an effective way of inhibiting seizure activity. This is possible with some of the newer biofeedback systems.50 We have tried this recently in 3 drug-resistant patients, with early success in reducing seizures. We do not know if the effects will be long-lasting. 13 FIGURE 1 Thalamic Activity Low frequency burst mode Single action potential mode (specific thalamocortical) Projection to neocortical Layer 1 EEG DELTA THETA Projections to neocortical Layer 4 ALPHA BETA GAMMA Hyper-coupled States Hyper-coupled States Optimal Coupling Seizure Likelihood Greater than Normal Normal Greater Than Normal (“Safe zone” ≈ 10 -18 Hz) Behavioral States Hypnosis Normal Increase attention Light Sleep consciousness Learning Visualization Propagation of seizures Onset of seizures (focal) 14 CASE REPORT: A 31-year-old man had recurrent partial complex and secondarily generalized seizures and frequent episodes of status epilepticus. Treatment with nine different anticonvulsants singly, and in combination, did not control his seizures or prevent status episodes. He was not a surgical candidate because of depth electrode monitoring revealed independent foci of onset in the left and right temporal lobes. A vagal nerve stimulator was implanted and did not reduce his seizures or prevent status episodes. METHOD A QEET (John database) revealed an increase in absolute theta at F7 and T4 and increased relative theta at FP1, FP2, F7 and F8. Decreased coherence of theta was found at T3/T5, O1/O2 and C3/C4. Decreased beta coherence was found at O1/O2. Training was done three times per week. First, theta coherence was down trained at T3/T5 (10 sessions), then O1/O2 theta coherence was down trained (5 sessions). Next, beta coherence was down trained at O1/O2. Then theta coherence was down trained at C3/C4. Finally, theta was down trained and SMR (12-15 Hz) was up trained at T3. RESULTS Since completion of the first protocol, there have been no further episodes of status epilepticus. Generalized seizures decreased to one per week after the first protocol and none have occurred since the second protocol was done. Partial complex seizures decreased to 10 per week after the first protocol, to 5 per week after the second and 1 per week after the third. When an increase in seizure frequency was noted with the 15 fourth protocol, it was discontinued and the fifth protocol was done. Training was completed December 1, 1998. The patient has had no more seizures. He was able to return to work full time and to begin driving again. Phenytoin was discontinued in October of 1998. His speech is no longer slurred and his memory is improved. He chose to continue his other medication, not wanting to risk a seizure while driving. CONCLUSION This case illustrates the powerful potential of QEEG-guided neurofeedback in the management of severe epilepsy. The training is noninvasive and was effective in a relatively short time. The cost is quite low compared to epilepsy surgery or vagal nerve stimulation. It may prove to be particularly important for drug resistant epilepsy. The combination of coherence training with traditional power training may work better than either approach alone. 16 1 Sterman, M B, Basic concepts and clinical findings in the treatment of seizure disorders with EEG operant conditioning. Clinical Electroencephalography, 1990; 32(1):45-55.* 2 Finley, W W, Smith H A, Etherton MD, Andrews, D J, Schonfeld, W H, Ellertson, B, & Klove, H., et al., Reduction of seizures and normalization of the EEG in a severe epileptic following sensorimotor biofeedback training; preliminary study, Biol. Psychology 2:189, 1975. 3 Lubar, J F, Shabsin, H S, Natelson, S E, Holder, G S, Whitsett, S F, Pamplin, W E, et al, EEG operant conditioning in intractable epileptics, Arch. Neurol, 1981;38(11):700-704. 4 Wyler, A. R., Lockard, J. S., & Ward, A. A., et al, Conditional EEG desynchronization and Seizure occurrence in patients, Electroencephalography and Clinical Neurophysiology 41:501, 1976. 5 Ayers, M., Long-term follow up of EEG Neurofeedback with absence seizures, Abstract, SSNR meeting, 1995. 6 Kuhlman, W. N., EEG feedback training of epileptic patients: clinical and electroencephalography analysis, electroencephalography Clin. Neurophysiology 45:69, 1978 7 Lantz, D. and Sterman, M.B., Neuropsychological assessment of subjects with uncontrolled epilepsy: Effects of EEG feedback training. Epilepsia 1988;29(2):163-171.* 8 Sterman, M. B. and Macdonald, L. R., Effects of central cortical EEG feedback training on seizure incidence in poorly controlled epileptics. Epilepsia 1978;19(3):207-222. 9 Kotchoubey, B., Busch S., Strehl, U., & Birbaumer, N., Changes in EEG power spectra during biofeedback of slow cortical potentials in epilepsy, Applied Psychophysiology and Biofeedback 1999;24(4):213-233.* 10 Kotchoubey B, Strehl U, Uhlmann C, Holzapfel S, Konig M, Froscher W, et al., Modification of slow cortical potentials in patients with refractory epilepsy: A controlled outcome study. Epilepsia 2001;42(3):406-416.* 11 Rockstroh B, Elbert T, Birbaumer N, Wolf P, Duchting-Roth A, Reker M, et al. Cortical self-regulation in patients with epilepsies. Epilepsy Research 1993;14:63-72.* 17 12 Uhlmann C, & Froscher W, Biofeedback treatment in patients with refractory epilepsy: Changes in depression and control orientation. Seizure 2001;10(1):34-38.* 13 Walker, J E, Weber, R and Norman, C, Importance of QEEG guided coherence training for patients with mild closed head injury, J. Neurotherapy 6:31-42, 2002. 14 Walker, J E and Davidson, D, Long term Remediation of Seizures in Refractory Epilepsy with QEEG- Guided Neurofeedback training, Abstract SNR, 2003. 15 Thatcher, R W, Neuroguide 16 Hudspeth, R W, Hudspeth database 17 Nuwer, M.R., Frequency analysis and topographic mapping of EEG and evolved potentials in epilepsy, Electroenceph. Clin. Neurophysiol. 69:118, 1988. 18 Hanner, R.N., et al, Completed EEG topography in epilepsy, Neurology 143: 457, 1987. 19 Baldy-Moulinier, M. et al, Report des cartographies de l’EEG quantrifiee et du déloit sanguine cerebral dans les epilepsies partielles complexes, in P. Rondot et al (eds.) Cartographie EEG. Frisson-Roches David, 1987, pp. 129-137. 20 Diaz, G.F. et al, Generalized background qEEG abnormalities in localized systematic epilepsy, Electroenceph. Clin. Neurohysiol. 106: 501, 1998. 21 Walczak, T S, & Jayahar, P, Interictal EEG, in Engle, J & Pedley, T. 22 Liljenstrom, H., Neural Stability and Flexibility: A Computational approach Neuropsychophaumacology 28: 564-573, 2003. 23 Kudela, P., Changing excitation and inhibition in simulated neural networks: effects on induced bursting behavior, Biol. Cybernetics 88:276, 2003. 24 Nyikos L, Lasztoczi B, Antal K, Kovacs R, Kardos J. Desynchronisation of spontaneously recurrent experimental seizures proceeds with a single rhythm. Neuroscience. 2003;121(3):705-17. 25 Bruce J. Gluckman, Hanh Nguyen, Steven L. Weinstein, and Steven J. Schiff. Adaptive Electric Field Control of Epileptic Seizures Journal of Neuroscience, 21 (2), 590-600 (2001) 26 Bragin, A, Engel Jr,, J Wilson, C., Fried, I., Buzsáki, G. Hippocampus 1999; 9(2):137-142. 27 Medvedeu, A.V., Epileptiform spikes desynchronize and diminish fast (gamma) activity of the brain: An anti-binding mechanism, Brain Research Bulletin 58:115, 2002. 18 28 Fisher, R. S., Webber, WR, Lesser, RP, Arroyo, S. & Uematsu, S., High-frequency EEG activity at the start of seizures, J. Clin. Neurophysical 9:441, 1992. 29 Engel, J, et al, Epileptic activation of antagonistic systems may explain prardoxical features of experiemtnal and human epilepsy : a review, In Wada, J. A. (ed), Kindling Raven Press, 1981. 30 Dichter, M., and Wilcox, K., “Excitatory Synaptic Transmission”, In: Epilepsy, Eugel, J., and Pedley, T. (eds.), Lippincott-Raven, Philadelphia, 1998, pp. 257-265. 31 McDonald, R., Inhibitory Synaptic Transmission, ibid, pp. 265-277. *Ibid is no longer used, need reference 32 Dulek, F., Wuarin, J., Tasker, K., Yang, I., ad Peacock, W., Neurophysiology of cortical slices resected from children undergoing surgical treatment for epilepsy, J. Neurosci. Methods. 59: 49-58, 1995. 33 Nosworthy, E., The role of clinical neurophysiology in the management of epilepsy, J. Clin. Neurophysiol. 15: 96-108, 1998. 34 Stevens, J, Signal of information transmission?, Ann Neurol. 1: 309-314. 35 Finnorty, G., and Jefferys, J., Investigation of neuronal aggregate generating seizures in the rat tetanus toxin model of epilepsy, J. Neurophysiol. 88: 2919-2927, 2002. 36 Lieb, J., Hogue, K., Skomer, C., and Song, X., Inter-Hemispheric propagation of human mesial temporal lobe seizures: a coherence/phase analysis, Electroencephalography and Clin. Neurophysiol. 67: 101-119, 1987. 37 Steriade, M., Gloor, P., Llinas, R., Lopes da Silva, F., & Mesulam, M., Basic mechanisms of cerebral rhythms activities, Electroencephalopathy and Clin. Neurophysical, 67:101-119, 1987. 38 Bartolomei F, Wendling F, Vignal JP, Kochen S, Bellanger JJ, Badier JM, Le Bouquin-Jeannes R, Chauvel P., Seizures of temporal lobe epilepsy: identification of subtypes by coherence analysis using stereo-electro-encephalography, Clin Neurophysiol, 110:1741-1754, 1999. 39 Gotman J., Interhemispheric interactions in seizures of focal onset: data from human intracranial recordings, Electroenceph. Clin. Neurophysical, 67:120-133, 1987. 40 Chavez, M., Quyen, M., Navarro, V., Baulac, M., & Martinerie, J., Spatio temporal dynamics prior to neocortical seizures: amplitude vs. phase couplings”, IEEE Trans., 50:571-583, 2003. 19 41 Lubar J. F., Neocortical dynamics: implications for understanding the role of neurofeedback and related techniques for the enhancement of attention, Applied Physiol. Biofeedback, 22:111-126, 1997. 42 Hughes J. R., & John, E. R., Conventional and quantitative electroencephalography in psychiatry, J Neuropsychiatry Clin Neurosci. 11:190-208, 1999. 43 Masuoka L.K., Spencer S.S., Clinical neurophysiology in epilepsy, Curr Opin Neurol. 7:148, 1994. 44 R.D. Pascual-Marqui, M. Esslen, K. Kochi, D. Lehmann. Functional imaging with low resolution brain electromagnetic tomography (LORETA): a review. Methods & Findings in Experimental & Clinical Pharmacology 2002; 24C:91-95. 45 Schomer D. L., Bonmassar G., Lazeyras F., Seeck M., Blum A., Anami K., Schwartz D., et al, EEG- Linked functional magnetic resonance imaging in epilepsy and cognitive neurophysiology, J Clin Neurophysiol, 17:43-58, 2000. 46 Dumermuth G., Gasser T., Germann P., Hecker A., Herdan M., Lange B., Studies on EEG activities in the beta band, Eur Neurol. 16:197-202, 1977. 47 Towle VL, Carder RK, Khorasani L, Lindberg D., Electrocorticographic coherence patterns, J Clin Neurophysiol, 16:528-547, 1999. 48 Lopes da Silva, F.H., Blanes, W., Kalitzin, S. N., Parra, J., Suffczynski, P., Velis, D. N., Dynamical diseases of brain systems: different routes to epileptic seizures, IEEE Trans 50:540-547, 2003. 49 Lehnertz, K and Elger, C E , Can Epileptic Seizures be Predicted? Evidence from Nonlinear Time Series Analysis of Brain Electrical Activity. Physical Review Letters 1998;80(22):5019-5022.* 50 Collura, T. (2002) Application of Repetitive Visual Stimulation to EEG Neurofeedback Protocols, Journal of Neurotherapy Vol. 6(2) 47-70. 20 21