Content uploaded by Siegfried Othmer
Author content
All content in this area was uploaded by Siegfried Othmer on Apr 30, 2021
Content may be subject to copyright.
Toward a Theory of Infra-Low Frequency Neurofeedback
Siegfried and Sue Othmer
ORCID ID: 0000-0003-1381-5279
This monograph is an augmented version of Chapter 3 in Restoring the Brain, Second Edition;
Hanno Kirk, editor; Taylor and Francis (2020)
Abstract
In Infra-Low Frequency Neurofeedback one is gifted with a cornucopia of compelling
clinical data, along with a paucity of quantifiability and sparseness of theory for their
interpretation. The real-time signal is meaningful only to the brain that produces it, and for the
real-time response one is dependent on client report. The training process is entirely self-
referential, i.e. endogenous, operating at the limit of subtlety at which good neuro-regulation
must necessarily take place. Nevertheless, clear patterns of responding have been observed.
Two primary failure modes have been identified and linked to two principal protocols that are
broadly impactful for the clinical population. Precision is available to us in the frequency
domain, and that has yielded a basis for understanding how the different brain regions
coordinate in the frequency domain. The identified pattern holds over the entire range of
frequencies of therapeutic relevance. Infra-Low Frequency Neurofeedback engages with a slow
control system that was first identified in animal research. This takes us to the foundations of
the developmental hierarchy, and as such facilitates recovery from early childhood trauma at
any time in life, as well as enabling the re-direction of developmental trajectories in infancy and
early childhood.
3.1 Introduction
Infra-low frequency neurofeedback is not readily subjected to formal evaluation by way
of group studies in the classic mode. Not only does the procedure have to be individualized to a
degree that is likely unprecedented in clinical practice, but the training procedure has to remain
adaptive throughout the training process. The multi-dimensional discovery process involved
here could only flourish in the clinical realm, and ILF training will likely remain a clinical frontier.
Undoubtedly, we are only at the beginning of the exploitation of this new modality.
1
At the same time, any arbitrary threshold of validity can be met even in this context by
the principle of Bayesian inference, without resort to group studies. The body of knowledge
reflected in this book testifies to that. The clinical model evolves incrementally in a manner
similar to the growth of the conventional lava flow at Mount Kilauea: plasticity at the frontier of
clinical practice, but leaving a trail of progressive solidity behind. Such plasticity is rare and
precious in the sciences that bear on human health. ILF Neurofeedback also offers intrinsic tests
of validity, by virtue of its parametric specificity in the frequency domain. And finally,
supportive evidence is starting to be furnished by more basic studies.
The more intractable barrier to acceptance is at the theoretical level. The scientific mind
resists being asked to take seriously data for which there is no agreed upon operative model.
The frontier of ILF NF has at times challenged our own belief systems, because the
clinical findings were so startling that they pushed the limits of our own credulity. Clients
respond rapidly and consistently to a relatively featureless low-frequency waveform. How is
this to be explained? In consequence, we have only the greatest sympathy for critics outside of
the field. If we hadn’t been confronted with our own ineluctable case data, and if we had not
had a chance to see systematic patterns emerge over the years, we would have been right
among them. Our challenge in this chapter, then, is to shape the data into a compelling
narrative, and to present a credible theoretical model.
3.2 Principal Categories of Brain Dysfunction
The top-level criterion for a clinical practice on the path to becoming an accepted
discipline is that of systemization. Can the basic facts of a field be categorized, systematized,
and manualized? In this regard, we face the challenge of working at the bottom of the
regulatory hierarchy, intervening with the lowest frequencies at which brain function is
dynamically regulated in a frequency-based manner. There are no established terms of
discourse for this problem. There are no measures that track the approach to this objective. We
are thrust back upon the self-report of clients with respect to their own self-regulatory status.
We are reduced to the client’s own ‘observables,’ by and large. The best we can do, therefore,
is to rely on the terminology that is used with clients. In addition, we also take advantage of
indices of autonomic function from physiological measures, and whatever we can discern from
trends in the EEG. Over the longer term, we can track improvements in cognitive function with
formal tests.
We begin by regarding the brain as a generic control system, one that has to satisfy the
requirements of any self-regulatory control system. Thinking of the brain in such holistic terms
takes us back to the very beginnings of the neurosciences. The neurologist Hughlings Jackson
originally propounded the concept of a “cerebral global function” that governs the
susceptibility to seizures. The fundamental burden of the brain is to assure its own categorical
stability, sufficient to sustain basic functionality. Macro-instabilities violating this criterion
include seizures, migraines, panic attacks, asthmatic episodes, and narcoleptic events. These
instabilities constitute the primary objective of a training strategy toward functional
normalization, as in their presence all subtlety of regulation is lost.
At the second level, the burden is to maintain setpoints of activation with respect to
various functional domains. The most global variable in this category is central arousal. This
concept also traces back to the origins of the neurosciences. As far back as 1895, Freud and
Breuer asserted that “a certain measure of arousal exists in the conductive pathways of the
resting, waking, engagement-capable brain.” This whole-brain property of necessity could serve
only a heuristic function then, and the same holds true now. Central arousal, in turn, modulates
sympathetic arousal of the autonomic nervous system, and vigilance, which comprises the
alertness of the attentional system, engagement of the executive function domain, and the
poise of motor system. None of these can be readily quantified. We are therefore compelled to
operate clinically with what is termed “ipsative trend analysis,” the discernment of change
induced via the training process in correlates of the categories of interest. In the absence of
measures, we are compelled to live with how the client interprets terms bearing on arousal,
vigilance, and excitability that the clinician brings into the discussion. In addition, there are
terms characterizing the affective state, the state of the autonomic nervous system, the quality
of sleep, and the status with respect to drives such as hunger, pain, and cravings.
This unsettling lack of quantifiability notwithstanding, one may still ask the question:
Just how we are thinking about terms such as arousal and vigilance? In the early years of our
work, we understood the term arousal in the Yerkes-Dodson sense. Central arousal is whatever
is under the management of the reticular activating system. (Even Yerkes-Dodson referred to
motivation before it became arousal in their famous plot!) That view is coming into some
question with the descent into the extreme low-frequency regime. The time constants of
responsiveness of the reticular activating system are fairly short—seconds to minutes.
Something else seems to be in a controlling role at the longer time constants, and with respect
to our clinical agenda, that appears to be the most relevant aspect. One can think of the fast
response of the reticular activating system as being precipitated by interactions with the
outside world. For our purposes, the more salient issue is the maintenance of the internal
milieu, our ambient or so-called resting state, where the time constants are typically longer.
Our clinical experience can be distilled into a primary concern with brain instability and
with arousal dysregulation. These two categories are foundational to the whole enterprise of
neural regulation, and both benefit from being readily describable by most clients. They map
into our two primary protocols for the initiation of the clinical work. This is possible because the
entire class of brain instabilities is responsive to the same protocol, or at least to the same class
of protocols. Similarly, arousal dysregulation primarily responds to a single protocol.
The context out of which brain instabilities arise tends to be one of neuronal hyper-
excitability. The roots of this concept also go back to Freud, who as early as 1894 referred to
the “sum of excitation,” which he saw as relevant to psychopathologies such as hysteria and
hallucinatory psychoses. As we now conceptualize the issue, there is a cellular (synaptic or
other membrane) aspect to hyper-excitability, one that is typically addressed by anti-epileptic
drugs, and there is a network aspect. The ILF training impinges on the network aspect in first
instance, and over time appears to also affect the setpoints of excitability at the cellular level.
That proposition is testified to by the observation that often levels of anti-convulsants can be
reduced or even eliminated through the course of neurofeedback training.
The context out of which arousal dysregulation arises tends to be trauma-based, where
this term is to be regarded in its most encompassing scope. Trauma does not have to be life-
threatening (or its equivalent) in order to wreak havoc with the course of early neuronal
development or to result in its dysregulation in maturity. The loss of a sense of safety, of
personal security, or even of status at any level tends to move the nervous system to a state of
over-arousal. This may become so well established as the new ambient that it is not perceived
as such. A new comfort zone develops such that a return to calm states may even give rise to a
sense of insecurity and loss of safety.
Physical injury such as concussions and other minor head injuries also disturb the
integrity of neuronal network relations, which likely constitutes their primary failure mode.
Commonly these are also observed to heighten neuronal excitability, particularly among those
with that vulnerability. Our greatest clinical challenges are those for whom both arousal
dysregulation and neuronal excitability prevail. In the case of the latter, we are typically dealing
with a genetically-mediated propensity. Brain instabilities such as migraines tend to run in
families, but they may not be evoked until triggered by a minor brain injury or other trauma.
That is also the case for seizures, where the vulnerability may well remain latent in the absence
of compounding events.
As over-arousal conditions tend to be trauma-related, they are typically environmental
in causation. With affect dysregulation and the fear response as mediators, some considerable
commonality in the neuronal failure modes is not unexpected. This may in turn account for the
fact that a standard protocol goes a long way to resolving these issues. Already noted is that
brain instabilities also tend to respond favorably to a single protocol, despite all of the variety in
which they manifest, and that likewise suggests a common failure mode.
3.3 The Trauma Model
The term trauma is at risk of becoming hackneyed and trivialized from over-
generalization. It is particularly at risk when apparently minor traumas are discussed in the
same context as major ones. But the fact is that even minor traumas can have major
consequences. This holds true for both minor emotional traumas and minor brain injuries. The
explanation is the obvious one. When even minor traumas afflict a vulnerable nervous system,
it can be tipped into major dysfunction. When we look at personal histories in such cases, a
consistent story can usually be told of a progressive vulnerability that finds its origins in early
childhood. This means that minor traumas cannot be considered in isolation in the general case,
or judged ‘on their own merits.’
This shifts our perspective from minor traumas, which are in fact ubiquitous in our
upbringing, to the matter of recovery. What really matters here is the dispersion that exists in
the distribution of recovery potential across the population. The wide variation in vulnerability,
in the lack of resilience, compels us to see the connection among apparently disparate events.
This process, in which a concatenation of apparently minor traumatic events can lead
progressively to major dysfunction, we refer to as the “Dysregulation Cascade.” Tracing such a
cascade to its causal origins commonly takes us back to a perilous environment for early
upbringing.
Perhaps the best exemplar of a Dysregulation Cascade is a boxing match, in which the
overt objective is the disruption of the neural integrity of the opponent. During the match,
there is little opportunity for functional recovery, so injury is cumulative. Full functional
recovery may eventuate, however, with nothing more than the tincture of time between
matches. What matters most in accounting for attrition in the careers of boxers is not the
magnitude of the blows received but the variation in recovery capacity. This observation
generalizes to the population at large.
This is such a critical issue that it bears further discussion. It has been found that even a
single change in domicile in a 14-year-old teenager doubles the cumulative risk (to middle age)
of attempted suicide. Roughly the same holds for the increased risk of substance abuse, violent
offending, and of any psychiatric diagnosis. What appears to be a minor traumatic episode may
in fact be quite significant in some children’s lives. The only reasonable explanation is that this
apparently ‘minor’ trauma has outsize consequences for a subset of the teenage population
that is already at risk by virtue of prior history.
2
What first brought this issue broadly into public awareness was the “Adverse Childhood
Experiences (ACE) Study”.
3
Here the focus was on overt traumatic experiences or contexts of
living: psychological, physical, or sexual abuse, etc. Evaluation of some 9500 adult
questionnaires yielded the finding that “Persons who had experienced four or more categories
of childhood exposure had 4- to 12-fold increased health risks for alcoholism, drug abuse,
depression, and suicide attempt.” Even more surprising was the correlation with chronic
medical disease: ischemic heart disease, cancer, lung disease, liver disease, and diabetes. With
four or more ACE’s the relative risk was elevated by about a factor of two.
Results were similar for a Scandinavian study that also evaluated the somatic health
impact of psychological stress. In tracking the increased incidence of disability-related pensions
post age 65, a dramatic correlation with early psychological stressors was brought to light. 3-7
stress factors yielded a doubling of the incidence of disability-related pension in the subsequent
five years. 9-12 stress factors yielded a risk multiplier of four. Overall, “over a quarter of…
disability pensions granted for somatic diagnoses could be attributed to psychological distress.”
Even more concerning, “…even mild psychological distress was associated with later onset of
long-term disability”.
4
Our first solid indication of the relationship between early development and general
health was furnished by the so-called Grant study on 256 Harvard students that began in the
late 1940’s.
5
Some fifty years later, the remaining 160 were evaluated for the usual diseases of
aging. If the person had grown up in a positive emotional environment, the incidence of such
diseases was 25%. If they had grown up in an adverse emotional climate, the incidence was
89%. The risk multiplier was an astounding 3.6. This is a more reliable figure than that furnished
by the ACE study, where the more severely impacted may already have attritioned out of the
population by the time it was questioned.
The Harvard study, on the other hand, has different limitations. It is representative of
the segment of the population that was not in economic straits. Socio-economic status is
known to be the largest single risk factor with respect to overall health and mortality. So, a
more inclusive view would assign even higher import to emotional wellbeing than is implied in
the Harvard study. A later study found that those with “six or more ACE’s died nearly twenty
years earlier than those without ACE’s”.
6
The clear implication is that our general health status—mental and physical well-being—
are closely correlated with early emotional upbringing. That brings us, then, to the question of
mechanisms. We distinguish between event trauma and a steady-state adverse living
environment. Event trauma transiently heightens memory function. The salience of an event
renders it state-stamped rather than date-stamped. It is registered as a whole-body memory,
with cognitive, affective, autonomic, and somatic responses fused into a unitary configuration,
bound together with the historical memory of the event. Irrespective of whether the individual
is personally at risk or is a mere witness, the event alters the setpoint of the threat response,
and it does so relatively permanently. This may well be protective of the individual, but it comes
at a cost to the physiology over the longer term. It has implications for the subsequent
development of neural network relations, and of neuroimmune and neuroendocrine system
activation.
In the matter of steady-state exposure to toxic living environments, we turn to the
research of Martin Teicher and colleagues at Harvard. This group uses the term maltreatent
trauma to characterize this population, which includes both overt mistreatment and abject
neglect, physical as well as sexual and emotional abuse. The impact is so substantial that in the
characterization of mental disorders, those with a maltreatment history may well constitute a
distinct ecophenotype.
7
Evidence for altered brain development is now coming into view.
8
Of particular interest
to us is evidence for altered functional connectivity, as illustrated in Figure 3.1. The evidence is
compelling that in unexposed brains one sees a healthy network under the aegis of the left
anterior cingulate, directing a confident interaction with the outside world. By contrast, in the
exposed case the well-elaborated network emanates instead from the right precuneus, which
has primary responsibility for securing safety, in collaboration with the right anterior insula. At
the same time, the frontal circuitry is impoverished in comparison with the unexposed
individuals. The trauma history, irrespective of its nature, has fundamentally altered how the
brain engages with the outside world and manages its internal regulatory regime.
9
What about the 25% of the remaining Harvard students who suffered chronic disease
but had a benign emotional upbringing? If neural network dysregulation, and dysregulation
more generally (e.g, autonomic), lies prominently within the causal chain of chronic disease, has
something been overlooked? Indeed. It is minor traumatic brain injury, which is an equal
opportunity immiserator, afflicting the rich and poor alike. Minor traumatic brain injury has
been just as much neglected in research as minor emotional trauma. Originally, the study of
traumatic brain injury was largely a military matter, and those who did not have a bullet in their
brain or a skull fracture were assigned to the category of minor traumatic brain injury (mTBI).
Figure 3.1. Network differences are identified for maltreated versus unexposed young adults. The green dots
indicate regions of interest for the left anterior cingulate cortex, the right anterior insula, and the right precuneus.
Connectivity analysis yielded the primary nodes that were common within each cohort. These are nodes with
direct connections to the region of interest (shown as salmon-colored dots). Secondary nodes (blue dots) are
linked to the region of interest only via the primary nodes. Jointly these linkages determine the ‘centrality,’ the
relative importance, of the three regions of interest for the two cohorts. (Source: Teicher, M.D., Samson, J.A., et al,
2016; Original source: Teicher, M.H., Anderson, C.M., 2014.)
This tragically misrepresented the clinical realities. mTBI has been a stealth condition
that has contributed to declining health status largely by the same mechanism of neural
network dysregulation. Symptoms often get worse over the first six months post-insult, which
means that—just as in the case of emotional trauma—the endogenous remedies that serve a
short-term purpose may exact a longer-term penalty, a process referred to as maladaptive
plasticity. Symptoms also get worse through cumulative insults, confirming the hypothesis of
the Dysregulation Cascade. It is finally being recognized that sub-concussive injury must be
taken seriously as well, particularly if it is part of a repetitious pattern, as in soccer play or
football.
10
None of these events need to rise to the level of creating organic injury within the
brain, as that term is ordinarily understood (i.e., neuronal shearing, etc.). They may not even
rise to the level of detectable symptoms. The primary mechanism of injury lies in the functional
domain. This proposition is attested to by the finding of a tendency toward hyper-connectivity
in mTBI.
11
Rapid recovery effected through neurofeedback training is also persuasive on this
point.
3.3.1 ILF Neurofeedback: Rescue Remedy for the Trauma Response
The trend to lower training frequency that has been underway since around the turn of
the century has, to all appearances, been driven by the imperatives of working with the
traumatized population, which presents the greatest clinical challenge to the neurofeedback
clinician. And in this project success has largely been achieved. Whereas there have always
been clients who could not be helped with these protocols, that has become largely a non-issue
with the available palette of training options, the extension into the deep ILF regime in
particular. But a larger reality has also been uncovered. It is not only those who bear the scars
of trauma that benefit from the two starting protocols. It is nearly everyone who comes for
training.
One must conclude that these protocols are redressing failure modes that our species
largely has in common. The tendency is for a challenged nervous system to move toward over-
arousal, with hyper-excitability an additional consequence if that is a vulnerability. When that
status cannot be sustained, the system may slide into under-arousal or crater into functional
collapse. The languor and effort fatigue that we associate with mTBI is a case in point. The time
courses seen in the anxiety-depression spectrum are another. Cratering is often masked by the
kindling of a disease process, being then naturally associated with the latter rather than with its
antecedent.
The prominence of early childhood adverse events in the life history of our most
challenging clients implies that neural network development is affected in its earliest stages. By
working at extremely low frequencies, we are addressing the foundations of the regulatory
hierarchy in three aspects: 1) the developmental hierarchy; 2) the functional hierarchy (from
the more general to the more specific; from the more distributed to the more localized); and 3)
the hierarchy in the frequency domain. In the latter, the lower frequencies set the context for
the dynamics unfolding at higher frequencies, and this hierarchy extends into the gamma range
of frequencies. These three constructs jointly inform the therapeutic hierarchy. The implication
of our clinical success is that even the intrinsic connectivity networks are sufficiently plastic so
that re-normalization of function can be mediated by way of ILF Neurofeedback.
Moreover, this appears to be possible at any age. The clinical agenda has become one of
re-normalization of the regulatory hierarchy from the bottom up with every client. It appears
that the residue of challenges to our early development resides in all of us at some level, and
that the mature nervous system can aid its own cause of functional enhancement by way of ILF
neurofeedback at these extremely low frequencies. Higher training frequencies then attend to
other levels in the hierarchy, to which the entire history of the neurofeedback field attests.
It may be helpful at this point to draw on a law of physics to elucidate the agenda: “The
Law of Least Action.” This is the principle that there should be minimal expenditure of energy
consistent with the ends to be achieved. The most efficient operation of the human brain
transpires at the levels of central arousal and of sub-system activation just sufficient for the
demand, but no more. Experts in the martial arts, meditators, chess and Go players are likely
well acquainted with this principle. Clients may become aware of it through the training
process, as they experience higher levels of vigilance, of alertness and mental clarity, even as
the nervous system is moved toward calmer states within a session.
3.4 Mechanisms of Regulation: Historical Roots of the Slow Cortical Potential
By 1935 the study of the EEG was substantially aided with the introduction of electronic
(tube) amplifiers by Hubert Rohracher. This called for capacitive coupling, which blinded us to
the slow potentials that lay beneath the cutoff frequency. In consequence, the world of EEG
research went dark on the Slow Cortical Potential (SCP) for about three decades. There was a
re-awakening in 1964 with the discovery by Grey Walter of the expectancy wave
(Bereitschaftspotenzial), and the publication by Nina Aleksandrovna Aladjalova of her extensive
animal studies on the tonic slow cortical potential in book form.
12
(Aladjalova, 1964). Soon
followed the engagement with evoked potentials and contingent negative variation, all of
which focused on the transient properties of the SCP. The tonic SCP was followed up by Joe
Kamiya, Karl Pribram, and Juri Kropotov. Intimating its importance, Karl Pribram referred to the
tonic SCP as “the second language of the brain.”
Aladjalova’s research has turned out to be of the greatest relevance to our present
purposes. She studied the infra-slow rhythmic potential oscillations (ISPOs) at great length. “A
single stimulation of the reticular formation immediately elicits an arousal reaction in the EEG
of the cortex, but has no effect on infraslow activity,” she writes. “This reaction is apparently
regulated by the rapid regulatory system. Stimulation of the ventromedial part of the
hypothalamus…intensifies infraslow cortical activity within 30-40 minutes. This reaction is
presumably regulated by the slow regulatory system.”
“…infraslow activity is intensified by certain actions after a long latency period, 30-100
and 120-200 minutes later. We conjectured that this phenomenon reflects the activity of the
slow control system of the brain…not only to automatically adjust the system to keeping
internal environment constant but actively to establish a new level of activity.”
It does not take a great leap to connect our training in the deep infra-low frequency
region with the slow control system Aladjalova identified. This system appears to be centrally
regulated by the hypothalamus, known to govern our internal milieu—autonomic function,
sleep-wake cycle, ultradian rhythms, etc. Thus, it makes sense that the ILF training extends
down to the circadian frequency.
It should be mentioned in the interest of completeness that the electrical stimulation of
yet other hypothalamic nuclei can also induce rapid state shift—even rapid de-activation—in
the same systems governed by the slow control system. For example, torpor could be suddenly
induced in a cat with suitable stimulation. This was the work of Walter Rudolf Hess, published in
1954.
13
ISPOs did not generate much interest again until fMRI imaging refocused attention
through the discovery of the intrinsic connectivity networks, or resting state networks, around
the turn of the century, some three decades later. They have been an intense area of study
over the past fifteen years.
14
,
15
,
16
Raichle and He have illuminated the connection between the
fMRI signal and the slow cortical potential.
17
It is these dynamics, predominantly in the range of
0.005 to 0.2 Hz, that are engaged in ILF neurofeedback. However, we impinge on this activity
indirectly via the contribution the extremely low frequencies make to their generation.
In 2017 we became aware that Giovanni Piantoni, of Mass General Hospital, had
identified slow brain rhythms of one to two-hour periodicity in extended recordings on epileptic
patients using depth electrodes. He identified these with the Basic Rest and Activity Cycle
(BRAC), a hypothesis that we entertained as well until we found it necessary to use even lower
target frequencies, and were thus compelled to broaden our perspective. One hypothesis does
not necessarily dispose of another. Over the entire ILF frequency range, we are no doubt
engaging with a number of core regulatory mechanisms, including in particular the BRAC.
3.4.1 The Regulatory Hierarchy
Seen from the vantage point of our neurofeedback challenge, the regulatory hierarchy
looks like what is shown in Figure 3.2. Early neurofeedback, including our own protocols,
conformed to the interests of cognitive neuroscientists by engaging with the attentional and
executive function domains, effectively the top of the regulatory hierarchy, as well as the motor
system, which was also well characterized. Our adoption of an adaptive procedure for protocol
refinement shifted the process from being prescriptive to being observational. That led ever so
gradually to the brain guiding us to its own priorities, namely the bottom of the regulatory
hierarchy, one client at a time, over the course of nearly two decades. This meant a progression
to lower training frequencies, which took place at a pace of about one decade in frequency
space per year for a number of years. This also meant a shift toward right-hemisphere rather
than left-hemisphere priority, as right hemisphere function is the first to develop and has
primary responsibility for personal safety as well as internal integrity and harmony.
18
It is also
where the primary vulnerability to psychopathology is lodged.
19
The sword-and-shield hypothesis is often invoked to concretize the above dichotomy.
The sword is typically wielded by the right hand, governed by the left hemisphere, whereas the
shield is borne by the left, governed by the right hemisphere. Left-handers most likely reverse
this pattern. A less graphic version is the approach/withdrawal dichotomy. We bestride the
world confidently with the left hemisphere in the lead, while the right is doing its best to keep
us alive and healthy.
Figure 3. 2. The Hierarchy of Regulation as seen in the perspective of the neurofeedback therapist. Of primary
interest is central arousal, which is intimately coupled to affective state, autonomic regulation, and interoception.
All are seen as principally under the management of the right hemisphere, which therefore demands the earliest
attentions. Executive function and cognition lie highest in the hierarchy, and are typically attended to later in the
training, if that is still necessary.
Of course, this simple dichotomy has not gone without challenge, but we spare the
reader a recitation of the counter-vailing evidence by reason of the following observation: In ILF
neurofeedback hemispheric reversal has not been observed. This finding is consistent now over
more than a decade, involving hundreds of thousands of clients, all being trained according to
the identical schema. We are engaging with the core architecture, beneath the level where
lateralized dominance—of eye, hand, and foot—is organized. In the early days of exclusively
EEG-band training, reversals of right and left were observed with several sets of mirror image
twins, as might well be expected. Since entering the ILF regime, however, we have not become
aware of work with mirror image twins.
We may well observe a reversal of dominance with the training, and we understand this
within the frame of birth trauma. Fetal thumb-sucking is the earliest indicator of laterality, and
on that basis a substantial laterality shift with the natural delivery process has been
documented (from ~95% right laterality to ~85% post-partum). That shift appears to be
reversible with the training at any time in the person’s life. The protocols with which that
laterality reversal is achieved, however, are themselves invariant with respect to laterality! The
resolution of the conflicting data appears to lie in the supposition that whatever observables
are relied upon to counter the universal assignment of core LH and RH control functions are
themselves confounded by laterality issues.
With reference to Figure 3.2, the clinical priority in almost all cases is to train the
regulatory arc of interoception, autonomic regulation, affect regulation, and central arousal. All
of these are highly correlated, reflecting a high level of functional integration. This objective
involves two protocols, targeting the posterior and the anterior aspect of the right hemisphere
sequentially. The priority lies with the posterior aspect, by virtue of its coupling to the posterior
hub of the Default Mode Network (DMN), which is the first to develop in infancy, and hence the
first to bear the scars of a non-nurturing environment.
20
The anterior placement yields our
most direct engagement with the Salience Network (SN) and the affective domain.
21
The posterior placement is the primary site for calming over-arousal of the nervous
system—of central arousal and of sympathetic activation. The latter should be demand-
responsive, leaving one in the general case in a state of sympathetic-parasympathetic balance,
or even of parasympathetic dominance—just as the lions of the veldt have modeled for us. If a
steady state of sympathetic over-arousal prevails, it is costly to our physiology, and it is the
right parietal placement that allows the system to de-escalate most efficiently and with the
greatest persistence.
Brain stability has to be promoted concurrently whenever that issue arises. This calls for
inter-hemispheric placement. Here the objective is good regulation, not merely the absence of
overt instabilities. This is most readily observable in the autonomic nervous system. Thus,
interhemispheric placement is called for not only to redress dysautonomia, or to tame
asthmatic episodes (which can be thought of as parasympathetically mediated paroxysms), but
to achieve good ambient autonomic regulation more generally. The subtle coordination
between the hemispheres turns out to be key to the objective of a dynamic balance between
the sympathetic and parasympathetic arms in the steady-state condition when the organism is
not under overt challenge or duress.
The DMN must be seen as the primary target of ILF NF, in that it is our resting state (i.e.,
task-negative) network, which governs our level of function and bears our dysfunction.
22
The
cortical resources it manages account for nearly all of the energy expended by the brain. The
Salience Network is the secondary priority.
23
The Salience network mediates between the task-
negative and the task-positive control network, the central executive.
24
It has a dual monitoring
role, one in which the insula attends to our internal status (interoception), and the anterior
cingulate portion tends to the interface with the external world. This activity is largely right-
lateralized.
25
This right-lateralization is most readily demonstrated in the description of Default Mode
connectivity relationships by Buckner et al (2008)
26
, as shown in Figure 3.3. Observe that the
connectivities linked to the right lateral temporal cortex, in the immediate neighborhood of the
right insula, are much more elaborated than those to the left. This is the essential point. The
lateralized hubs of the DMN (T3 and T4 and P3 and P4) have been our primary training sites,
but it should be noted that that has been the case since the late nineties, well before the DMN
was first characterized.
27
Figure 3.3 The principal hubs of the Default Mode Network are shown with their respective connectivities
reflected in the thickness of the lines between them. The right lateral temporal cortex (R LTC) is shown to be more
intimately connected to other hubs of the DMN than the left lateral temporal cortex (L LTC). These two sites are
involved in all lateralized placements and thus constitute the primary training sites. The other primary training
sites are the left and right inferior parietal lobules (L IPL and R IPL). The principal midline hubs of the DMN are the
posterior cingulate and retrosplenial cortex (PCC/Rsp) and the ventromedial prefrontal cortex, along with the
dorsomedial prefrontal cortex (vMPFC and dMPFC). HF refers to the hippocampal formation and PHC refers to the
parahippocampal cortex. (Buckner, Andrews-Hanna, and Schacter, 2008)
Despite the major shift in our clinical priorities over the years, the specific placements
have remained substantially invariant over that time. The shift from EEG-band priority to ILF-
priority has involved mainly some shift from upper tier to lower tier sites, e.g. from central to
temporal sites (C3 to T3, and C4 to T4). But in truth that shift may well have been more at the
conceptual level than the practical. For example, the standard placement for Sterman’s and
Ayers’ early work was C3-T3. So temporal placement has been in the picture since the
beginning of research with human subjects. However, the Sterman model concerned itself with
the sensorimotor strip exclusively, whereas presently the model concerns itself primarily with
the temporal sites, with T3 and T4 present in all lateralized placements.
The rationale for the shift to temporal priority emerged only after the clinical reality had
been thoroughly established. The constellation of principal training sites lined up with the
multi-modal association areas. This made sense since these areas are the most integrative in
character, and they rank highest in terms of functional plasticity. For both reasons, they should
therefore rank highest in training efficiency. This integrative character has been nicely
demonstrated in a determination of connectivity gradient from the primary sensory areas to
the multi-modal sites.
28
This is shown in Figure 3.4, which has been adapted from the original.
The connectivity gradient is minimal in the primary sensory areas and maximizes in the multi-
modal association areas. The loci of warmer colors identify our primary training sites: lateral
temporal cortex (T3 and T4); angular gyrus (P3 and P4), and the inferior frontal gyrus (Fp1 and
Fp2). At these sites, the Default Mode Network is accessible to us at the cortical surface for the
purpose of lateralized training.
Figure 3.4. The gradient of connectivity exhibits a distribution that minimizes in the primary sensory areas of cortex
(cool colors) and maximizes in the multi-modal association areas (warm colors). These connectivity maxima
identify our primary training sites for lateralized training: medial temporal cortex (T3 and T4); angular gyrus (P3
and P4); inferior frontal gyrus (Fp1 and Fp2). They also correspond to the regions in which the Default Mode
Network is accessible to us at the cortical surface for lateralized placements. A1, primary auditory cortex; S1
somatosensory cortex; M1, primary motor cortex; V1, primary visual cortex; mfg, medial frontal gyrus; infs, inferior
frontal sulcus; sfg, superior frontal gyrus; phf, parahippocampal formation; pmc, posteromedial cortex; cing,
cingulate; vmpfc, ventromedial prefrontal cortex. (Daniel S. Margulies et al, PNAS 2016: 133:44:12574-12579)
The Default Mode is accessible to us also at midline sites, and the primary linkage in the
DMN, between the posterior and the anterior hubs (The PCC and MPFC in Figure 3), beckons for
clinical attention. But this linkage turns out to train according to different rules. The differential
training with bipolar montage, which was so clarifying in lateralized training, so unambiguous in
its imperatives, turned out to be minimally productive when applied to midline sites. This
repeatedly side-lined our attentions to this critical linkage for some years. Here the primary
need is the enhancement of the coordination of the posterior and the anterior hubs of the
DMN. Alpha band synchrony training, long popular within the field, may already have been
serving this objective. To this agenda we have now added synchrony training in the ILF regime.
In some of the most severe cases of early trauma, ILF synchrony training can be the keystone
for functional restoration. In those cases where it is observed to be highly beneficial, it appears
also to be indispensable.
In the extreme cases of early maltreatment and/or neglect, the intact core self does not
have a chance to emerge because it is formed in relationship (the burden of the posterior hub)
before it is consolidated in agency (the burden of the frontal hub). Neglect disrupts the orderly
maturation of the posterior hub of the DMN in the first year of life, and the early phases of the
coordination with the frontal hub. That state of maladaptation is then further consolidated over
the course of development. Applying the ecophenotype model to this aspect, the mental health
universe can be said to divide between those in whom the front-back axis of the DMN is
profoundly dysregulated, and everyone else. Development takes us either on a boot-strapping
path of repair and recovery, or of further consolidation of dysfunction. The reed bends as it
lists, as it were. We end up with a bimodal distribution with little middle ground.
The Harvard group has looked at their bimodal distribution, distinguished as susceptible
versus resilient, and perhaps surprisingly found the same array of brain abnormalities in both.
29
Reduced nodal efficiencies distinguished the resilient cohort, particularly with respect to the
amygdala, and these were deemed to be neuroprotective. From our current vantage point, one
is tempted to conjecture that the differences may show up more prominently in the dynamics
than in static measures. Our clinical findings indicate that the remedy is to be found there as
well, and that whatever brain abnormalities exist do not present a categorical barrier to
recovery.
Some progress has recently been made in identifying a possible neurophysiological
representation of the core self. This emerged out of extensive studies of sleep by Andreas
Ioannides and colleagues in Japan.
30
A prominent characteristic of non-REM sleep is the high
degree of variability associated with these states. They are not homogeneous in character. In
that context, the stability and context-independence exhibited by two small regions attract
attention. They are located anteriorly and posteriorly on the left side of the dorsal midline
fissure, and are characterized by high levels of gamma-band activity. Curiously, this activity
level increases progressively from awake state to light sleep to deep sleep, and maximizes
finally in REM sleep.
Ioannides proposes that these two regions constitute the neural representation of the
core self. They are identified as the Midline Self-Representational Core (MSRC1 and 2).
Evidence for this is provided in the waking state, in which mental activities that are self-
referential and autobiographical tend to evoke activity in the penumbra of the MSRC1 and 2.
The Default Mode can therefore be characterized as a three-layer onion: The bulk is committed
to managing the resting or baseline state of the brain, and is most active when the brain is in a
non-engaged state; then there is the penumbra of the core self; and finally there is the core
self. The penumbra mediates between the core and the bulk of the DMN. The core self,
meanwhile, is preserved from ready alteration, particularly during the waking state.
It is during sleep states, in which the brain is largely non-engaged, that the opportunity
maximizes for accommodation by the intrinsic self to new realities that have been assimilated
during the waking state. Dreaming may be an essential part of this process. The inherent bias,
however, remains one of stability and of resistance to ready alteration on the part of the core
self. A threat to the survival or integrity of the self suffices, no doubt, to surmount this barrier
to change.
3.5 ILF Neurofeedback in the Frequency Domain: The Frequency Rules
Whereas nearly all of the terms of discourse utilized in connection with ILF
neurofeedback resist rigorous quantification, there is one singular exception, namely the
relationships among the optimal response frequencies that prevail at the different training
sites.
31
It is found that the ORFs for right-lateralized placements stand in harmonic relationship
Figure 3. 5. The frequency rules governing lateralized training are shown here. A harmonic relationship prevails in
the ILF region, whereas an arithmetic relationship applies in the EEG domain. The crossover is necessarily in the 2-4
Hz region of the delta band. Inter-hemispheric placement follows RH rules.
to the ORFs in left-lateralized placements. The ratio is universally a factor of two, with the left
hemisphere at the higher frequency. This is found in the context that harmonic relationships
are not commonplace in the EEG world. And indeed, a non-harmonic relationship for the ORFs
prevails in the EEG range, where the left hemisphere optimizes at a frequency 2 Hz higher than
the right. This is shown in Figure 3 5. Significantly, inter-hemispheric placements follow right-
hemisphere rules.
The crossover between the two regions is necessarily where the two criteria converge,
which is at a LH frequency of 4Hz and a RH frequency of 2 Hz. The 2-4 Hz range therefore
represents a major transition region between the domains where the ILF rules and the EEG
spectrum rules apply. The distinction has long been made between the delta band and the
theta band, and 4 Hz has been broadly accepted as the dividing line. So, we now have reason to
associate the delta band with the ILF regime in this critical respect.
The implications of the frequency relationships for model-building are likely profound,
although they can only be intimated at the present state of knowledge. First of all, the fact that
inter-hemispheric training follows RH rules confirms RH primacy in organizing the frequency
hierarchy. It is therefore more foundational in the regulatory hierarchy. This is consistent with
the observed dominance of our right-hemisphere placements with respect to the regulation of
the resting state.
Frequency rules have also been discerned for inter-hemispheric placements at
homotopic sites. With respect to the central strip sites of C3/C4 and T3/T4 that have garnered
most of the clinical attention in the history of EEG neurofeedback, frontal and pre-frontal sites
train at ORFs that are 2 Hz lower, and posterior sites train at frequencies that are 4 Hz lower. In
the ILF range a harmonic relationship once again applies, as frontal sites train a factor of two
lower than central, and posterior sites a factor of four lower. There has been little opportunity,
however, to explore these relationships in the ILF regime. This is for two reasons. First of all, we
have the historical circumstance that over most of the period of development of ILF NF, the
lowest frequency available in the software ended up being the preferred frequency, rendering
submultiples unavailable. The primary reason, however, is that most of the inter-hemispheric
training has defaulted the T3-T4 placement, and there has been little incentive to date to
explore other homotopic site pairs. This remains a task for the future.
The solidity of the findings with respect to frequency rules more than compensates for
the manifest shortcoming of the ORF phenomenology, namely that the entire basis rests on the
self-reports of clients. No other evidence in support of the model has ever been found.
Nevertheless, the reproducibility of the ORF from session to session, the one just as blinded as
the other, places the whole matter beyond dispute, unsatisfactory as that may be in the eye of
the critical researcher. Typically, the ORF only undergoes subtle migration over the course of
training. Moreover, one protocol has been found to alter the ORF slightly and, to all
appearances, systematically.
The ORFs are dynamically regulated, and the ILF training clearly impinges on that
process, subtly re-organizing the frequency-based properties of the neural networks. This
constitutes the most rigorous proof of validity over the entire frequency regime, in that the
observed frequency rules hold consistently over eight orders of magnitude, from 10-6 to 100 Hz.
Is there anything that renders the unitary quality of our regulatory regime more obvious than
this? When it comes to brain function, we are confronted with dynamics on all relevant
temporal and spatial scales—a continuous modulation of activation, of successive affiliation
and dissociation of neural assemblies—but certain relationships can be invariant and stable,
and so they appear to be.
3.6 A Resonance Phenomenon
We first published the observation that the behavior of the training process in the
vicinity of the ORF was reminiscent of a resonance phenomenon in 2008, so this model of the
process has engaged us for a long time.
32
The usual handicap prevails, namely the limits on
performing experiments in clinical settings. The standard resonance curve is shown in Fig. 6,
showing only the real component. It reflects the major features of the clinical experience. The
training appears to be more impactful at the center frequency, and also more unambiguously
positive, than training at nearby frequencies.
Figure 3.6. A standard resonance curve is shown to illustrate the dependence of the response on frequency with
the assumption of an operative biological rhythm at the ORF. Shown also is the effect of perturbations, including in
particular fluctuations in the ORF itself. If the spectral filter is mistuned with respect to the ORF, large errors can
creep into the signal that the brain may then misinterpret. These error signals are much smaller when the filter is
aligned with the ORF.
The approach to the ORF in actual practice can be analogized to a piece of music going
from dissonance to resolution. In the immediate vicinity of the ORF, the training experience can
be more complex, confounding, and even adverse. When this is encountered in a clinical
setting, the clinician feels obliged to find resolution in the ORF promptly. One does not linger to
establish reproducibility of such a phenomenon. Clients are not research subjects. Once the
ORF is found, it is not abandoned for the sake of scientific exploration. Hence the handicap. All
of the observations along these lines remain singular events for which patterns of
reproducibility could not be established.
The resonance phenomenon does not lend itself to ready evaluation in that it is not
being studied in isolation, as would be the case for electronic apparatus, for example. The
‘system under test’, so to speak, is the client’s brain in interaction with the signal. We are
seeing the response of a control loop with a sentient being in a controlling role. Once signal
acquisition occurs, in the sense that the brain has detected the correlation of the signal with its
own internal state, the brain undertakes to refine and particularize its response to that signal,
an essential part of the learning process. The control loop becomes a function of time, and
repeatability is not available to us in any event. As the training proceeds to lower frequencies
within the first session in the search for the ORF, a return to higher frequencies yields a
different response than was observed before—even after just a few minutes. The process is so
impactful that the caution of the Buddhist meditator prevails: “You never train the same brain
twice.”
Yet another confound is that the strongest responses to the training are observed with
the most dysregulated brains which are on their own unpredictable journey—particularly under
the provocation of the feedback loop. That is sufficient all by itself to obliterate any
expectations of reproducibility. Matters become much more manageable, however, once the
ORF is found, and a systematic path forward can usually be charted on the basis of ongoing
client reports. This ORF may have to be targeted within five percent or even less.
What could account for such a degree of parametric specificity? It is the requirement to
operate at the very top of the resonance curve, which is flat. Under these conditions the subtle
modulations on which the training depends are least compromised by all the confounding
factors that prevail in this experimental design. This is illustrated in Figure 6, in which an
arbitrary perturbation of the system is shown as a sinusoidal excursion on the frequency axis.
When the target frequency is mistuned to the skirts of the resonance curve, a large fluctuation
is expected in the signal output. At the ORF, the fluctuation is much smaller. This can explain
the much greater ‘turbulence’ that prevails when training on the slope of the curve near the
peak. A related issue is that of phase, which varies strongly in the vicinity of the resonance
peak. This phase combines with the phase shift induced by the filter to yield the phase of the
loop response function. A narrow workspace in the frequency domain follows. The resonance
curve giveth, and it taketh away—all within the scope of a few percent variation in the training
frequency around the ORF.
We are closing in on a model in which dynamically-organized biological rhythms exist
within the EEG and ILF realms that play key roles in organizing frequency-based relationships on
the large spatial scale—interhemispheric and lateralized. In the ILF realm, they govern resting
state dynamics on the longer time scales. Nevertheless, they must be responsive at the speed
of life. The resulting modulations are detectable if the brain itself is the detector, and they
become meaningful as the brain assigns meaning. The feedback loop effectively becomes
internalized, and thus makes possible the subtle refinement of resting state temporal
organization ‘within paradigm’, i.e. entirely within the resting state framework.
This process is capable of refining the regulatory regime to the limit of subtlety at which
it should ideally function. That cannot be accomplished via externalities such as reinforcers any
more than one could hope to improve Hilary Hahn’s violin playing by such means. This process
is also an answer to Karl Friston’s open-ended question back in 2009: Just how does one do
experiments on resting state organization (i.e., without disrupting what one is trying to study).
33
The answer is to let the brain operate within paradigm, absent any external challenge, and just
monitor the unfolding process. That is ILF neurofeedback.
3.7 Implications for Inter-hemispheric Coordination
The case has been made for the primacy of the right hemisphere in managing resting
state activation and coordination. All along, however, there has been a latent concern that a
shift to right hemisphere priority in the training, along with a shift from the EEG band to ILF,
might entail the neglect of what had previously been our priority, namely the training of
vigilance with left-central and left-frontal placements. That concern has been laid to rest.
Continuous performance tests have been done throughout this period of protocol evolution,
and they documented that no price was being paid with the shift in clinical priorities. The
quality of resting state organization governs even those functions that we associated with the
left hemisphere, and involve the engaged rather than the resting brain.
To illuminate the role of the left hemisphere, we turn to a seminal paper that looks at
information flow among the principal hubs of the Default Mode Network. These were originally
identified in the study of microstates by the Lehman group in Switzerland, well before there
was any talk of resting state networks.
34
These are shown in Figure 3.7. Each of the microstates
is identified with one of four hubs of the DMN, three in the posterior region and one with an
anterior locus. Two are lateralized, and two lie along the midline.
Figure 3.7. Shown schematically here are the four microstates that were originally identified by Lehman et al. Two
are lateralized; two are on the midline; three are posteriorly centered; only one has a frontal bias. (Ignore
polarities.) This tends to support the parietal bias in ILF neurofeedback. (Adapted from Lehman et al, 2014)
Information flow among these hubs was determined by means of a measure of directed
coherence on the alpha and low beta bands, leading to the finding that information flow was
dominant from the left hemisphere to the right, as well as from the left to the midline hub,
relative to the flows the other way. The imbalance can be substantial. This is shown in Fig. 8.
The clear implication is that with respect to the regulatory role of the lower EEG bands, the left
hemisphere is in a commanding position with respect to the right hemisphere.
Figure 3.8. The relative magnitude of information flow between the posterior hubs is shown here. This is derived
from calculations of directed coherence in the EEG bands of alpha and low beta. The information flow from left to
right vastly exceeds that from right to left, which implies that in the EEG range the left hemisphere is playing the
dominant role in organizing inter-hemispheric communication. (Adapted from Lehman et al, 2014).
A division of responsibilities is indicated. The right hemisphere bears the primary burden
of organizing our resting states while the left hemisphere supervises our engagement with the
outside world. The ILF regime plays a primary role in organizing the resting state configuration,
whereas the EEG regime handles the complexity, coordination, and temporal precision required
for our interface with the outside world. The delta band falls in the middle ground.
3.8 Foundational Research in Characterization of ILF Neurofeedback
ILF neurofeedback attracted the attentions of Dr. Olga Dobrushina of the International
Institute of Psychosomatic Health in Moscow, and of the Treatment and Rehabilitation Center
of the Russian Federation, also located in Moscow. Over the last several years they have
collaborated on a large-scale study of ILF NF using functional magnetic resonance. The objective
was to identify the networks engaged in this process of covert neurofeedback in a comparison
of veridical with sham training. 52 healthy volunteers were recruited to a single session of ILF
NF under uniform conditions, and resting state fMRI data were acquired immediately prior to
and again following the session.
35
Significant changes were observed in both groups, and there were systematic findings
among the members of each group, despite their heterogeneity. In the veridical training group,
“increased connectivity was observed through a network consisting of the right and left inferior
lateral occipital cortex, right dorsolateral prefrontal cortex and striatum nuclei.” The sham
training group, by contrast, showed increased involvement of the salience network but not of
the striatum. The authors proposed that whereas the salience network is responsible for the
conscious perception of rewards, the striatum plays more of a role in reward that lies beneath
consciousness.
This major study also contributes to the accreting body of evidence testifying to the
proposition that sham neurofeedback is not to be considered a neutral process. In order for the
control to be meaningful at all, the sham training group has to be given the same instructions as
the veridical training cohort. Both enter the study under the assumption of undergoing an
active process. In the search for persistent correlations that results, the brains in the veridical
group experience closure and get to settle down to an actual feedback process, whereas the
brains in the sham remain in a search status, one that is never graced with success. This can
account for the greater role of the salience network in the sham group, which is expected to be
largest when the salience question on the table cannot be satisfactorily resolved over an
extended period of time.
The theoretical aspects of neurofeedback are also starting to attract academic
attention.
36
Among the several models for neurofeedback, the skill learning model is also
discussed.
3.9 Summary and Conclusion
The state of our current thinking with respect to the basic mechanisms underlying ILF
neurofeedback has been presented in narrative fashion, appropriate to the state of knowledge
derived largely from clinical practice, which does not lend itself to experimentation for ethical
and other reasons. The basis has been laid for further fundamental studies of the ORF
phenomenology using the ILF feedback scheme as a probe of resting state functional
organization.
ILF neurofeedback is likely the most prominent exemplar of endogenous
neuromodulation—i.e., covert and continuous neurofeedback—in clinical practice. Whereas in
the ILF frequency domain there is no alternative, the advantages carry over to EEG-band
training as well. After all, it was in the EEG range that the ORF principle was first discovered
more than twenty years ago—also by way of covert, continuous feedback that was provided for
within the then-standard operant conditioning design. This approach allows the brain to
assume control through internalization of the process. Only endogenous neurofeedback can
take the process to the limits of subtlety and refinement at which brain regulation must
necessarily take place.
This skill learning modality, for which there is no known alternative, has import for the
entire realm of human functioning that involves the nervous system. It holds the greatest
significance for those contending with severe functional deficits acquired in the early stages of
development. Moreover, it offers a remedy available at any age when a need for the training is
identified, even down to early infancy. This is possible because the training imposes no
cognitive demand and is not contingent on conscious awareness on the part of the trainee.
References
1
Othmer, S., Othmer, S.F., Kaiser, D.A., Putman, J. (2013). Endogenous Neuromodulation at Infra-Low Frequencies.
Seminars in Pediatric Neurology, 20(4): 246-260.
2
Webb, R.T., Pedersen, C.B., Mok, P.L.H. (2016) Adverse Outcomes to Early Middle Age Linked with Childhood Residential
Mobility. Am. J. Prev. Med., 51(3), 291-300
doi: 10.1016/j.amepre.2016.04.011
3
Felitti VJ, Anda RJ, Nordenberg D, et al. Relationship of childhood abuse and household dysfunction to many of
the leading causes of death in adults. The adverse childhood experiences (ACE) study. Am J Prevent Med
1998;14:245-58.
4
Rai, D., Kosidou, K., Lundberg, M., et al (2011). J. Epidemiological Health, doi:10.1136/jech.2010.119644
5
Vaillant, G., Mukamal K. (2001). Successful Aging. American Journal of Psychiatry, 158:839–847.
6
Brown, D.W., Anda, R.F., Tiemeier, H., Felitti, V.J., & Edwards, V.J., Croft, J.B. Giles, W.H. (2009). Adverse
Childhood Experiences and the Risk of Premature Mortality, Am. J. Preventive Med. 37(5), 389-396 DOI:
10.1016/j.amepre.2009.06.021
7
Teicher, M. H. & Samson, J. A. (2013) Childhood maltreatment and psychopathology: a case for ecophenotypic
variants as clinically and neurobiologically distinct subtypes. Am. J. Psychiatry 170, 1114–1133 (2013).
8
Teicher, M.H. Samson, J.A., Anderson, Carl M., Ohashi, K. (2016). The effects of childhood maltreatment on brain
structure, function, and connectivity. Nature Reviews/Neuroscience 17, 652-666
9
Teicher, M. H., Anderson, C. M., Ohashi, K. & Polcari, A. (2014). Childhood maltreatment: altered network
centrality of cingulate, precuneus, temporal pole and insula. Biol. Psychiatry 76, 297–305
10
Talavage, T.M., Nauman, E.A., Breedlove, E.L., Yoruk, U., Dye, A.E., Morigaki, K.E., Feuer, H., and Leverenz,
L.J. (2014). Functionally-Detected Cognitive Impairment in High School Football Players without Clinically-
Diagnosed Concussion. J. Neurotrauma 31:327-338. DOI: 10.1089/neu.2010.1512
11
Muller, A.M., Virji-Babul, N. (2018). Stuck in a State of Inattention? Functional Hyperconnectivity as an Indicator
of Disturbed Intrinsic Brain Dynamics in Adolescents With Concussion: A Pilot Study, ASN Neuro 2018, 1-17 DOI:
10.1177/1759091417753802
12
Aladjalova, N.A. Slow Electrical Processes in the Brain. Elsevier Publishing Company, 1964. Elsevier.
13
Hess, W.R. (1954) The Diencephalon: Autonomic and Extrapyramidal Functions. Grune and Stratton, New York
ASIN: B015AFG0SU.
14
Vanhatalo, S., Palva, J.M., Holmes, M.D., Miller, J.W., Voipio, J. & Kaila, K. (2004) Infraslow oscillations modulate
excitability and interictal epileptic activity in the human cortex during sleep. Proc. Natl. Acad. Sci. USA, 101, 5053–
5057.
15
Monto, S., Palva, S., Voipio, J. & Palva, J.M. (2008) Very slow EEG fluctuations predict the dynamics of stimulus
detection and oscillation amplitudes in humans. J. Neurosci., 28, 8268–8772.
16
Palva, J.M. & Palva, S. (2012) Infra-slow fluctuations in electrophysiological recordings, blood-oxygenation-level-
dependent signals, and psychophysical time series. NeuroImage, 62, 2201–2211
17
He, B.J., Raichle, M.E.(2009). The fMRI signal, slow cortical potential and consciousness. Trends Cogn Sci., 2009
Jul, 13(7), 302-9
18
Chiron, C., Jambaque, I., Nabbout, R., Lounes, R., Syrota, A., & Duklac,, O. (1997). The right brain hemisphere is
dominant in human infants. Brain, 120, 1057-1065
19
Schore AN (1997). Early organization of the nonlinear right brain and development of a predisposition to
psychiatric disorders. Development and Psychopathology, 9 (1997), 595–631
20
Gao, W., Alcauter, S., Smith, J.K., Gilmore, J.H., Lin, W. (2014) Development of human brain cortical network
architecture during infancy. Brain Structure and Function, pp. 1-14 DOI: 10.1007/s00429-014-0710-3
21
Sridharan D, Levitin DJ, Menon V (2008). A critical role for the right fronto-insular cortex in switching between
central-executive and default-mode networks. Proc Natl Acad Sci U S A., 105, 12569-74
22
Broyd SJ, Demanuele C, Debener S, Helps SK, James CJ, Sonuga-Barke EJ (2009). Default-mode brain dysfunction
in mental disorders: a systematic review. Neurosci Biobehav Rev. 33, 279-96
23
Menon, V. (2011). Large-scale brain network and psychopathology: a unifying triple network model. Trends Cogn
Sci, 10:483-506
24
Menon, V., Uddin, L.Q. (2010). Saliency, switching, attention and control: a network model of insula function.
Brain Struct Funct 214(5-6):655-67 DOI 10.1007/s00429-010-0262-0
25
Sridharan D., Levitin D.J., Menon V., 2008, op. cit.
26
Buckner RL, Andrews-Hanna JR, Schacter DL (2008). The brain's default network: anatomy, function, and
relevance to disease. Ann N Y Acad Sci. 1124(1), 1-38.
27
Raichle, M.E., MacLeod, A.M., Snyder, A.Z., Powers, W.J., Gusnard, D.A., Shulman, G.L. (2001). A default mode of
brain function. Proceedings of the National Academy of Sciences of the United States of America. 98 (2), 676-
682. DOI: 10.1073/pnas.98.2.676
28
Margulies, D.S., Ghosh, S.S., Goulas, A., Falkiewicz, M., Huntenburg, J.M., Langs, G., Bezgin, G., Eickhoff, S.B.,
Castellanos, F.X., Petrides, M., Jefferies, E., & Smallwood, J. (2016). Situating the default-mode network along a
principal gradient of macroscale cortical organization. Proceedings of the National Academy of
Sciences, 113 (44) 12574-12579 DOI:10.1073/pnas.1608282113
29
Ohashi K., Anderson C.M., Bolger E.A., Khan A., McGreenery C.E. & Teicher M.H., Susceptibility or Resilience to
Maltreatment Can Be Explained by Specific Differences in Brain Network Architecture, Biological Psychiatry (2018),
doi: https://doi.org/10.1016/j.biopsych.2018.10.016.
30
Ioannides, A.A., (2018). Neurofeedback and the Neural Representation of Self: Lessons From Awake State and
Sleep, Frontiers in Human Neuroscience, 12, 1-20. doi: 10.3389/fnhum.2018.00142
31
Othmer, S., Othmer, S.F. (2017). Toward a Frequency-based Theory of Neurofeedback. Chapter 8 in Rhythmic
Stimulation Procedures in Neuromodulation, James R. Evans and Robert A. Turner, Eds., Academic Press (London),
2017, pp. 254-307
32
Othmer, S. (2008) Neuromodulation Technologies: An Attempt at Classification.
Chapter 1 in Introduction to QEEG and Neurofeedback: Advanced Theory and Applications (Second Edition), Thomas
Budzynski, James R. Evans, and Andrew Abarbanel, Eds, Elsevier, pp. 3-26.
33
Friston, K.J. (2009) Modalities, Modes, and Models in Functional Neuroimaging, Science, 326(5951), 399-403
DOI: 10.1126/science.1174521
34
Lehmann, D., Pascual-Marqui, R.D., Milz, P., Kochi, K., Faber, P., Yoshimura, M., Kinoshita, T. (2014). The resting
microstate networks (RMN): cortical distributions, dynamics, and frequency specific information flow. Available
from: http://arxiv.org/abs/1411.1949.
35
Dobrushina, O.R., Pechenkova, E.V., Vlasova, R.M., Rumshiskaya, A.D., Litvinova, L.D., Mershina, E.A., Sinitsyn,
V.E. (2018). Exploring the brain contour of implicit infra-low frequency EEG neurofeedback: a resting state fMRI
study. International Journal of Psychophysiology 131S (2018) S69–S184 doi:10.1016/j.ijpsycho.2018.07.217
36
Sitaram, R., Ros, T., Stoeckel, L, Haller, S., Scharnowski, F., Lewis-Peacok, J., Weiskopf, N., Blefari, M.L., Rana, M.,
Oblak, E, Birbaumer, N., & Sulzer, J. (2017). Closed-loop brain training: The Science of Neurofeedback. Nature
Reviews of Neuroscience, 18, 86-100
