Infra-low Frequencies and Astrocytes – Brief Summary

by Siegfried Othmer | August 21st, 2013

EEG Info NewsletterBy David Kaiser PhD – August 2013

Our brain is autorhythmic and demonstrates amazing stability over a vast array of rhythms spanning multiple time frames ranging from a few milliseconds to several hours. Infra-low frequencies (ILF), those at or below 0.1 Hz (which we call the Omega range) correspond to periods of many seconds to many hours. An ILF of 0.18 mHz, for instance, corresponds to a 90-minute cycle, the time it takes most of us each night to shift between non-REM sleep (also called slow wave sleep) to REM sleep and back again, from restorative cortical activity free of brainstem influences to periods of brainstem-generated excitability and plasticity known as rapid eye movement (REM) sleep. These cortical excitability cycles are apparent during the day as well in our EEG.

EEG Info Newsletter

Circadian rhythm, those which peak and trough once a day, corresponds to an ILF of 0.0112 mHz, a slow but very common rhythm observed in biological cells. Even under free-running dark conditions, circadian rhythms are remarkably stable in duration, though not in phase as we use the sun’s light to entrain us daily. Ultradian rhythms, those lasting from many seconds to a few hours, are not an evolutionary adaptation to the earth’s rotation but rather a response to metabolic challenges. A waxing and waning of the brain’s energy state is adaptive, and common to homeodynamic systems, those maintaining homeostasis in an energetically-variable environment. It is adaptive to run fast (be more active) periodically throughout the day, even if the cost is being slower or less active occasionally. Ultradian rhythms are present in every aspect of biology, from algae to house flies to humans. We observe ultradians in blood flow, oxygenation, neural firing rates, brainwave activity, sleep arousals, and even epileptic seizures.

EEG Info NewsletterWhat in the brain is cycling so slowly to account for our 90-minute or longer sleep cycle? Infra-low frequencies between 0.06 to 0.1 Hz correspond to blood flow and cortical network dynamics, but these are fast compared to the slow arousal fluctuations we observe. The thalamus shows oscillations as slow as 0.005 Hz, but still too fast to explain why we find strong effects from training ILFs as slow as 0.1 mHz. What in the brain cycles this slowly?

The answer may be glia. Humans are glial brains more than any other species. Glia take up nearly 90% of our cortex, and 80% of the cortex of our genetic relative the chimpanzee, but only 60% of rodents, and 20% of fruit flies. Across species, astrocytes, an important type of glia, increase in prevalence proportionally with the complexity of the brain. The astrocyte—neuron ratio is 1 to 25 for invertebrates but 7 to 1 in the human neocortex. The size and complexity of astrocytes has also increased in larger brain species. Astrocytes produce long-term fluctuations in energy, synaptic plasticity, as well as glutamate and calcium availability, jointly constituting a multi-faceted capability of regulating cortical excitability and plasticity. Thalamic astrocytes generate spontaneous and often highly rhythmic intracellular calcium oscillations as slow as 0.003 Hz. Astrocyte networks influence an even slower arousal cycle, the sleep-waking cycle. They control sleep pressure accumulation in part by inhibiting awake-state-promoting cholinergic neurons in the basal forebrain. Some believe that astrocyte activity may contribute to conscious modulation of brain rhythms in neurofeedback. They are likely responsible for the ILFs we train therapeutically.

For more information, and citations, see the full article at

Siegfried Othmer, PhD

9 Responses to “Infra-low Frequencies and Astrocytes – Brief Summary”

  1. Hi Siegfried and hi David (we met 17 years ago in Encino):
    I find the article fascinating and quite in synch with our work. Yet there is something that bothers me about it. If I remember correctly there are three kinks of biorhythms in living organisms all clumped together under the term RHEOSTASIS.
    1. The first are the CIRCADIAN RHYTHMS which literally means a rhythm that repeats itself around or close to every 24 hours
    2. The second are the ULTRADIAN RHYTMS which literally move themselves beyond a day like menstruation in women
    3. The third are the INFRADIAN RHYTHMS which literally move themselves beneath a day (repeat themselves several times per day) like the ones written about in this article.
    So if this is correct, and I think it is, David you should mend this mistake in terms.
    I hope I was able to of service. Please send me your feedback to

    Ricardo Rojas Bedoya
    Neuroscientist of Consciousness
    Neurofeedback Facilitator
    Lima Peru

  2. Ricardo—
    David’s article is adhering to the conventional usage of the term ultradian rhythm.
    It is a question of whether one is thinking in terms of the frequency domain or the time domain.
    In frequency terms, rhythms with shorter periods than 24 hours are characterized by higher frequencies than the 24-hour periodicity. Hence the term ‘ultra.’
    Since the term ‘ultra’ is commonly associated with frequency (e.g., ultraviolet versus infrared), there is overall consistency here.
    Such consistency, however, does not always exist.
    We are currently promoting the use of the term ‘infra-low’ frequency for the domain of EEG frequencies less than 0.1 Hz, but sometimes one sees the term ultra-low frequency in use as well.
    We hope that this usage will experience a well-deserved extinction.

    • Ok Siegfried. Thanks a lot but this usage sure is a very very shady area. In science we do strive to find a black and white definition of things even if they are different shades of gray. But I am able to see your point. Anyways I take the opportunity to ask you, I get the inkling that the next version of the CYGNET will be addressing at least two more notches beneath the current 0.1 mHZ. Is this correct? And if so, can you tell me when will it be out for public use? Please be notified that I want it!!!! Thanks Siegfried and I want you to know I am minding the admonition you sent my way the day we said goodbye. I love you and your family Siegfried. Be well.

  3. We don’t expect to go below the current range in frequency. We feel that way both for theoretical and practical reasons, and also because our clinical experience is consistent with this.
    Arguing from the ultradian rhythms takes us to this limit but no lower. That’s the theoretical reason.
    It is extraordinarily difficult to work at even lower frequencies. That’s the practical reason.
    And we have not had anyone—anyone at all—give any indication of needing to go below the present range for optimal response.
    Finally, even if we were to encounter such a person, the optimum would not be far away, one suspects, and one can manage with the existing range.

  4. Ok thanks. There is one thing “bothering me” and I am hoping you can shed some light on it. For argument´s sake let us say we are wanting to deal with an ANXIETY symptom, such as, “lack of bodily awareness” which appears in the RIGHT-BACK quadrant. And let us say this condition is accompanied by a given state of arousal: say, under arousal or the other way around, over arousal. How is this AROUSAL INFORMATION used to set the REWARD FREQUENCY in ILF? How is it used if it is an OVER AROUSED state or an UNDER AROUSED state? What does knowing this do to the reward frequency you decide to start with?
    Can you please help me with this? Thanks

  5. You are reflecting here the model that was operative when we were still working in the SMR/beta range, where our estimation of whether we were dealing with over- or under-arousal would govern our initial choice of reward frequency.
    In the ILF region that is no longer workable. Our judgment about whether someone is over- or under-aroused no longer gives us a clue as to where the training will optimize.
    The question is still relevant, however, in the following sense. If we move up in frequency from what we believe to be near the optimum, then the person will move in the direction toward over-arousal, and if we move down from that frequency, the person will move in the direction of lower arousal.
    So the issue is still relevant when it comes to fine-tuning the reward frequency.

  6. Ok great Sigfried. Your answer gives me the space where I can deploy my wings -which by the way I have been holding snug against my body in my communications with you.
    I will now attempt to be bold as I share something with you. The fact that you are no longer considering AROUSAL as a criterion to make decisions is valid in so far as building and trying out the latest PROTOCOLS you need to develop in order to handle your technology and its outcomes. That is cool. I respect that.
    BUT… this fact does NOT MEAN the real brain stops being underaroused or overaroused just because you decide not to take this aspect of its behavior into consideration. Any person that comes to session brings with him an under / over aroused brain: he walks with it and not only does he do everything in his life with such states but often times BECAUSE of such states!
    This decision just means you are not taking that into consideration anymore while you pursue other avenues; and I posit not only that you are making a conceptual mistake but that you are incurring in contradiction with your own words (not the ones stated here but the ones stated along the course and practicum) and wasting a valuable opportunity to do DATA MINING with the data you have so painstakingly gathered.
    If you would allow me I want to present to you my train of thought but I think it better if I do so via the private route unless you say otherwise. In the mean while I will start translating into english some of the documentation I have been developing and with which I will support this train of thought.
    I have been looking for a chance to carry on an amicable and intelligent conversation with you and here it is. May I proceed?

  7. You’re right of course that the brain coming in the door can still be described as under- or over-aroused. In the infra-low frequency regime, this is just not predictive of the target frequency, that’s all.
    We pay attention to what happens with arousal regulation just as much as we ever did before.

    The determinant of target frequency in this regime is something else. We are wholly constrained by the properties of the EEG spectrum, which is a characteristic of the person, and it is not not a matter of what ails him. It is the system that prescribes for us how it must be trained. It is in all cases optimum system functioning that is the objective of our training. We are not targeting the dysfunction.

    • Man do I get off on your distinctions. They are very well placed:
      “optimum system functioning is the objective of our training. We are not targeting the dysfunction”.
      “We are wholly constrained by the properties of the EEG spectrum, which is a characteristic of the person, and it is not not a matter of what ails him.” The EEG characteristics of the person versus the EEG characteristics of what ails him.
      THis is great. I shall process this as I prepare the material I spoke to you about. Thanks for your kindness and patience.

Leave a Reply