Related Topic(s):


A chapter from the Textbook of Neurofeedback, EEG Biofeedback and Brain Self Regulation
edited by Rob Kall, Joe Kamiya and Gary Schwartz
The E-book is Available on CD Rom


M. Barry Sterman, Ph.D.

School of Medicine, University of California, Los Angeles


Even with dramatic technological advances, measuring the "internal behavior" of the brain is a far more complex task than measuring the body systems that it controls. At the present time the EEG signal recorded from the scalp is the only reliable, real-time, and totally risk-free method for estimating this internal behavior. Because the EEG reflects neuronal behavior, the method which uses this signal as a source for physiological feedback has been given the separate label that we use here, namely neurofeedback.

Studies in animals have taught us a great deal about the EEG signal and its origins within the brain. Since we cannot conduct similar studies in humans we must depend upon the insights provided by this animal research to guide our understanding of the EEG, and to provide the strategies to shape some aspects of the internal behavior of the brain. Fortunately, there are strong parallels between the EEG patterns seen in higher mammals and in humans. Moreover, evidence provided by disorders of the human nervous system and by accidental injuries suggests that the mechanisms underlying these patterns are also similar. Therefore, much of the picture that we present here relating to the human EEG was, of necessity, gained through careful and painstaking research conducted in higher animals. In my experience, the models which have emerged apply effectively to the interpretation of the human EEG.

Mapping the brain

The EEG measures small electrical potentials that can be recorded at the greatly expanded surface of the brain, which is called the cerebral cortex. The cerebral cortex is the highest level of nervous system organization. Cells here store our memories, generate our thoughts, express our feelings and intentions, and determine our conscious awareness of ourselves and of the outside world. They are organized into separate systems, some of which are concerned with the regulation of specific functions, such as sensations and movements, and others with more integrative processes, such as behavior sequencing and memory. This organization allows us to position EEG electrodes in a pattern that maps the surface of the brain and examines the status of these different functions. The method used is called quantitative EEG topographic mapping.

The small electrical potentials that make up the EEG are generated by summated extracellular currents in pools of cortical cells that are modulated by signals arriving from the outside world and from other parts of the nervous system. These changes are recorded by referencing a cortical electrode against a presumed "silent" site such as the earlobe (referential recording), or against a different pool of neurons in a different brain area (bipolar recording). For example, let’s say that we are sitting at a table reading a boring book. If nothing of interest is found in this material, we will become progressively inattentive and eventually be unable to "see" what we are reading. We will read the same sentence over and over and still not be certain about what it says. This behavior occurs because the sensory pathway and cortical cells that convey this information to consciousness have begun to assume a "resting mode". This change will be especially reflected in the EEG recorded from electrodes placed over the visual part of our cortex. However, if a spider unexpectedly crawls onto the left side of the page, the pathway that conveys this information from the eyes to the brain will suddenly excite specific cells in the area of our cortex that records and interprets that portion of our visual world. The EEG will show this change. Depending on our comfort with spiders this event may also elicit automatic responses in other pathways which may cause changes in our internal state, focus of attention, and motor behavior as we orient behaviorally to the spider and take some action. These responses will be reflected by other characteristic changes in the EEG.

By placing sensors or electrodes over appropriate areas of the cortex we can measure the pattern of such changes in ongoing EEG signals. These changes are expressed as shifts in the frequency and amplitude of the EEG signals. The term frequency refers to the timing of the average interval between successive oscillations in the EEG, as we shall see below. The term amplitude refers to the average height or "voltage" of the oscillations in a given frequency range.

Traditionally, people in the EEG field have given Greek letter labels to various frequency ranges seen in the EEG signal. This practice, which was useful at first, has come under some question due to the fact that the same frequency patterns can be seen under very different functional circumstances. For example, hypoxia, drowsiness, epilepsy, ADD and head injury can all produce 4-8 Hz patterns which are not functionally distinguished by the label theta. Similarly, brain excitation is associated with an EEG pattern referred to as beta activity. In my opinion, this label is also unfortunate, since it oversimplifies actual events.


Exciting the Pathways of the Brain

What causes the EEG to show such changes? To understand the answer to this question, we must understand the different ways in which cortical cells (or neurons) are influenced by incoming signals. Neurons are like little batteries, in that there is a separation between negative ions, which are kept in abundance inside the cell, and positive ions which are kept outside of the cell. This separation is produced by the electrochemical properties of the cell’s outer layer, or membrane. Resting neurons are thus said to be charged, or polarized, and this charge can be measured with sensitive instruments.

Receptive sites can be excited by incoming signals that release chemicals called excitatory neurotransmitters on the neuronal membrane. These chemicals act like keys that then open doors, or channels, in the cell membrane and allow positive ions into the cell and let negative ions out. This produces a small depolarizing current flow that moves across the membrane and sets up ionic perturbations throughout the cell and its surrounding extracellular space. It also reduces the charge across the membrane. This current flow is called an excitatory pre-synaptic potential, or EPSP. If many such pre-potentials accumulate they can depolarize, or discharge the cell, an event referred to as an action potential. Action potentials flow to the cells’ axon and are thus carried to other neurons with which the cell is connected, causing local release of neurotransmitters on their membrane surfaces. The cell will rapidly charge up again to be ready for the next bit of information.

Neurons can also show an increase in polarization, a phenomenon called hyperpolarization. This response is caused by the action of a different class of pre-synaptically released chemicals, called inhibitory neurotransmitters. These transmitters carry negative chloride ions into the cell and/or keep another positive ion, potassium, out of the cell. The opposite current flow produced makes the membrane more polarized and, therefore, the cell more difficult to excite. The combination of currents produced by depolarizing and hyperpolarizing dynamics and their interaction between deep and superficial layers of the cortex are responsible for generating the field potentials that give rise to the EEG.

During the processing of relevant sensory information, such as the spider mentioned above, excitation reaches the cell from many sources, including the incoming sensory information, its interpretation and integration by other cortical systems, and its affective associations. In the sensory pathway itself excitation may be selective, with only certain cells discharging. This selectivity conveys the distinguishing features of the stimulus, such as its location, shape, movements, etc. However, other nearby cells that are not directly excited may actually be hyperpolarized. This process can produce an area of cellular inhibition around the excited cells called "surround inhibition", which is believed to help focus and resolve incoming signals. Because many sources of excitation stimulate relevant cortical cells at many sites, while nearby cells are unaffected or inhibited, the complex field potentials produced tend to cancel each other when recorded from our distant surface electrode. This will produce a recorded EEG trace that shows little frequency modulation or amplitude, which we call an activated or desynchronized EEG pattern, and others call Beta. When this kind of EEG trace appears suddenly in the visual cortex we know that the individual has seen the spider. If it spreads rapidly to other cortical areas we can assume that they were startled and/or not necessarily delighted by this perception.


Reduced excitation: Making Brain Waves

When they are not processing information, cortical cells can behave quite differently. They are under a different kind of control. The pathways between external sensory and internal regulatory systems and the cerebral cortex pass through a structure deep in the brain called the thalamus. The thalamus, which in Greek means "inner room" and in Egyptian "antechamber", is made up of clusters of functionally related cell groups called nuclei. The Egyptian antechamber designation is very appropriate, since these nuclei serve to regulate the access of information to the brain from either sensory receptor systems concerned with the external world, such as the eyes and ears, or from other structures in the brain concerned with internal regulation, such as motor coordination, emotions, and other higher functions.

Thus, when the spider entered the picture, cells in the thalamic nucleus that relay left visual field information from the retina of the eyes to the visual nucleus of the thalamus were stimulated. The cells in this nucleus were also excited by internal neuromodulatory responses to this unexpected perception that arise from such factors as arousal, startle or fear. As such, they readily conveyed this information to the appropriate part of the visual cortex and initiated an alerting sequence of events. However, in the absence of the spider, the cells in both the thalamic and cortical elements of the visual pathway would tend to go off-line if the individual was becoming uninterested.

In this case, a different EEG pattern would appear in the visual cortex and eventually in many other cortical areas as well. This different pattern is caused by changes in both the thalamus and cortex when we become inattentive. During attention, stimulation from the world around us and our interest in certain things in that world combine with our general state of alertness to excite the thalamus and cortex. This makes information conduction efficient. However, during inattention these sources of excitation are all reduced. The thalamic relay cells and their cortical targets are less excited and may even be inhibited by competing brain processes. This causes these cells to become relatively hyperpolarized, and to enter a "resting" or "stand-by" state. As mentioned above, hyperpolarization makes these cells harder to excite.

Unlike many other brain cells, thalamic relay cells don’t just remain silent when they are hyperpolarized. This is because the currents which develop within these cells are changed by hyperpolarization, a property which makes their behavior "voltage dependent". In thalamic relay cells this property causes the hyperpolarized membrane to let certain positive ions creep in. As a result, instead of not firing when hyperpolarized, these cells display a "built-in" slow depolarizing current which gradually brings them to their firing threshold, even when they are at rest. When they fire in this state they show unique bursts instead of single action potentials. Further, these bursts are repetitive.

The repetitive burst discharge results from another unique characteristic of thalamic organization. The axons of relay neurons simultaneously convey these bursts of excitation to the cortex and to another nearby thalamic nucleus. The exclusive role of this nucleus is to regulate the relay cells, and each relay cluster has its own special relationship with adjacent segments of this nucleus. Bursts in relay cells cause sister cells in this nucleus to burst also. When they burst, these cells feed back to the thalamic relay cells that activated them, but instead of exciting these cells they release an inhibitory neurotransmitter that reinstates the initial hyperpolarization. The re-hyperpolarized relay cells begin to slowly depolarize again and the whole process is repeated. This circuit organization results in an oscillation between the sister cells in these two thalamic nuclei that produces the repetitive thalamic relay cell bursts. Such bursting behavior can last for only a fraction of a second or for several seconds. Details about this organization and its functional consequences for the EEG can be found in an abundant neurophysiological literature 8-12.

The trains of thalamic relay cell bursts are directly relayed to associated cortical cells and, if they are also relatively unexcited, can cause them to burst in a similar pattern. As inattention is prolonged this oscillation encompasses more and more relay cells and triggers the networks between cortical neurons to draw more cells into the oscillation. The growing pool of oscillating cortical cells becomes synchronized over a widespread area. Our surface EEG electrode will now record synchronized alternating currents set up by this oscillatory pattern of discharge as repeated wave-like potential changes, thus giving rise to the term "brain waves" .

Since these waves are repeated, the interval between them, or their frequency, can be measured over time. In most adults the frequency of these waves in the inattentive state is between 8 and 12 cycles per second. When sustained over the visual cortex this pattern tells us that visual attention has decreased. We know that the individual is not processing the words on the page of their book. They are not necessarily sleepy, they are just inattentive. These oscillations in the EEG during visual inattention were first described by the pioneering German scientist Hans Berger 13, who considered them to be "first-order brain waves", and thus called them the "alpha rhythm". The kind of activated EEG pattern seen with attention (as when the spider was detected) was considered by Berger to represent "second-order brain waves", which he called the "beta rhythm". In the alert and visually attentive individual these "beta" waves briefly dominate the EEG from the posterior cortical surface. However, when we are visually inattentive "alpha" waves will be dominant. If we become generally unaroused and inattentive this activity will spread more generally over the entire cortex, as other systems go "off-line". Neuro-feedback training for the enhancement of this rhythm will increase visual inattention and can lead, eventually, to general relaxation.

In addition to the alpha and beta rhythms there is another important EEG rhythm that we need to discuss. This is a distinctive, faster rhythm that is recorded over a specific brain area. The rhythm is in the 12-19 cycle/second range, and the brain area where it is dominant is concerned with our conscious awareness of our body and the condition of our muscles and joints. It is thus called the sensorimotor cortex. Accordingly, the rhythm recorded here has been labeled as the "sensorimotor rhythm", or SMR. My laboratory was the first to study this rhythm in animals and later in humans 14,15. The SMR is seen clearly in higher mammals who are alert but in a state of sustained motionlessness, a condition some have described as "expectant". It is suppressed by movement and only occurs when the animal begins to shut down its voluntary motor pathway. That is, when the animal resists any impulse to move. When this happens the thalamic relay cells in the pathway that convey sensations from the muscles and joints to the sensorimotor cortex show a similar characteristic oscillatory bursting as was described earlier. However, because this condition represents motor inattention only, in an otherwise alert state, the interval between bursts is shorter, giving rise to the faster SMR. The reason for this faster frequency will be discussed later. The SMR behaves very much like the alpha rhythm but is related to a specific suppression of movement rather than to a reduction of visual attention 15, 16. Neurofeedback enhancement of the SMR thus produces a quiet body in the context of an alert mind.

It should be pointed out that the SMR is more difficult to see in the surface EEG of humans than are slower rhythms. There are several reasons for this discrepancy. First, in animal studies EEG electrodes are placed directly onto the brain. Interference in the conduction of brain wave signals through connective tissues, bone, and skin has been found to affect faster frequencies more than slower ones. Secondly, as we have said, the faster SMR is associated with relative alertness. In this state fewer thalamic and cortical cells are drawn into synchronized oscillation, resulting in a lower voltage and a more limited spread within the cortex. However, when electrodes are placed directly on the brain in humans, for example after the surgical removal of the bone over the sensorimotor cortex), this higher frequency activity is clearly apparent and resembles that seen in cats 17. Thus, under normal circumstances we must rely on sensitive frequency analysis equipment to aid us in the detection of the SMR.


Modulating the Frequency of Brain Waves

By now some of the reader’s own visual relay cells are beginning to show bursting behavior. If we don’t counter this effect this process will not only expand to include more and more cells but may lead, ultimately, to your falling asleep. If we were recording from you, we would know this because the characteristics of your brain waves would begin to change. Before this happens you might want to know how it happens.

The complex regulation of brain waves is directed by many different functional influences, including 1) the information from our senses and other important pathways that coordinate motor and internal responses, 2) influences from the various areas of the brain which express our learned biases and direct our particular interests at the moment, and 3) influences from other areas in the brain which set our general level of alertness. Thus, when we are attentive, are emotionally motivated, or plan a particular action, excitatory influences act on appropriate thalamic relay and cortical cells and promote selective depolarization in the service of our needs. Such events are signaled in the EEG by an activated pattern which can be either localized to a particular system or generalized to many areas of the cortex during complex cognitive activities.

However, if we are unmotivated, confused, or exhausted, which could certainly be the case here, a growing number of relay and cortical cells begin bursting. This will lead at first to localized alpha rhythm activity, usually in the visual cortex. As inattention is sustained, this rhythmic pattern increases in amplitude, becomes more widespread, and includes a slower component of alpha rhythm activity. Our model attributes these changes to the fact that the relay cells have become more hyperpolarized, more of them have joined the oscillation and, because of an associated and generalized reduction in neuromodulatory excitation, rhythmic activity has emerged in other relay systems as well. Eventually, the frequency of these rhythms begins to slow and rhythmic trains in the range of 4-7 cycle/sec. are seen. This is one example of slower activity commonly referred to as the "theta rhythm". It tells us that the individual has become drowsy and is falling asleep. You will know that this has happened when your face hits your desk!

We suggest that the passive reduction of excitatory influences from all sources in this state, together with conditioned inhibitory responses, results in a progressively greater hyperpolarization of thalamic relay cells. Such an increasing hyperpolarization has been shown to change cell membrane characteristics and to slow the built-in depolarizing current. This effect lengthens the interval between relay cell bursts 18. This happens because of the previously mentioned voltage dependent changes in membrane ionic currents.

Thus, in the normal brain, very fast low voltage EEG activity is associated with active excitation, the SMR is associated with reduced motor excitability and the absence of movement, intermediate alpha rhythm frequencies are associated with inattentiveness, and slower theta rhythm frequencies with drowsiness. Very slow delta frequencies are normally generated in infancy and early childhood 19, 20, and during stage 3-4 sleep in adults. While the SMR is localized, faster and slower frequencies can be either localized or generalized, depending on the pathways that are activated, placed in stand-by mode, shut-down, or disrupted.

Despite some contradictions in the EEG literature it is abundantly clear that all of these rhythms depend primarily on the integrating and oscillatory functions of the thalamus. It has long been known from clinical findings that structural injuries or degenerating processes that separate the thalamus from the cortex produce a featureless EEG, with only unstable slow activity 22. If the pathology is severe the pattern seen is persistent "polymorphic delta activity" or PDA. In less severe or more localized pathology the slow activity seen may be episodic and suppressed by arousal or cognitive engegement. This is called "intermittent rhythmic delta activity", or IRDA. These pathological rhythmic patterns occuring in relation to lesion-induced isolation of cortical tissues appear to be generated by intrinsic currents in these isolated cell populations.

The closest thing one sees to this activity in the normal brain occurs during stage 3-4 sleep, when thalamocortical interactions are functionally attenuated due to a deep hyperpolarization of thalamic relay neurons during this phase of sleep. Steriade et al. 21,23 postulate " a progressive hyperpolarization of thalamocortical neurons with the deepening of the behavioral state of EEG-synchronized sleep". One can presume that this state is different from pathological isolation, since findings from these studies indicate that cortical cells contribute to thalamic pacemaker functions, setting up a thalamocortical-corticothalamic oscillation in the case of sleep-related delta activity.

Perhaps the most convincing evidence for the critical role of the thalamus in EEG rhythm generation comes from a series of studies by Villablanca and associates in the 1970’s 24-26. With great skill and precision these investigators surgically removed the entire thalamus in a series of cats. They called this preparation the "athalamic cat". In other studies they simply separated the thalamus from its cortical projections in one hemisphere. In the latter preparations normal EEG rhythmic activity could be recorded from the intact hemisphere but the isolated cortex of the operated side showed only slow rhythmic activity resembling clinical PDA. However, EEG recordings from the separated thalamic relay nuclei showed normal rhythmic activity.

But the coup de gras was the athalamic cat. Initially, in these animals, regardless of the behavioral state determined by other physiological markers, the only cortical EEG pattern seen was the same slow PDA. This was true even during REM sleep (normally associated with an activated EEG), and was completely unaltered by the administration of barbiturate drugs. Only after a prolonged recovery period was active arousal accompanied by attenuation of these slow waves, likely due to the functional restoration of extra-thalamic excitatory neuromodulatory pathways.

I think that these considerations are very important in that they focus our thinking concerning the primary mechanisms underlying relevant EEG rhythms to the thalamus. For example, they provide a compelling argument for the fact that delta activity in the EEG can only occur under rather special circumstances involving deep hyperpolarization of thalamic and/or cortical neurons. Such hyperpolarization requires a significant reduction in excitatory input or a profound source of functional inhibition. What then are people talking about when they report quantitative increases in delta activity in normal awake and responsive subjects? Do they provide actual EEG tracings showing this delta activity? Can they show that this activity has the required amplitude and rhythmicity of true delta activity? Such "delta activity" is often indicated as occuring simultaneously with normal alpha activity. Can the mammalian nervous system do that? These reports come only from extracted numerical data derived by mindless computer-based frequency analysis. Such analysis is fraught with potential artifact and mathematical distortions in the low frequency domain. There will never be a zero output in any defined frequency range with current analysis methods. To those who think that these changes are relevant I say "show me the delta, baby".

As we learn more and more about the principles of thalamocortical and corticothalamic interactions, and how they relate to normal and abnormal circumstances, we acquire the functional tools to understand what we are seeing in the EEG. In fact, in order to assist the reader in remembering the position expressed here I have composed the following poems.

The Athalamic Cat

by M. Barry Sterman

"This is not the pussy in the hat, this is the athalamic cat!

He can’t make alpha all the day, he can’t make theta any way!

Its delta, delta on and on, because, you see, his thalamus is gone!"



The Happy Cat

by M. Barry Sterman

"This is indeed the cat in the hat,

He is healthy, happy and even a bit fat,

He can make alpha when he is out of touch,

He can make theta when he is sleepy or such,

But only makes delta when he is in bed,

or if he suffers a knock on his head."


What can this physiological model tell us about our patients?

From this perspective we can presume that the presence or absence of sustained EEG rhythms, as well as their location and their frequency, can tell us a great deal about how the brain is directing attention and planning behavior. We can also interpret deviations from a more informed perspective. For example, when adults close their eyes or become visually inattentive the dominant 8-12 cycle/second alpha rhythm emerges over the back of the head. This dominant activity is usually greatest and rather symmetrical over this region where conscious visual perception is localized. This response is normally well developed by the second decade of life. Prior to this age, however, the dominant frequency is slower 19, 20. Thus, at age 5 it is in the range of 4-7 cycle/second. Why is this dominant frequency slower in these younger children? As we saw above, the frequency of EEG rhythms is believed to reflect the combined influence of many functional systems acting on the thalamic and cortical generators of EEG rhythms. The more these generators are excited the faster the rhythmic oscillation produced by the thalamic machinery described above. It is reasonable to assume, therefore, that the immature status of the pathways which mediate general alertness and excite thalamic relay cells in early childhood results in less excitation and, accordingly, slower EEG rhythms. This condition is not abnormal but instead reflects the normal course of brain maturation.

Using modern EEG brain mapping it is possible today to identify both topographic and frequency deviations from this normal course of maturation. Thus, if a 15 year old child has a localized or pervasive dominant frequency in the 5-7 cycle/second range, we can presume that something is wrong. Brain mapping of the EEG has in fact demonstrated just such abnormalities in many children with ADD, and this information provides the primary rationale for neurofeedback therapy. In our experience EEG brain mapping is disclosing a number of different patterns of abnormality in these children, a finding which may aid in providing a more precise and reliable physiological differentiation of the various patterns of this disorder.

Remembering that neurofeedback can be viewed as a focused exercise program for selected pathways within the central nervous system and that its potential effects are driven by the Law of Effect, we can speculate as to how this treatment might facilitate therapeutic adjustments within thalamocortical networks. EEG electrodes placed over areas of EEG abnormality can be used to guide functional reorganization within these loops. Once the primary cortical area(s) showing this disturbance are identified (frequency and localization may vary with age, severity, or idiosyncrasies of the disorder), a neurofeedback strategy can be developed to promote the progressive normalization or optimization of EEG patterns throughout the area. Neurofeedback alone cannot accomplish this objective. As with any other behavioral treatment, it is essential to establish a close working relationship with the client, and to function as a personal trainer and source of motivation and educational feedback. With many disorders this will involve a program which seeks theoretically to reduce cellular hyperpolarization in thalamocortical circuits, and to raise the frequency of EEG rhythmic activity. Thus, the SMR protocol requires the patient to gradually adjust thalamocortical excitation through increased attention (suppression of slow activity) in the context of a quiet body (enhancement of the SMR.) The desire for rewards will initially promote transient changes in functional state. However, with repeated facilitation of normal interactions within relevant circuits, this exercise may also elicit progressive and more permanent changes in both functional and structural characteristics. We know that the neural plasticity required to achieve such changes exists. We have only to harness it effectively with solid guiding principles, methodology, and clinical expertise.

A focused and directed functional exercise effect, which we know can alter underlying systems if applied correctly, seems to be the most unique contribution of neurofeedback. It is hoped that these basic considerations will provide the reader with some framework for interpretation of the interesting findings to be encountered in the many and diverse papers which follow.


1. Sterman, M.B., Lo Presti, R.W. and Fairchild, M.D. Electroencephalographic and behavioral studies of monomethyl hydrazine toxicity in the cat. 1969, Technical Report, AMRL-TR-69-3, Air Systems Command, Wright-Patterson Air Force Base, Ohio.

2. Sterman ,M.B., Howe, R.D. and Macdonald, L.R. Facilitation of spindleburst sleep by conditioning of electroencephalographic activity while awake. 1970, Science, 167: 11461148.

3. Sterman, M.B. Neurophysiologic and clinical studies of sensorimotor EEG biofeedback training: Some effects on epilepsy. (1) In: L.Birk (Ed.), Biofeedback: Behavioral Medicine, 1973, Grune and Stratton, New York (Chapter 9), 1973: (2) Seminars in Psychiatry, 5(4):507525.

4. Sterman, M.B. and Friar, L. Suppression of seizures in epileptic following sensorimotor EEG feedback training. 1972, Electroenceph. Clin. Neurophysiol., 33:8995.

5. Lubar, J.F. and Shouse, M.N. EEG and Behavioral changes in a hyperkinetic child concurrent with training of the sensorimotor rhythm (SMR): A preliminary report. 1976, Biofeedback and Self-Regulation, 3 :293-306.

6. Hull, C.L. Principles of Behavior. Appleton-Century-Crofts, New York, 1943.

7. Thorndike, E.L. Educational Psychology: Volume II. The psychology of learning. Teachers College, Columbia University, New York, 1913.

8. Bal, T. And McCormick, D.A. Ionic mechanisms of rhythmic burst firing and tonic activity in the nucleus reticularis thalami: a mammalian pacemaker. 1993, J. Physiol. (Lond), 468: 669-691.

9. Sherman, M.S. and Guillery, R.W. Functional organization of thalamocortical relays. 1996, J. Neurophysiol., 76(3): 1367-1395.

10. Steriade, M. and Llinas, R. The functional states of the thalamus and the associated neuronal interplay. 1988, Physiol. Rev., 68: 649-742.

11. Steriade, M., McCormick, D.A. and Seinowski, T.J. Thalamocortical oscillations in the sleeping and aroused brain. 1993, Science, 262: 679-685.

12. Von Krosig, M., Bal, T. and McCormick, D.A. Cellular mechanisms of a synchronized oscillation in the thalamus. 1993, Science, 261: 361-364.

13. Berger, H. On the electroencephalogram of man, II. 1930, J. Psychol. and Neurol., 40: 160-179.

14. Sterman, M.B. and Wyrwicka, W. EEG correlates of sleep: Evidence for separate forebrain substrates. Brain Res., 6:319329, 1967.

15. Sterman, M.B. Physiological origins and functional correlates of EEG rhythmic activities: implications for self-regulation. 1996, Biofeedback and Self-reg., 3-33.

16. Kuhlman, W.N. Functional topography of the human mu rhythm. Electroenceph. clin. Neurophysiol., 1978, 44: 83-93.

17. Sterman, M.B. Studies of EEG biofeedback training in man and cats. 1972, Highlights of 17th Annual Conference: VA Cooperative Studies in Mental Health and Behavioral Sciences. VA Dept. of Medicine and Surgery.

18. McCormick, D.A. & Huguenard, J.R. A model of the electrophysiological properties of thalamocortical relay neurons. 1992, J. Neurophysiol., 68: 1384-1400.

19. Matousek, M. And Pertersen, I. Frequency analysis of the EEG in normal children and adolescents. 1973, In: P. Kellaway & I. Pertersen (Eds.) Automation of Clinical Electroencephalography. Raven Press, New York, pp. 75-102.

20. Matsuura, M., Yamamoto, K., Fukuzawa, H. et al. Age development and sex differences of various EEG elements in healthy children and adults - Quantification by a computerized waveform recognition method. 1985, Electroenceph. clin. Neurophysiol., 60: 394-406.

21. Steriade, M. Cellular substrates of brain rhythms. In: Niedermeyer, E. and Lopes da Silva, F. (eds.), Electroencephalography: Basic Principles, Clinical Applications, and Related Fields. 3rd Edition, Williams & Wilkins, Baltimore, 1993, pp. 1164.

22. Schaul, N. Green, L., Peyster, R. And Gotman, J. Structural determinants of electroencephalographic findings in acute hemispheric lesions. 1986, Ann. Neurol., 20: 703-711.

23. Steriade, M., Curro Dossi, R. and Nunez, A. Network modulation of a slow intrinsic oscillation of cat thalamocortical neurons implicated in sleep delta waves: Cortically induced synchronization and brainstem cholinergic suppression. 1991, J. Neurosci., 11: 3200-3217.

24. Villablanca, J. and Ruiz de Pereda, G. The athalamic cat, Sleep-wakefulness patterns. Arch. Biol. Med. Exp., 7: 45-46, 1971.

25. Villablanca, J. and Marcus, R. Sleep-wakefulness, EEG and behavioral studies of chronic cats without neocortex and striatum: The diencephalic cat. Arch. Ital. Biol., 110: 348-382, 1972.