ArticlePDF Available

Computer Enabled Neuroplasticity Treatment: A Clinical Trial of a Novel Design for Neurofeedback Therapy in Adult ADHD

Authors:
  • NewPsy Psychoanalytic Institute

Abstract and Figures

Background We report a randomised controlled clinical trial of neurofeedback therapy intervention for ADHD/ADD in adults. We focus on internal mechanics of neurofeedback learning, to elucidate the primary role of cortical self-regulation in neurofeedback. We report initial results; more extensive analysis will follow.Methods Trial has two phases: intervention and follow-up. The intervention consisted of neurofeedback treatment, including intake and outtake measurements, using a waiting-list control group. Treatment involved $sim$40 hour-long sessions 2-5 times per week. Training involved either theta/beta or sensorimotor-rhythm regimes, adapted by adding a novel 'inverse-training' condition to promote self-regulation. Follow-up (ongoing) will consist of self-report and executive function tests.SettingIntake and outtake measurements were conducted at University of Helsinki. Treatment was administered at partner clinic Mental Capital Care, Helsinki.RandomisationWe randomly allocated half the sample then adaptively allocated the remainder to minimise baseline differences in prognostic variables.BlindingWaiting-list control design meant trial was not blinded.Participants54 adult Finnish participants (mean age 36 years; 29 females) were recruited after screening by psychiatric review. 44 had ADHD diagnoses, 10 had ADD.MeasurementsSymptoms were assessed by computerised attention test (T.O.V.A.) and self-report scales, at intake and outtake. Performance during neurofeedback trials was recorded.ResultsParticipants were recruited and completed intake measurements during summer 2012, before assignment to treatment and control, September 2012. Outtake measurements ran April-August 2013. After dropouts, 23 treatment and 21 waiting-list participants remained for analysis.Initial analysis showed that, compared to waiting-list control, neurofeedback promoted improvement of self-reported ADHD symptoms, but did not show transfer of learning to T.O.V.A. Comprehensive analysis will be reported elsewhere.Trial RegistrationComputer Enabled Neuroplasticity Treatment (CENT), ISRCTN13915109.Partly funded by Finnish science agency TEKES, project #440078.
Content may be subject to copyright.
CLINICAL TRIAL
published: 09 May 2016
doi: 10.3389/fnhum.2016.00205
Frontiers in Human Neuroscience | www.frontiersin.org 1May 2016 | Volume 10 | Article 205
Edited by:
Soledad Ballesteros,
Universidad Nacional de Educación a
Distancia, Spain
Reviewed by:
Lutz Jäncke,
University of Zurich, Switzerland
Juliana Yordanova,
Bulgarian Academy of Sciences,
Bulgaria
*Correspondence:
Benjamin Cowley
benjamin.cowley@ttl.fi
Received: 13 January 2016
Accepted: 22 April 2016
Published: 09 May 2016
Citation:
Cowley B, Holmström É, Juurmaa K,
Kovarskis L and Krause CM (2016)
Computer Enabled Neuroplasticity
Treatment: A Clinical Trial of a Novel
Design for Neurofeedback Therapy in
Adult ADHD.
Front. Hum. Neurosci. 10:205.
doi: 10.3389/fnhum.2016.00205
Computer Enabled Neuroplasticity
Treatment: A Clinical Trial of a Novel
Design for Neurofeedback Therapy in
Adult ADHD
Benjamin Cowley1, 2*, Édua Holmström3, Kristiina Juurmaa2, Levas Kovarskis3and
Christina M. Krause2
1BrainWork Research Centre, Finnish Institute of Occupational Health, Helsinki, Finland, 2Cognitive Brain Research Unit,
Cognitive Science, Institute of Behavioural Sciences, University of Helsinki, Helsinki, Finland, 3Formally affiliated with Faculty
of Behavioural Sciences, University of Helsinki, Helsinki, Finland
Background: We report a randomized controlled clinical trial of neurofeedback therapy
intervention for ADHD/ADD in adults. We focus on internal mechanics of neurofeedback
learning, to elucidate the primary role of cortical self-regulation in neurofeedback. We
report initial results; more extensive analysis will follow.
Methods: Trial has two phases: intervention and follow-up. The intervention consisted
of neurofeedback treatment, including intake and outtake measurements, using a
waiting-list control group. Treatment involved 40 h-long sessions 2–5 times per week.
Training involved either theta/beta or sensorimotor-rhythm regimes, adapted by adding
a novel “inverse-training” condition to promote self-regulation. Follow-up (ongoing) will
consist of self-report and executive function tests.
Setting: Intake and outtake measurements were conducted at University of Helsinki.
Treatment was administered at partner clinic Mental Capital Care, Helsinki.
Randomization: We randomly allocated half the sample then adaptively allocated the
remainder to minimize baseline differences in prognostic variables.
Blinding: Waiting-list control design meant trial was not blinded.
Participants: Fifty-four adult Finnish participants (mean age 36 years; 29 females) were
recruited after screening by psychiatric review. Forty-four had ADHD diagnoses, 10 had
ADD.
Measurements: Symptoms were assessed by computerized attention test (T.O.V.A.)
and self-report scales, at intake and outtake. Performance during neurofeedback trials
was recorded.
Results: Participants were recruited and completed intake measurements during
summer 2012, before assignment to treatment and control, September 2012. Outtake
measurements ran April-August 2013. After dropouts, 23 treatment and 21 waiting-list
participants remained for analysis.
Initial analysis showed that, compared to waiting-list control, neurofeedback
promoted improvement of self-reported ADHD symptoms, but did not show
transfer of learning to T.O.V.A. Comprehensive analysis will be reported elsewhere.
Cowley et al. Clinical Trial of Novel Neurofeedback Design
Trial Registration: “Computer Enabled Neuroplasticity Treatment (CENT),”
ISRCTN13915109.
Keywords: neurofeedback, attention deficit/hyperactivity disorder, attention deficit disorder, adult, randomized
controlled trial, waiting list control, learning curves, learning transfer
1. INTRODUCTION
Attention Deficit/Hyperactivity Disorder (ADHD) is a
neurobiological condition which can strongly affect several
areas of life: lower socio-economic status, less satisfaction with
employment and marriage, as well as common co-occurrence of
conditions like addiction and depression. Epidemiology research
has estimated the prevalence of ADHD among adults to be 4.4%.
Subtypes of ADHD have been identified including Inattentive
(ADHD-I), Hyperactive (ADHD-H) and combined (ADHD-C).
However, the nature of the disease and most effective method of
treatment are still not well understood.
This open-label clinical trial is a test of neurofeedback (NFB)
as a treatment intervention for ADHD or ADD-diagnosed
adults1. The intervention is based on the NFB training regimes
“theta-beta” (TB) and “sensorimotor rhythm” (SMR). These
regimes train self-regulation of power in specific bands of the
EEG frequency spectrum, through the principle of operant
conditioning. The trial also includes the novel addition of
a third type of training, “inverse training, designed to help
investigate the relationship between NFB learning performance
(the requirement to self-regulate) and the specific effects of each
training regime. This follows recent calls for increased focus on
the internal mechanics of the NFB process by Gevensleben et al.
(2014) and Zuberer et al. (2015), in order to attempt more than
just another simple test of efficacy.
We test intervention efficacy in a between-subjects manner
using a waiting-list control (WLC) group. The trial thus relies
on three different levels of measurements: neurophysiological,
cognitive, and behavioral; and examines their relationship across
multiple time-points. Transfer of learning is measured by:
questionnaires tapping ADHD/ADD symptoms that participants
filled out at four time-points during the intervention; and
by performance on a continuous performance test. More
importantly, given our focus on mechanics, we also test specific
effects of the NFB training regimes by analysing within-subjects
learning curves (LCs) of the treatment group, which represent
how they learn to self-regulate their EEG-band activity.
Abbreviations: ADHD, Attention Deficit/Hyperactivity Disorder; ADD, Attention
Deficit Disorder; BCIA, Biofeedback Certification International Alliance; CENT,
Computer Enabled Neuroplasticity Treatment; CNV, Contingent Negative
Variation; IBS, Institute of Behavioral Science; MCC, Mental Capital Care
(clinic); NFB, neurofeedback; RCT, randomized controlled trial; SCP, Slow
Cortical Potentials (neurofeedback training regime); SMR, SensoriMotor Rhythm
(neurofeedback training regime); TB, Theta/Beta (neurofeedback training regime);
T.O.V.A., Test Of Variables of Attention; UoH, University of Helsinki; WLC,
waiting list control.
1The ICD-10 diagnosis classification system is in use in Finland where the study
was conducted. Both ADD and ADHD are used as diagnoses, the former referring
to the predominantly inattentive type, as opposed to predominantly hyperactive or
combined types.
At the time of writing this is an ongoing two phase
trial, where phase 1 included intake measurements, NFB
intervention, and outtake (all complete), and phase 2 will include
follow-up measurements and WLC treatment opportunity
(pending). In addition to the protocol structure, we report
an initial group-comparison result from the first phase,
showing mixed outcomes for efficacy. This result is primarily
included to illustrate the importance of the planned within-
subjects analysis for obtaining clear insights from the data
gathered.
1.1. Attention Deficit/Hyperactive Disorder
Models seeking to explain the cognitive neuropsychological
problems associated with ADHD include disturbance of
attention, cortical arousal, and executive functions (for review
see e.g., Sergeant et al., 2003; Seidman, 2006). However, a
meta-analysis by Huang-Pollock and Nigg (2003) discarded the
explanatory value of attention, at least in individuals diagnosed
with the combined subtype of ADHD. Increasingly, ADHD is
not seen as a disorder of attention at all but as a disorder in
key aspects of self-regulation and executive functions (Nigg,
2005). One caveat is the growing consensus that the executive
function “single deficit” model cannot sufficiently explain ADHD
(Nigg, 2005; Pennington, 2005; Sonuga-Barke, 2005). Studies
indicate that not all persons with ADHD have executive function
deficits — at least as measured by laboratory tests.
Cortical arousal models in ADHD are closely related to
the attentional concept of alerting as proposed by Posner and
Petersen (1990), reflecting right-lateralized vigilance network
with noradrenergic involvement. These models emphasize
deficiencies in the early stages of information processing as
a result of under-arousal in cortical systems (Sergeant et al.,
1999). EEG and ERP findings tend to support this model in
that they reveal excess slow-wave activity in adults with ADHD
(Bresnahan et al., 1999). Support also comes from consistent
findings of deficit in the continuous performance test (CPT) d-
prime parameter, which can be considered a consensus index
of arousal (Losier et al., 1996). Epstein et al. (2003) found that
the d-prime demonstrated very robust relationships to the 18
DSMV-IV ADHD symptoms.
EEG studies suggest that ADHD might stem either from
a maturational lag or a developmental deviation (Barry et al.,
2003). Maturational lag models require that EEG measures
from an individual with ADHD would be considered normal
in a younger person, and implies that ADHD adults grow
out of their immature EEG activity with increasing age (Mann
et al., 1992). Whereas in the developmental deviation model,
ADHD is conceptualized as resulting from an abnormality in
the functioning of the central nervous system, unlikely to change
without targeted intervention.
Frontiers in Human Neuroscience | www.frontiersin.org 2May 2016 | Volume 10 | Article 205
Cowley et al. Clinical Trial of Novel Neurofeedback Design
Longitudinal studies, reviewed by Bresnahan et al. (1999), that
followed participants up till adulthood revealed that although
there is a significant reduction of slow wave activity in both
the ADHD and the control group with increasing age, absolute
and relative theta activity remained elevated through adolescence
into adulthood (Bresnahan and Barry, 2002). Interestingly, with
increasing age, the level of beta activity produced by adults
with ADHD was normalized in the frontocentral regions. The
most consistent finding from EEG studies of ADHD in adults is
increased absolute power in theta, clearly visible in frontocentral
areas (Bresnahan et al., 1999; Lazzaro et al., 1999; Clarke et al.,
2001). Such findings contradict the maturational lag model, as the
difference in slow activity does not disappear with increasing age.
By contrast, the hypo-arousal developmental deviation model
originally proposed by Satterfield et al. (1974) has been supported
by cerebral blood flow and positron emission tomography studies
(Lou et al., 1989; Zametkin et al., 1990). This model proposes
that ADHD results from cortical under-arousal, and the observed
atypical slow wave activity confirms the existence of altered brain
activity among adults with ADHD. Hypo-arousal is thought
to correlate with both beta and SMR (sensory motor rhythm,
also called low beta), because in normal functioning, increased
beta is associated with mental activity, and decreased SMR with
physical activity.
Much of the existing research has identified maturational lag
or hypo-arousal as the underlying cause of ADHD. Although
these models have initiated extensive research, they have failed to
clarify the aetiology of the disorder (Bresnahan et al., 1999). Thus,
the literature suggests that, at least in adult ADHD, aetiological
specificity is lacking; with the consequence that traditional
treatments targeted at ADHD as a single disorder are unlikely
to be reliable. In contrast, personalized medicine emphasizes
heterogeneity within a given disorder, relying on biomarkers or
endophenotypes to guide different treatments.
1.2. Neurofeedback
NFB, also called EEG biofeedback, is operant conditioning of
specific temporal, spatial and frequency features extracted from
scalp-recorded electrical potentials (Lubar and Shouse, 1976).
Feedback is presented to the treated individual in the form of
positive and negative reinforcers (in this study: visual reinforcers)
whenever their ongoing EEG features meet or fail to meet a
predefined criterion. The aim of NFB is to learn to gain control
of those EEG features over time.
Literature supports the efficacy of NFB for children with
ADHD (Arns et al., 2009, 2014b; Micoulaud-Franchi et al., 2014).
Part of its value is that NFB can be personalized to suit the specific
clinical presentation, provided that there is requisite theoretical
and observational data to guide the personalization.
NFB has been described as a mechanism that can stimulate
cortical arousal and/or regulate cortical oscillations, which in
turn may influence such cognitive activity as attention (Vernon
et al., 2003). The specific effect has been described variously as
following one of two models, termed by Gevensleben et al. (2014)
as “conditioning- and-repairing model” vs. “skill-acquisition
model.” This implies that the effect of NFB may be to repair
a presumed cause of disorder to normalize behavior, or instead
may be a tool to enhance cognitive performance (see Gevensleben
et al., 2014 for a thorough discussion).
It has been suggested that, besides the neurophysiological
aspects of NFB, treatment outcome depends greatly on the
subjective involvement of the patient. Calderon and Thompson
(2004) have conceptualized biofeedback as a three-step process
that consists of
becoming aware of a physiological response,
learning to control the response, and
transferring control of the response to everyday life.
The first two steps of the model — becoming aware and learning
to control the electrical activity of the brain — constitute NFB
learning. The third step refers to transfer of the NFB learning,
measured here by performance on a neurocognitive test as well
as self-reported ADHD related symptoms.
This study employs two kinds of separate NFB training
regimes: theta/beta and sensimotor-rhythm. Additionally, we
included a novel “inverse mode” of training, which is a
modification of each of these two regimes. Finally, transfer
trials during which the patient is given no visual feedback were
included toward the end of the trial.
Theta/beta (TB) training regime assumes a theta power that
is elevated above normal, and therefore uses an inhibition target
for theta power and a reinforcement target for beta power. EEG
recording is often at a frontal site. The rationale behind TB
training has been described in at least two different ways: as the
rectification of cortical hypoarousal (Barry et al., 2003), and as
the reinforcement of working memory (Vernon et al., 2003).
Sensimotor-rhythm (SMR) training regime reinforces beta
power, usually low or mid beta, often with an inhibition target for
theta. The site is above the sensorimotor strip, often lateral, such
that the beta oscillations correspond to the sensorimotor rhythm.
The rationale for SMR training has been proposed as either
facilitating attention (Vernon et al., 2003), or the improvement
of sleep through an increase in beta spindles, with concomitant
effects on cognitive function (Arns et al., 2014a).
It is important to note, that NFB learning is anchored in
two scientific theories, but occurrence of NFB learning as such
tests only one of these. On the one hand, NFB learning relies
on the cortical arousal model of ADHD that emphasizes under-
arousal in the cortical systems with excess slow wave activity
affecting information processing (Sergeant et al., 1999; Barry
et al., 2003). Based on this model, the NFB training aims at
increasing fast-wave activity (in this study: SMR and beta bands)
and decreasing slow-wave activity (in this study: theta) (Barry
et al., 2003). However, NFB learning as such does not test whether
the underlying problem of ADHD is under-arousal; what it does
test is the operant conditioning of EEG activity. NFB learning
is conceptualized in terms of changes in the amount of time a
patient manages to move his/her EEG features in the required
direction during training sessions as a result of learning to self-
regulate cortical oscillations. Zuberer et al. (2015) argue that NFB
outcomes should be tested by examination of the learning curves.
Thus, it is clear that, although aetiological and thus clinical
specificity for ADHD is lacking, all NFB treatment regimes share
a common goal of promoting self-regulation. On the other hand,
Frontiers in Human Neuroscience | www.frontiersin.org 3May 2016 | Volume 10 | Article 205
Cowley et al. Clinical Trial of Novel Neurofeedback Design
some more modern NFB regimes have a more explicit approach
to self-regulation than their older cousins. For example, Slow
Cortical Potentials (SCP) training uses two opposed cortical
regulation targets (Mayer et al., 2013), to be trained in random
consecutive order. The two most common NFB training regimes
TB and SMR do not include such an explicit set of counter-poised
targets to induce self-regulation, relying instead on a single target
of reinforcement/inhibition, which is trained repeatedly.
The target of SCP training is the Contingent Negative
Variation (CNV) Event Related Potential, which Mayer et al.
defined as a slow negative shift over central sites that develops
following the presentation of warning stimulus while expecting
an imperative stimulus that requires a response (Mayer et al.,
2015). Thus, while SCP directly addresses self-regulation, it does
so only for a single correlate of attention. Other cortical correlates
of attention processes are addressed by other NFB training
regimes, e.g., cortical hypo-arousal in TB, or spontaneous motor
activation in SMR. However, these training regimes have no
specific component designed to promote self-regulation.
Therefore, to the standard TB and SMR training regimes
we have introduced a mode of “inverse training” (denoted iTB
and iSMR), in order to explore the effect of adding an SCP-
like approach to these unidirectional training regimes. This
takes the form of an extra target in each regime, where the
reinforcer/inhibitor is the exact opposite of the norm (see
Methods).
TB and SMR training regimes are based on sub-second
frequency-band features, so they are not directly comparable with
SCP which feeds back the time domain DC component. However,
from our point of view, there are at least three motivating
reasons to test the “inverse training” self-regulatory modes of TB
and SMR.
First, the neurological effect of TB and SMR remains unclear,
due to the competing explanatory models. This study will not
lay all questions to rest, but it does pursue a novel line of
inquiry. Second, “inverse training” should aid the subjective
experience of self-regulation, as it does in SCP; after all from the
clinical point of view, NFB training in any training regime relies
on the patient’s own “mental strategy,” reinforced by feedback-
free transfer trials. Third, combining the above two issues,
the participants have the opportunity to learn the experiential
correlate of the inverse neural state, and thus learn to be able
to activate OR deactivate cortical resources at will. If we accept,
for example, that TB trains the activation of cortical arousal,
then the experiential correlate of inverse TB (iTB) might be
more appropriate when the individual needs to enter a state
of calm reflection. Similarly, SMR implies activation of the
sensorimotor strip, which in turn implies quietude of bodily
motor-neuron activity; however inverse SMR (iSMR) might be
more appropriate when particular task activity calls for so-called
kinaesthetic intelligence. This conceptualization of the process
would follow the “skill-acquisition” model of Gevensleben et al.
(2014). That is, the patient would gain a tool to enhance cognitive
performance, as opposed (or in addition) to repairing a presumed
cause of disorder.
Although earlier work (Lubar and Shouse, 1976; Monastra
et al., 2005), including meta-analysis by Snyder and Hall (2006),
has shown support for a single-trait model of ADHD (an elevated
theta-beta ratio), others have argued that research results and
clinical application should be interpreted with more regard for
variability of individuals (Arns et al., 2008). Hammond (2010)
goes into this issue in detail, illustrating the heterogeneity in
quantitative EEG (qEEG) patterns associated with symptoms
and discussing the requirements and need for qEEG analysis
guided by normative databases. Johnstone et al. (2005) provided
a review of such databases, along with a review of qEEG profiles,
which are manifestations seen between genome and behavioral
that they term “intermediate” EEG endophenotypes. They called
for QEEG endophenotype-guided NFB treatments to provide
non-pharmacological interventions to help the subgroup of non-
responders to traditional treatments, or complement traditional
treatments in certain cases.
Especially in adults, who are subject to maturation effects
across a broad age range, ADHD is a heterogeneous disorder
with an uncertain treatment situation. In other words, some
might have executive function deficits and might possibly benefit
from TB over the prefrontal cortex; while some might benefit
more from the characteristic behavioral correlate of SMR, that
is, immobility as well as reduction of muscular tension (Chase
and Harper, 1971; Howe and Sterman, 1972), thus facilitating
the self-regulation of attention through mechanisms similar to
mindfulness meditation (Zylowska et al., 2008). For this reason,
in this study TB and SMR training regimes are assigned in
a personalized fashion based on EEG spectral profiles (see
Methods).
1.2.1. Outline
In the rest of this paper, following the CONSORT guidelines, we
first document the Methods and design of the trial, including
participant criteria, intervention details, objectives, outcome
measures, sample size calculation, randomization procedure and
other allocation details, plus statistical analysis. Next, we provide
existing results from the trial as it stands, primarily regarding
how the treatment stage was run, along with preliminary
analyses of group comparison outcomes. Finally we discuss issues
arising from the trial design and implementation, as well as the
implications of the preliminary analyses and future work.
2. METHODS/DESIGN
2.1. Participants
Inclusion criteria were scores on Adult ADHD Self Report Scale
(ASRS) (Kessler et al., 2005), and Brown -ADHD scale (BADDS)
(Brown, 1996) indicating presence of ADHD, as well as:
pre-existing diagnosis of ADHD or ADD,
nonexistence of neurological diagnoses,
age 18–60 years,
IQ score >80 measured by a qualified psychologist using WAIS
IV (Wechsler, 2008)
Exclusion criteria included extreme outlier scores in the
scales of
Generalized Anxiety Disorder (Spitzer et al., 2006),
Frontiers in Human Neuroscience | www.frontiersin.org 4May 2016 | Volume 10 | Article 205
Cowley et al. Clinical Trial of Novel Neurofeedback Design
Beck Depression Inventory (Beck et al., 1996),
Alcohol Use Disorders Identification Test (Saunders et al.,
1993),
the Mood Disorder Questionnaire (Hirschfeld et al., 2000),
test of prodromal symptoms of psychosis (Heinimaa et al.,
2003), and
the Dissociative experiences scale (Liebowitz, 1992) for
dissociative symptoms.
Thresholds for exclusion were not fixed but at the discretion of
the consulting psychiatrist. Use of medication for ADHD was not
an exclusion criterion but participants were asked not to make
changes in medication during the time of the training. Informed
consent was obtained from each subject in accordance with the
Declaration of Helsinki.
Self-report tests were distributed and submitted by mail.
Psychiatric consultations were performed at the private practice
of the psychiatrist; IQ testing was performed at the testing
room of the University of Helsinki (UoH) Institute of Behavioral
Sciences (IBS); both in central Helsinki.
2.1.1. Ethical Approval
Written informed consent for participation was obtained from
all participants before entering the study. The protocol followed
the Declaration of Helsinki for the rights of the participants
and the procedures of the study. An ethical approval of the
present research protocol for all participants was obtained from
The Ethical Committee of the Hospital District of Helsinki and
Uusimaa, 28/03/2012, 621/1999, 24 §. Participants were not
remunerated.
2.1.2. Clinic of Treatment
The clinic for intervention sessions was required to be centrally
located in Helsinki, to be staffed by licensed psychiatrist in case
of emergencies, and to have recognition by the Association for
Finnish Work. Technicians were required to have a primary
degree in a discipline related to human psychology; they were also
required to take a 3-month training course provided at the UoH
based on principles of the Biofeedback Certification International
Alliance (BCIA).
2.2. Intervention
The experimental treatment was a novel neurofeedback
(NFB) intervention, based on the well-known operant
conditioning NFB training regimes “theta-beta” (TB) and
“sensorimotor rhythm” (SMR); with the novel addition of a
self-regulatory component designed to address the heterogeneity
and aetiological uncertainty of ADHD in the adult population.
The comparator was a WLC group. The WLC design places a
randomly assigned control group on hiatus, while the active
treatment is applied to the randomly assigned treatment group.
The WLC group should receive treatment after the follow-up
assessment at 24 months post-treatment, without experimental
oversight.
Participants who volunteered in response to advertisements
were recruited at time T0. Contact with the psychiatrist and
psychologist for screening followed at time T1.
Successfully screened participants were taken to the intake
measurement at time T2, where they performed the T.O.V.A.
test along with eyes-open and closed baselines, while scalp
EEG was recorded. The individual alpha peak frequency (IAPF)
of each participant was estimated from band power analysis
of eye-opened and eye-closed baseline conditions (Lansbergen
et al., 2011). The boundaries of each EEG frequency band for
each participant are defined with respect to IAPF, e.g., theta is
IAPF×0.4 to IAPF×0.6.
After randomization between treatment and WLC groups at
time T3, we assigned participants in the NFB treatment group
to either TB or SMR training based on their IAPF-adjusted
theta/beta ratio. Those with theta/beta ratio >1 (n=9) received
reinforcement for simultaneous increase in beta and decrease
in theta (over power estimated from per-session baseline) at
electrode Fz. The rest (n= 16) got reinforcement for increase in
SMR and decrease in theta at electrode C4. Band powers within
the NFB training regimes are adjusted by IAPF.
Treatment was administered using the following hardware
and software set up. The EEG amplifier was the Enobio
ambulatory device (Neuroelectrics SL, Barcelona)2, with
streaming Bluetooth connection to standard Windows 8 desktop
computers. The software was developed within the project, as
described in Cowley et al. (forthcoming). Briefly, the system is
based on OpenViBE signal acquisition framework3, with a Qt
frontend, and is available open source4.
NFB interventions were standardized by scheduling of the
training sessions: session duration was fixed; and training blocks
per session, sessions per week, timing of the break from training,
and total duration of training were all constrained to equalize
the intervention. At time T4, treatment group participants
began their treatment by being briefed about all aspects of the
NFB training regimes, e.g., length, frequency, purpose. Finally
for the first phase, outcome measures were taken at time T5,
when all participants in the treatment group had completed 40
sessions NFB.
In the first phase the care providers were monitored
by both the lead researcher and responsible psychiatrist on
separate occasions, with interviews to ascertain their self-
assessment of performance. Both care providers and patients
were given self-assessment questionnaires to describe their
working relationships.
The second phase of the trial is ongoing at time of writing.
Beginning with re-recruitment at T6, phase two consists of a
follow-up measurement for all, and treatment option for WLC
participants. Follow-up measurements include the ASRS and
BADDS self-reports. Treatment for the WLC group will not
be NFB, but will consist of a game-like computerized attention
training intervention, without concurrent recording of EEG.
2.3. Objectives
Our RCT research questions (RQs) follow Calderon and
Thompson (2004), since we first examine NFB learning “within
subjects,” and second examine the transfer of NFB learning
2http://www.neuroelectrics.com/.
3http://openvibe.inria.fr/.
4https://github.com/CBRUhelsinki/CENTplatform.
Frontiers in Human Neuroscience | www.frontiersin.org 5May 2016 | Volume 10 | Article 205
Cowley et al. Clinical Trial of Novel Neurofeedback Design
comparing the treatment group to a WLC group. Within this
paper, we report only first-stage analysis, namely the comparison
between groups addressed by H2a and H3c below (see Results).
2.3.1. Learning in NFB
The NFB learning metric reflects the proportion of time during
training when EEG signals are in the target state; the “learning
curve” is thus characterized by a signal evolving over blocks
and sessions of training. The shapes or slopes of participants
LCs are rarely reported in the NFB literature: analysis tends
to focus on transfer outcomes compared to a control group.
However, clinical observations commonly indicate that learning
occurs. Also, NFB learning in this study was manipulated with
the addition of the “inverse-training” mode. The LCs resulting
from normal, inverse and transfer training blocks should each
be slope-positive, because they each require a similar act of
concentration which the participants are practicing throughout
training. Finally, the profile of the LCs over sessions which
combine all training types should be slope-positive, because
training with counter-poised targets increases the need to self-
regulate. Thus, we propose the following hypotheses:
H1a: “normal” NFB training results in positive-slope LCs.
H1b: “inverse” NFB training results in positive-slope LCs.
H1c: “transfer” NFB training results in positive-slope LCs.
2.3.2. Transfer: Attention Test and Self-Reported
Symptoms
Due to meta-analyses that find that NFB is efficacious for
reduction of inattention (Arns et al., 2009), transfer of learning
is expected to result in more improvement-over-baseline of
the treatment group, compared with a control group, at
the continuous performance test (CPT) Test Of Variables
of Attention (T.O.V.A.) applied before and after training.
Furthermore, those participants who perform better in baseline
T.O.V.A. (lower scores), are expected to learn quicker during the
NFB training.
H2a: the treatment group will achieve better T.O.V.A.
performance, and improve more after training, than the WLC.
H2b: a better baseline T.O.V.A. score will predict better
baseline NFB performance and better NFB learning.
Severity of subjective symptoms of ADHD/ADD should be
reduced by the transfer of NFB learning to the ability to self-
regulate. We also expect this effect to be dose-dependent, such
that participants with better NFB performance (steeper positive
slope) should present a higher rate of change in reported
symptoms (steeper negative slope). Finally, the treatment group
is expected to report fewer symptoms than the control group in
the outcome measurement.
H3a: NFB training will result in a negative linear trend in
reported ADHD/ADD symptoms.
H3b: the NFB LC profile will correlate with reported
ADHD/ADD symptoms.
H3c: the treatment group will report greater improvements in
ADHD/ADD symptoms than the WLC.
2.4. Outcomes
Primary outcome measures include learning curve assessment,
T.O.V.A., ASRS, and Digit span. These measures fall into
two categories: (1) between-group comparison of NFB and
WLC groups; and (2) within-subjects tests for treatment group.
Primary measures are all comparative, truly experimental, and
hypothesis-driven.
Secondary outcome measures include pre- and post-treatment
vigilance measurement with an EEG protocol (Olbrich et al.,
2012); also per-session self-report of circadian patterns, mood,
excitement, effort and frustration; and the Pittsburgh Sleep
Quality Index (PSQI) administered pre-, post- and at two
intermediate points during treatment. These measures were
taken to explore the additional research question of the
relationship between sleep quality and NFB performance.
Additional methods taken to assess the quality of
measurements include
placebo expectation reports (Borkovec and Sibrava, 2005)
asked before and at halfway through the treatment,
the novel participant-technician interaction questionnaire,
designed to assess the participants’ experience of their
interaction with the technician. Based on the guidelines of
the “Framework for Measuring Impact”5, we created what
they term “specifically developed questions.” This option
was especially appropriate given the requirement to translate
any existing measure, which would limit validity. The
questionnaire was a set of ten simple questions such as “has
my trainer supported me during my neurofeedback training?”
Responses were made on a six-point Likert-scale from “0-not
at all” to “5-enough," for a scoring range of 0–50.
2.5. Sample Size
The power calculation of N=60 was based on an estimated
effect size for neurofeedback of 0.9, alpha at 0.05 and Power
at 0.95 (for an independent samples t-test, estimate based on
studies including Egner and Gruzelier, 2004; Rossiter, 2004a,b;
Arns et al., 2009). Recent work has shown an effect size of 0.73
for a purely adult population (Mayer et al., 2012), but we can
still maintain N=60 by assuming power =0.85, since this is
equivalent to an phase 1 trial where the motivation is greater to
minimize Type I error than Type II. This also contributes to our
motivation to use a WLC along with the principle that NFB is
closer in nature to a behavioral intervention than a drug trial, and
thus double-blinding is unwarranted and unethical.
2.6. Randomization Procedure
We assign patients to test or control groups using a procedure
that controls for selection bias, and known and unknown sources
of external variance due to prognostic variables; the procedure is
based on combined randomization and adaptive allocation.
Simple randomization prevents biased selection and meets the
assumptions of uniform assignment probability made in standard
inferential procedures. However, it does not guarantee equal
group sizes and can lead to baseline imbalance in prognostic
variables such as age, gender or disease severity. Equal group sizes
5http://www.measuringimpact.org/home.
Frontiers in Human Neuroscience | www.frontiersin.org 6May 2016 | Volume 10 | Article 205
Cowley et al. Clinical Trial of Novel Neurofeedback Design
can be obtained using blocking, while baseline imbalance can
be helped by using blocking stratified over prognostic variables,
although the number of variables which can be used is small. This
limitation is surmounted by adaptive allocation, exemplified by
minimization methods: these are not strictly random but allow
tight control of balance of multiple prognostic variables (see e.g.,
Roberts and Torgerson, 1998 on methodological issues).
In their review of recommendations of assignment method,
(Scott et al., 2002, p. 671) found a general support for
minimization with a random element in smaller trials. Our
approach follows: X% random blocking followed by 100-X%
minimization (0<X<100). The algorithm is:
1. Simple blocked random allocation of X% patients (of those
currently recruited).
2. For the next 100X% of patients, reassign patients based on
minimization.
3. Using a distance measure, make the assignment that results in
the smaller inter-group distance.
4. For every new assignment decrement the number of possible
new assignments to that group, to maintain equal group sizes.
Assignment was balanced over age, sex, education, IQ, diagnosis
(ADHD vs. ADD), comorbities from diagnosis, comorbities from
administered scales, and ASRS subtype score, and tested between
groups to show no statistically significant differences (see Table 1
below). The same tests returned null when run after any change in
relative group composition, due to e.g., drop-outs. Thus, groups
did not differ in terms of symptom severity, diagnosis, IQ or
demographic features at assignment.
Given that the initial allocation of patients is random, further
non-random selections will also be random at the population
level. Even if the minimization assignments cannot be considered
random, the overall assignment retains a proportional amount
TABLE 1 | Demographic and clinical characteristics for NFB and WLC
groups.
Treatment Control Statistical testing
N=25 N =29 t df p
Age Mean 35.72 36.45 0.259 52 0.797
Std. Dev. 9.66 10.86
WAIS-IV: VCI Mean 113.68 113.59 0.035 52 0.972
Std. Dev. 10.33 9.37
WAIS-IV: POI Mean 113.44 112.69 0.194 52 0.847
Std. Dev. 11.87 15.87
χ2df p
Gender Female 14 (56%) 15 (52%) 0.99 10.753
Male 11 (44%) 14 (48%)
Education Primary 6 (24%) 5 (17%) 0.470 20.791
Secondary 15 (60%) 18 (62%
Tertiary 4 (16%) 6 (21%)
ADHD/ADD ADHD 21 (84%) 23 (79%) 0.196 10.658
ADD 4 (16%) 6 (21%)
of randomness equivalent to X, similar to more complex
biased-coin approaches. In this trial X was set equal to 50.
The most important caveat of the approach is that variables
used in minimization must be included as covariates during
analysis, to avoid potentially misleading results.
Technicians must be assigned to treatment and control group
by recruitment on-demand, as the group treatment phases were
separated in time.
2.6.1. Allocation Concealment
Participants were randomly assigned by use of computerized
algorithm detailed above, at one time, directly after all intake
measurements and before the start of NFB treatment.
2.6.2. Implementation
The allocation algorithm was designed and implemented by the
lead researcher; the direct contact with participants for enrolment
and assignment was handled by a technician working for the
MCC clinic.
2.6.3. Blinding
Due to the fact that the trial used waiting list for the control
group, assessment was not blinded. Thus, after assignment,
all participants and researchers/technicians had access to the
assignment information. As mentioned above, since NFB is closer
in nature to a behavioral intervention than a drug trial, double-
blinding would be unwarranted and unethical.
2.7. Statistical Methods
Independent variables for group comparison include NFB
training regime (TB vs. SMR), participant age and gender, and
assigned group (treatment vs. WLC group). Dependent variables
(DVs) for group comparison include the T.O.V.A. and the ASRS
self-report.
T.O.V.A. variables, based on response times (RT) and error
rates, include RT variability (RTV) indicating consistency; mean
RT; Omission errors (OM) indicating inattention; Commission
errors (COM) indicating impulsivity; as well as the D-prime
score. D-prime is described as a measure of “perceptual
sensitivity” and has been suggested as an index of arousal (Losier
et al., 1996). T.O.V.A. variables are standardized in the analysis.
To evaluate the effect of NFB on T.O.V.A., we created
five new variables by subtracting the baseline scores from
the outcome measurement scores: RTV-change, mean RT-
change, OM-change, COM-change, and D-prime-change. These
T.O.V.A. change scores are subsequently compared for the
treatment group and the WLC, using independent samples t-test.
The Levene’s test showed that the variances for the two groups
were similar. Consequently, the independent samples t-test was
run with equal variances assumed.
MANOVA was subsequently used to evaluate the effect of NFB
training, compared to the WLC, in the outcome measures on the
five dependent variables of T.O.V.A.
ASRS consists of 18 items tapping the frequency of recent
DSM-IV criterion symptoms of adult ADHD, including a scale
for Inattention (IA, max 36 points) and a scale for Hyperactivity-
Impulsivity (HI, max 36 points). We calculated differences of
Frontiers in Human Neuroscience | www.frontiersin.org 7May 2016 | Volume 10 | Article 205
Cowley et al. Clinical Trial of Novel Neurofeedback Design
scores between baseline and outcome measurements to create
IA-change and HI-change scores. These difference scores are
subsequently compared for the NFB and the control group, using
independent samples t-test.
The Levene’s test showed that the variances in IA-change were
statistically significantly different in the two groups (F=4.36, p
< 0.05). As a result, the independent samples t-test for IA-change
was run with equal variances not assumed. The opposite was the
case for the variable HI-change.
In all our random coefficient models the intercept and
slope are separately estimated for each participant. That is, the
coefficients are estimated for each participant for the linear
regression equation as follows: Score =Intercept +B (Session).
All participants were treated with NFB at a single center.
All technicians were equivalently trained and capable; therefore
clustering of participants per technician was based simply on
scheduling logistics.
3. RESULTS
Results primarily describe the specific details of the intervention
implementation. Also, as stated, a preliminary between-subjects
comparison was performed after phase one to assess intervention
efficacy under a conventional analysis model. Thus, we report the
two straightforward pre- to post-treatment outcome measures
which are comparable between-groups: T.O.V.A. and ASRS. The
ambiguous results of this conventional approach, especially in an
unblinded context, supports the motivation to extend the analysis
with LC modeling. We therefore include these results in the
protocol report because they constitute an informative part of the
trial design going into the second phase.
3.1. Participant Flow
Eighty-two adults were recruited through cooperating clinics
Mental Capital Care, Neuromental and YTHS; also by newspaper
advertisement and posting to online forums for the Helsinki-
based ADHD society of Finland. Of this, 19 dropped out of
the trial before completing the screening process, due to various
issues.
Sixty-three participants were screened by a psychiatrist and a
psychologist prior to the training, resulting in nine participants
screened out of the trial due to one or more failures to meet
the criteria. The remainder (n=54) consisted of 29 females, 25
males, mean age 36 std.dev. 10 years, with 44 ADHD and 10 ADD
diagnoses.
Participants were split equally between treatment (n= 27)
and control groups (n=27); however two switched from
treatment group (final n= 25) to control group (final n=29)
for personal reasons. From these assignments, eight dropped out
from the WLC group (including one participant whose pre-test
measurement data was then deleted by request), and two from
the treatment group. Thus, 23 treatment group, and 21 WLC
group cases (total n=44) were available for analysis. The trial
progression is shown in detail in Figure 1.
Out of those who completed NFB, five participants did not
complete treatment on the pre-defined schedule, of which three
exceeded by more than a week. Delays were due to personal
reasons, causing a number of cancelations of scheduled sessions,
which is a regularly observed phenomenon in this diagnostic
group.
3.1.1. Implementation of Intervention
In practice, NFB training consisted of 40 sessions (range: 38–
41) during 2–4 months. There was a mid-training pause of
nominally 2 weeks. Patients came to the sessions 2–5 times
a week. One session lasted 1 h, subdivided into self-report
of mood, excitement, hours slept and hours awake; electrode
attachment; baseline measurement; 5–7 units of 5 min NFB trials;
and debrief including self-report of effort and frustration. During
each session, patients played different NFB “game” trials during
which they got immediate visual reinforcement for classifier-
matching states in their EEG. The scores per game trial are
baseline-adjusted and averaged per session to form characteristic
LCs. The content and purpose of the training sessions followed a
phased timeline:
1. Tutorial stage, for becoming accustomed to NFB, two practice
sessions: participants were given normal NFB trials with
baseline thresholds adjusted by a constant factor to make the
training easier;
2. Beginner stage, for NFB training, 18 sessions up to halfway
break: normal NFB with non-adjusted baseline thresholds;
3. Intermediate stage, for learning to self-regulate, ten sessions
from half-way to session 30: normal training blocks were
gradually reduced in number to half per session, and inverse
training blocks introduced in their place;
4. Expert stage, for transfer training, ten sessions until session 40:
as Intermediate stage, but also with one to two “transfer” trials
with no feedback stimuli.
3.1.2. Recruitment
As shown in Figure 1, participants were recruited in May/June
2012, with intake measurements during July/August 2012.
Randomization took place in early September; treatment began
September 17, 2012. Outtake measurements ran from April until
August 2013.
Follow-up measurements are planned for start of 2016.
3.2. Baseline Data
The participants who began the trial (n=54, 29 females) had
mean age 36 years (std.dev. 10 years), with 44 ADHD and 10
ADD diagnoses. The characteristics per group are shown below
in Table 1.
All intervention sessions were performed at the Mental
Capital Care (MCC) clinic premises in Helsinki, which met all
requirements described above. Technicians were recruited to
administer the NFB from IBS and MCC; they were trained at
UOH over a 3 month period; four of the team also attended
the Biofeedback Certification International Alliance (BCIA)-
accredited introductory course at the Brainclinics education and
treatment center, Netherlands.
3.2.1. Numbers Analyzed
The analysis included 23 participants in the treatment group, and
21 participants in the WLC group. Given that dropouts were not
analyzed, this analysis was not by intention-to-treat.
Frontiers in Human Neuroscience | www.frontiersin.org 8May 2016 | Volume 10 | Article 205
Cowley et al. Clinical Trial of Novel Neurofeedback Design
FIGURE 1 | Flow chart of assessments and treatments in the CENT trial.
Frontiers in Human Neuroscience | www.frontiersin.org 9May 2016 | Volume 10 | Article 205
Cowley et al. Clinical Trial of Novel Neurofeedback Design
3.3. Outcomes and Estimation
For H2a we find no support after analysis of the five
T.O.V.A. indexes; Table 2 shows their mean differences between
baseline and post-training. Changes in these variables were not
significantly different for the NFB group than for the WLC.
Results of the MANOVA at the outcome measurement revealed
that on the Wilks’ Lambda the difference in means between the
NFB and WLC did not reach significance [F(5,38) =0.45, p
>0.05]. Thus, NFB training had no significant effect on the 5
indexes of T.O.V.A. at the end of the intervention.
Table 3 presents the mean difference of the two indexes
between baseline and post-training and the results of the
independent sample t-tests comparing IA-change and HI-change
between NFB group and WLC. The NFB group presented
a higher reduction of inattention symptoms than the WLC
t(36.03) = −2.14, p< 0.05. Similarly, while the NFB group
presented a reduction of HI symptoms from baseline to post-
training, the WLC presented an increase in HI symptoms
t(44) = −2.42, p< 0.05. Thus, we find statistically significant
support for H3c. That is, the treatment group reported greater
improvements in ADHD/ADD symptoms than the WLC.
A more detailed analysis concerning the rest of the hypotheses
will follow in a separate paper. As treating the hypotheses H1a-b,
H2b, and H3a-b requires substantial additional methodological
reporting, addressing them does not fit the scope of a clinical trial
report.
3.3.1. Adverse Effects
No adverse effects were observed in the treatment group. Further
investigation of this question is planned for follow-up, using the
state-oriented self-report items described above.
4. DISCUSSION
4.1. Interpretation
Regarding the relationship between NFB learning and
performance in the continuous performance test, H2a proposed
that the NFB group will achieve better T.O.V.A. performance,
and improve more after training, than the WLC. Results of
this study did not find evidence for such transfer. Patients
participating in the NFB training did not perform better than
WLC on the 5 indexes of T.O.V.A. in the outcome measurement.
This result can be interpreted in at least two ways. On the one
hand, it can be a sign of the all too common problem of transfer
of training (Green and Bavelier, 2012).
Lack of transfer is one of the most important of the several key
obstacles pertaining to the effect of NFB trainings (Gazzaniga,
2009, p. 94). Because brain plasticity is highly task specific,
training in a specific task shows little or no improvement on
related tasks. On the other hand, the results of this study can
mean that NF learning bears no relationship to performance
on any indexes of the T.O.V.A. test. This would contradict the
findings of Losier et al. (1996) who considered the D-prime index
of T.O.V.A. a consensus index of arousal, which is, in turn,
assumed to be a manifestation of excessive slow wave brain waves
in ADHD patients (Barry et al., 2003).
TABLE 2 | Groups statistics for RTV-change, mean RT-change,
OM-change, COM-change and D-prime-change scores from baseline to
outcome.
NMean Std. t df p
RTV-change 0.343 42 0.739
NFB group 23 5.56 40.13
WLC 21 9.44 34.15
Mean RT-change 0.132 42 0.896
NFB group 23 3.37 16.23
WLC 21 3.94 11.85
OM-change 1.19 39 0.240
NFB group 23 10.10 72.73
WLC 21 46.25 198.84
COM-change 0.517 42 0.608
NFB group 23 6.19 27.49
WLC 21 1.81 28.67
D-prime-change 0.016 42 0.987
NFB group 23 1.50 45.30
WLC 21 1.72 44.73
TABLE 3 | Groups statistics for IA-change and HI-change scores.
NMean Std t df p
IA-change 2.14 36.03 0.039
NFB group 25 1.2 2.17
WLC 21 0.14 1.06
HI-change 2.42 44 0.020
NFB group 25 1.08 2.31
WLC 21 0.38 1.65
H3c suggested that the NFB group will report greater
improvements in ADHD symptoms than the WLC. Results
show the change of IA was significant. Patients did perceive
a reduction of inattention symptoms over the course of the
training. Furthermore, this perceived reduction of inattention
symptoms differed significantly from the perceived reduction of
inattention symptoms of the control group. This supports the
meta-analysis by Arns et al. (2009) concluding that NFB has
large effect sizes on inattention. In the index of HI, no negative
linear trend was found. Interestingly however, patients in the
NFB group did perceive a significantly larger reduction of these
symptoms over the course of the training than the control group.
It might be that the training indeed resulted in some reduction
of IA and HI symptoms. This interpretation gets support
from theories claiming that ADHD is, in effect, pathology of
executive functions that cannot be tapped by neurocognitive
tests, but can instead be measured by self-reported questionnaires
(Rabbitt, 1997; Brown, 2009, pp. 81–116). Alternatively, the
results can also be interpreted as an example of the so called
Hawthorne effect (Green and Bavelier, 2012). Establishing the
presence of experience-dependent learning effects is not always
straightforward. It is well documented, that individuals who take
an active interest in their performance tend to improve more,
or evaluate their improvement more positively. The Hawthorne
Frontiers in Human Neuroscience | www.frontiersin.org 10 May 2016 | Volume 10 | Article 205
Cowley et al. Clinical Trial of Novel Neurofeedback Design
effect can lead to powerful subjective improvements that have
little to do with the specific cognitive training regimen being
studied reflecting motivational factors instead.
If the training caused the decreased symptoms, a higher rate
of learning should have a relationship to the rate of ADHD
symptom decrease under the training. Therefore further analyses
of this trial will examine the relationship between NFB LCs and
the trend of self-reported ADHD symptoms.
Also, the different learning performance levels across the
group should reflect different long term effects, to be measured
during the second phase in a within-subjects analysis. This
contrasts with the effect from the interaction-derived placebo
which is relatively constant for all participants (per group),
and presumably “fades away” quickly after the phase one
intervention. Thus, second phase measurements will be analyzed
with respect to NFB LCs.
Though this is a relatively small study, we believe the analysis
of learning curve questions is a novel and useful contribution. As
noted by Arns (personal communication, 2012):
“In any Neurofeedback study it is very important to track
and have an indication of ‘learning.’ If neurofeedback fails to
demonstrate a clinical effect and there is no indication that
learning actually took place, one can’t draw any conclusions
about neurofeedback. In an analogy, if one employs operant
conditioning to [teach] a rat to press a lever, and the rat does
not learn to press the lever, then it is incorrect to conclude that
‘operant conditioning’ does not work. This means that maybe the
operant conditioning procedure was not implemented effectively.
The same applies to neurofeedback, and this is further illustrated
by the study from Roger DeBeuss. His study employed sub-
optimal parameters e.g., auto-thresholding, ‘game’ feedback and
an ‘unconventional’ training regime (engagement index) making
it likely harder to learn. On the group level they found no effects
of neurofeedback in ADHD, however, when separating learners
from non-learners based on session data they did find an effect.”
Questions of efficacy on the other hand are still a matter of
controversy in the literature. WLC controls are not accepted as
sufficient evidence by some. Double blind control through sham
NFB was not chosen for this study because of the discussed lack
of general understanding of ADHD. This lack, combined with the
extremely contingent nature of NFB which depends heavily on
non-specific aspects of treatment, implies that even if a double
blind RCT showed large effect for NFB the causal mechanisms
would still not be clear. A true resolution to this issue is probably
only possible by running the kind of large sham NFB RCT called
for by others; we choose instead to side-step this debate and focus
on questions of internal comparisons within the method.
There is no blinding in this open-label WLC study, except at
random assignment to groups. Other control paradigms may be
preferred in pharmacological interventions; however there are a
number of arguments in support of WLC, in the context of NFB.
The WLC is a minimum viable control for non-specific effects of
history, maturation, repeated testing, instrument drift, statistical
regression, selection bias, and population inhomogeneity effects
(Mohr et al., 2009). There is no control for non-specific or
placebo effects; however this is still a valid experimental control
design. Among other things, WLC controls for expectation and
attention (Hawthorne) effects, whereby the notion that at some
future point treatment will be provided (and life will improve) is
by itself able to produce improvement. Additionally, longitudinal
(rather than parallel) designs control for maturation, regression
to the mean, instrument drift and practice effects; also time
threats to validity (the same effect occurring 2 years in a row
in different samples rules out external non-seasonal temporal
causes for the effect. Seasonal causes with a WLC group should
be ruled out by staggered application i.e., 2nd treatment starting
in spring).
Technicians can be considered equally skilled and expert. All
began as neurofeedback novices before the trial. Qualification
history was varied. Due to availability, only some began
their training with attendance at the Biofeedback Certification
International Alliance (BCIA)-accredited introductory course at
the Brainclinics education and treatment center, Netherlands.
However, all five then shared 3 months training at UoH
premises, including extensive peer review work, which helped to
disseminate and pool the knowledge across the group.
4.2. Generalizability
External validity asks the question of “generalizability”: to
what populations, settings, treatment variables and measurement
variables can this effect be generalized? While the question of
external validity is never completely answerable, it is of particular
interest for intervention research (Campbell et al., 1966, pp.
171–246). Campbell et al. (1966) note that there is a recurrent
reluctance among researchers to accept Hume’s truism that
induction or generalization is never fully justified logically. While
the problems of internal validity are solvable within the limits
of the logic of probability statistics, the problems of external
validity are not logically solvable in any neat, conclusive way
(Campbell et al., 1966, p. 17). Generalization always involves
extrapolation into a realm not represented in one’s sample. Here,
the issue of sample bias is of importance. If an experimental
study is conducted with voluntary patients from a given district,
they might have characteristics that cause the experimental
treatment to be more effective than it would be in other
populations. However, for ethical reasons, intervention studies
are impossible to conduct without the informed consent of the
research subjects. It is obvious, that a “true” experimental design
is in practice impossible with a fully representative sample of
a given country, let alone all ADHD patients in the world. It
must be emphasized that the results of an experiment “probe”
but do not "prove" a theory. An adequate hypothesis is one
that has repeatedly survived such probing, but it may always
be displaced by a new probe. Many findings in experimental
psychology gain generalizability not through the nature of the
setting in which they occurred, but through their ability to
establish a theory of basic mental processes that are implicated in
many tasks.
4.3. Overall Evidence
After preliminary analysis, the trial did not find evidence
for a transfer of learning that was the intended benefit of
the intervention. Since the intervention’s goal is symptomatic
Frontiers in Human Neuroscience | www.frontiersin.org 11 May 2016 | Volume 10 | Article 205
Cowley et al. Clinical Trial of Novel Neurofeedback Design
improvement outside the laboratory, the results underline how
the transfer-problem limits the potential benefits. Patients did
perceive a reduction of Inattention symptoms over the course of
training, yet this change was not reflected in better performance
in the continuous performance test (T.O.V.A.).
We propose that in the absence of any other evidence, one
should consider these self-report results as due to placebo by
default. The improvement in self-reported symptoms might not
be specifically due to NFB, but due to employment of goal-
oriented attention in general. At present, we can not rule out the
possibility that the individuals would also report improvement
if all other factors were held equal but NFB was swapped for
some other exercise that requires concentration. That is, our
preliminary results did not support the effectiveness of NFB in
alleviating the symptoms of AHDH/ADD.
Nevertheless, keeping in mind the core aim of studying
mechanisms and models of NFB, the CENT trial is on track
to provide the necessary evidence. As Seidman (2006) suggest,
an adequate neuropsychological model of ADHD should utilize
measures from multiple domains to be able to encompass
subtypes and multiple deficits. Combining neuropsychological,
neurophysiological and behavioral measures, we aim toward an
evaluation of structure-function relationships in NFB treatment
for adult ADHD.
AUTHOR CONTRIBUTIONS
BC designed and implemented the study and wrote the draft text;
EH conducted the statistical analyses and contributed to the draft;
KJ and LK contributed to the study design and implementation,
and the draft; CK contributed to the study design and the draft.
ACKNOWLEDGMENTS
The authors wish to thank Svetlana Kirjanen, Mona Moisala,
Marko Repo, Hanna Björkstrand, Tanja Hyttinen, Jari
Torniainen, Teemu Itkonen, Markus Kivikangas for their
part in running the trial; also Laura Hokkanen and Jari Lipsanen
for assistance with the work of EH. Partly funded by Finnish
science agency TEKES, project #440078.
REFERENCES
Arns, M., Feddema, I., and Kenemans, J. L. (2014a). Differential effects of
theta/beta and SMR neurofeedback in ADHD on sleep onset latency. Front.
Hum. Neurosci. 8:1019. doi: 10.3389/fnhum.2014.01019
Arns, M., Gunkelman, J., Breteler, M., and Spronk, D. (2008). EEG phenotypes
predict treatment outcome to stimulants in children with ADHD. J. Integr.
Neurosci. 7, 421–438. doi: 10.1142/S0219635208001897
Arns, M., Heinrich, H., and Strehl, U. (2014b). Evaluation of neurofeedback
in ADHD: the long and winding road. Biol. Psychol. 95, 108–115. doi:
10.1016/j.biopsycho.2013.11.013
Arns, M., Ridder, S. D., Strehl, U., Breteler, M., and Coenen, A. (2009). Efficacy
of neurofeedback treatment in ADHD: the effects on inattention, impulsivity
and hyperactivity: a meta-analyis. Clin. Eeg. Neurosci. 40, 180–189. doi:
10.1177/155005940904000311
Barry, R. J., Clarke, A. R., and Johnstone, S. J. (2003). A review of electrophysiology
in attention-deficit/hyperactivity disorder: I. Qualitative and quantitative
electroencephalography. Clin. Neurophysiol. 114, 171–183. doi: 10.1016/S1388-
2457(02)00362-0
Beck, A., Steer, R., and Brown, G. (1996). BDI-II, Beck Depression Inventory:
Manual, 2 Edn. Boston, MA: Harcourt Brace.
Borkovec, T. D., and Sibrava, N. J. (2005). Problems with the use of placebo
conditions in psychotherapy research, suggested alternatives, and some
strategies for the pursuit of the placebo phenomenon. J. Clin. Psychol. 61,
805–818. doi: 10.1002/jclp.20127
Bresnahan, S. M., Anderson, J. W., and Barry, R. J. (1999). Age-related changes
in quantitative EEG in attention- deficit/hyperactivity disorder. Biol. Psychiatry
46, 1690–1697.
Bresnahan, S. M., and Barry, R. J. (2002). Specificity of quantitative EEG analysis
in adults with attention deficit hyperactivity disorder. Psychiatry Res. 112,
133–144. doi: 10.1016/S0165-1781(02)00190-7
Brown, T. E. (1996). Brown Attention-Deficit Disorder Scales, Manual. San
Antonio: Harcourt Brace.
Brown, T. E. (2009). ADHD Comorbidities : Handbook for ADHD Complications in
Children and Adults. Washington, DC: American Psychiatric Pub.
Calderon, K. S., and Thompson, W. W. (2004). Biofeedback relaxation training:
a rediscovered mind-body tool in public health. Am. J. Health Studies 19,
185–194.
Campbell, D. T., Stanley, J. C., and Gage, N. L. (1966). Experimental and Quasi-
Experimental Designs for Research. Chicago, Il: R. McNally.
Chase, M., and Harper, R. (1971). Somatomotor and visceromotor correlates
of operantly conditioned 1214 c/sec sensorimotor cortical activity.
Electroencephalogr. Clin. Neurophysiol. 31, 85–92.
Clarke, A. R., Barry, R. J., McCarthy, R., and Selikowitz, M. (2001). Excess beta
activity in children with attention-deficit/hyperactivity disorder: an atypical
electrophysiological group. Psychiatry Res. 103, 205–218. doi: 10.1016/S0165-
1781(01)00277-3
Cowley, B., Torniainen, J., Itkonen, T., and Rabenel, R. (forthcoming).
Computer enabled neuroplasticity treatment: a modular, extensible platform
for neurofeedback with lightweight wearable EEG devices. PeerJ Comput. Sci.
Egner, T., and Gruzelier, J. (2004). EEG biofeedback of low beta band
components: frequency-specific effects on variables of attention and event-
related brain potentials. Clin. Neurophysiol. 115, 131–139. doi: 10.1016/S1388-
2457(03)00353-5
Epstein, J. N., Erkanli, A., Conners, C. K., Klaric, J., Costello, J. E., and Angold,
A. (2003). Relations between Continuous Performance Test performance
measures and ADHD behaviors. J. Abnormal Child Psychol. 31, 543–554. doi:
10.1023/A:1025405216339
Gazzaniga, M. S. (2009). The Cognitive Neurosciences, 4 Edn. Cambridge, MA: MIT
Press .
Gevensleben, H., Moll, G. H., Rothenberger, A., and Heinrich, H. (2014).
Neurofeedback in attention-deficit/hyperactivity disorder - different
models, different ways of application. Front. Hum. Neurosci. 8:846. doi:
10.3389/fnhum.2014.00846
Green, C. S., and Bavelier, D. (2012). Learning, attentional control and action video
games. Curr. Biol. 22, R197–R206. doi: 10.1016/j.cub.2012.02.012
Hammond, D. C. (2010). The need for individualization in neurofeedback:
heterogeneity in QEEG patterns associated with diagnoses and symptoms.
Appl. Psychophysiol. Biofeedback 35, 31–36. doi: 10.1007/s10484-009-
9106-1
Heinimaa, M., Salokangas, R. K. R., Ristkari, T., Plathin, M., Huttunen, J., Ilonen,
T., et al. (2003). PROD-screen–a screen for prodromal symptoms of psychosis.
Intern. J. Methods Psychiatric Res. 12, 92–104. doi: 10.1002/mpr.146
Hirschfeld, R. M., Williams, J. B., Spitzer, R. L., Calabrese, J. R., Flynn, L., Keck,
P. E. Jr., et al. (2000). Development and validation of a screening instrument for
bipolar spectrum disorder: the Mood Disorder Questionnaire. Am. J. Psychiatry
157, 1873–1875. doi: 10.1176/appi.ajp.157.11.1873
Howe, R. C., and Sterman, M. B. (1972). Cortical-subcortical EEG correlates
of suppressed motor behavior during sleep and waking in the cat.
Electroencephalogr. Clin. Neurophysiol. 32, 681–695.
Frontiers in Human Neuroscience | www.frontiersin.org 12 May 2016 | Volume 10 | Article 205
Cowley et al. Clinical Trial of Novel Neurofeedback Design
Huang-Pollock, C. L., and Nigg, J. T. (2003). Searching for the attention deficit in
attention deficit hyperactivity disorder: the case of visuospatial orienting. Clin.
Psychol. Rev. 23, 801–830. doi: 10.1016/S0272-7358(03)00073-4
Johnstone, J., Gunkelman, J., and Lunt, J. (2005). Clinical database development:
characterization of EEG phenotypes. Clin. EEG Neurosci. 36, 99–107. doi:
10.1177/155005940503600209
Kessler, R. C., Adler, L., Ames, M., Demler, O., Faraone, S., Hiripi, E., et al. (2005).
The World Health Organization Adult ADHD Self-Report Scale (ASRS): a
short screening scale for use in the general population. Psychol. Med. 35,
245–256. doi: 10.1017/S0033291704002892
Lansbergen, M. M., Arns, M., van Dongen-Boomsma, M., Spronk, D., and
Buitelaar, J. K. (2011). The increase in theta/beta ratio on resting-state EEG
in boys with attention-deficit/hyperactivity disorder is mediated by slow alpha
peak frequency. Progr. Neuro Psychopharmacol. Biol. Psychiatry 35, 47–52. doi:
10.1016/j.pnpbp.2010.08.004
Lazzaro, I., Gordon, E., Li, W., Lim, C. L., Plahn, M., Whitmont, S., et al.
(1999). Simultaneous EEG and EDA measures in adolescent attention deficit
hyperactivity disorder. Intern. J. Psychophysiol. 34, 123–134.
Liebowitz, M. R. (1992). Dissociative experiences scale. Am. J. Psychiatry 149, 719.
Losier, B. J., McGrath, P. J., and Klein, R. M. (1996). Error patterns on the
continuous performance test in non-medicated and medicated samples of
children with and without ADHD: a meta-analytic review. J. Child Psychol.
Psychiatry All. Discipl. 37, 971–987.
Lou, H. C., Henriksen, L., Bruhn, P., Bø rner, H., and Nielsen, J. B. (1989). Striatal
dysfunction in attention deficit and hyperkinetic disorder. Archiv. Neurol. 46,
48–52.
Lubar, J. F., and Shouse, M. N. (1976). EEG and behavioral changes in a
hyperkinetic child concurrent with training of the sensorimotor rhythm (SMR):
a preliminary report. Biofeedback Self Regul. 1, 293–306.
Mann, C. A., Lubar, J. F., Zimmerman, A. W., Miller, C. A., and Muenchen,
R. A. (1992). Quantitative analysis of EEG in boys with attention-deficit-
hyperactivity disorder: controlled study with clinical implications. Pediatric
Neurol. 8 30–36.
Mayer, K., Wyckoff, S. N., Schulz, U., and Strehl, U. (2012). Neurofeedback for
adult attention-deficit/hyperactivity disorder: investigation Of slow cortical
potential neurofeedback preliminary results. J. Neurother. 16, 37–45. doi:
10.1080/10874208.2012.650113
Mayer, K., Wyckoff, S. N., and Strehl, U. (2013). One size fits all? Slow cortical
potentials neurofeedback: a review. J. Attent. Disorders 17, 393–409. doi:
10.1177/1087054712468053
Mayer, K., Wyckoff, S. N., and Strehl, U. (2015). Underarousal in Adult ADHD:
how are peripheral and cortical arousal related? Clin. EEG Neurosci. doi:
10.1177/1550059415577544. [Epub ahead of print].
Micoulaud-Franchi, J.-A., Geoffroy, P. A., Fond, G., Lopez, R., Bioulac, S., and
Philip, P. (2014). EEG neurofeedback treatments in children with ADHD: an
updated meta-analysis of randomized controlled trials. Front. Hum. Neurosci.
8:906. doi: 10.3389/fnhum.2014.00906
Mohr, D. C., Spring, B., Freedland, K. E., Beckner, V., Arean, P., Hollon, S. D.,
et al. (2009). The selection and design of control conditions for randomized
controlled trials of psychological interventions. Psychother. Psychosomat. 78,
275–284. doi: 10.1159/000228248
Monastra, V. J., Lynn, S., Linden, M., Lubar, J. F., Gruzelier, J., and LaVaque,
T. J. (2005). Electroencephalographic biofeedback in the treatment of attention-
deficit/hyperactivity disorder. Appl. Psychophysiol. Biofeedback 30, 95–114. doi:
10.1007/s10484-005-4305-x
Nigg, J. T. (2005). Neuropsychologic theory and findings in attention-
deficit/hyperactivity disorder: the state of the field and salient
challenges for the coming decade. Biol. Psychiatry 57, 1424–1435. doi:
10.1016/j.biopsych.2004.11.011
Olbrich, S., Sander, C., Minkwitz, J., Chittka, T., Mergl, R., Hegerl, U., et al. (2012).
EEG vigilance regulation patterns and their discriminative power to separate
patients with major depression from healthy controls. Neuropsychobiology 65,
188–194. doi: 10.1159/000337000
Pennington, B. F. (2005). Toward a new neuropsychological model of attention-
deficit/hyperactivity disorder: subtypes and multiple deficits. Biol. Psychiatry
57, 1221–1223. doi: 10.1016/j.biopsych.2005.03.010
Posner, M. I., and Petersen, S. E. (1990). The attention system of the human brain.
Ann. Rev. Neurosci. 13, 25–42.
Rabbitt, P. (1997). Methodology of Frontal and Executive Function. Hove, UK:
Psychology Press.
Roberts, C., and Torgerson, D. (1998). Understanding controlled trials:
randomisation methods in controlled trials. BMJ 317, 1301–1310.
Rossiter, T. (2004a). The effectiveness of neurofeedback and stimulant drugs in
treating AD/HD: Part I. Review of methodological issues. Appl. Psychophysiol.
Biofeedback 29, 95–112. doi: 10.1023/B:APBI.0000026636.13180.b6
Rossiter, T. (2004b). The effectiveness of neurofeedback and stimulant drugs
in treating AD/HD: Part II. Replication. Appl. Psychophysiol. Biofeedback 29,
233–243. doi: 10.1007/s10484-004-0383-4
Satterfield, J. H., Cantwell, D. P., and Satterfield, B. T. (1974). Pathophysiology of
the hyperactive child syndrome. Arch. Gen. Psychiatry 31, 839–844.
Saunders, J. B., Aasland, O. G., Babor, T. F., de la Fuente, J. R., and Grant, M.
(1993). Development of the alcohol use disorders identification test (AUDIT):
WHO collaborative project on early detection of persons with harmful alcohol
consumption–II. Addiction (Abingdon, England) 88, 791–804.
Scott, N. W., McPherson, G. C., Ramsay, C. R., and Campbell, M. K. (2002). The
method of minimization for allocation to clinical trials. Controll. Clin. Trials 23,
662–674. doi: 10.1016/S0197-2456(02)00242-8
Seidman, L. J. (2006). Neuropsychological functioning in people with
ADHD across the lifespan. Clin. Psychol. Rev. 26, 466–485. doi:
10.1016/j.cpr.2006.01.004
Sergeant, J., Oosterlaan, J., and van der Meere, J. (1999). “Information processing
and energetic factors in attention-deficit/hyperactivity disorder,” in Handbook
of Disruptive Behavior Disorders SE - 4, eds H. Quay and A. Hogan (New York,
NY: Kluwer Academic/Plenum Publishers), 75–104.
Sergeant, J. A., Geurts, H., Huijbregts, S., Scheres, A., and Oosterlaan, J. (2003).
The top and the bottom of ADHD: a neuropsychological perspective. Neurosci.
Biobehav. Rev. 27, 583–592. doi: 10.1016/j.neubiorev.2003.08.004
Snyder, S. M., and Hall, J. R. (2006). A meta-analysis of quantitative EEG power
associated with attention-deficit hyperactivity disorder. J. Clin. Neurophysiol.
23, 440–455. doi: 10.1097/01.wnp.0000221363.12503.78
Sonuga-Barke, E. J. S. (2005). Causal models of attention-deficit/hyperactivity
disorder: from common simple deficits to multiple developmental pathways.
Biol. Psychiatry 57, 1231–1238. doi: 10.1016/j.biopsych.2004.09.008
Spitzer, R. L., Kroenke, K., Williams, J. B. W., and Löwe, B. (2006). A brief measure
for assessing generalized anxiety disorder: the GAD-7. Archiv. Int. Med. 166,
1092–1097. doi: 10.1001/archinte.166.10.1092
Vernon, D., Egner, T., Cooper, N., Compton, T., Neilands, C., Sheri, A., et al.
(2003). The effect of training distinct neurofeedback protocols on aspects of
cognitive performance. Inter. J. Psychophysiol. 47, 75–85. doi: 10.1016/S0167-
8760(02)00091-0
Wechsler, D. (2008). Wechsler Adult Intelligence Scale: Technical and Interpretive
Manual, 4th Edn. San Antonio, TX: Pearson.
Zametkin, A. J., Nordahl, T. E., Gross, M., King, A. C., Semple, W. E., Rumsey,
J., et al. (1990). Cerebral glucose metabolism in adults with hyperactivity of
childhood onset. New England J. Med. 323, 1361–1366.
Zuberer, A., Brandeis, D., and Drechsler, R. (2015). Are treatment effects of
neurofeedback training in children with ADHD related to the successful
regulation of brain activity? A review on the learning of regulation of brain
activity and a contribution to the discussion on specificity. Front. Hum.
Neurosci. 9:135. doi: 10.3389/fnhum.2015.00135
Zylowska, L., Ackerman, D. L., Yang, M. H., Futrell, J. L., Horton, N. L.,
Hale, T. S., et al. (2008). Mindfulness meditation training in adults and
adolescents with ADHD: a feasibility study. J. Att. Disorders 11, 737–746. doi:
10.1177/1087054707308502
Conflict of Interest Statement: The authors declare that the research was
conducted in the absence of any commercial or financial relationships that could
be construed as a potential conflict of interest.
Copyright © 2016 Cowley, Holmström, Juurmaa, Kovarskis and Krause. This is an
open-access article distributed under the terms of the Creative Commons Attribution
License (CC BY). The use, distribution or reproduction in other forums is permitted,
provided the original author(s) or licensor are credited and that the original
publication in this journal is cited, in accordance with accepted academic practice.
No use, distribution or reproduction is permitted which does not comply with these
terms.
Frontiers in Human Neuroscience | www.frontiersin.org 13 May 2016 | Volume 10 | Article 205
... Twelve documents explored Neurofeedback as an intervention for ADHD. These include randomised controlled trials [229][230][231], individual interventions [232,233], case-control studies [234,235], a single case study [236], and treatment guidance [118, [237][238][239]. Neurofeedback (NF) treatment models focus heavily on neurocognitive deficits as being the origin of ADHD behaviours. ...
... Theta/beta and theta/alpha waveform ratios (TBR) are considered a measure of differences in excess, slow-wave activity and epileptiform spike and wave activity [240], interpreted as abnormal brain processes indicating cortical under arousal, insufficient inhibitory control, and maturational delay in ADHD [241]; however recent studies have challenged TBR as a marker for ADHD diagnosis [235]. Sensory-motor rhythm (SMR) or low beta waveform ratios are thought to indicate cortical hypo-arousal, interpreted as deficiencies in the early stages of information processing [230]. Decreased contingent negative variation (CNV), a steady, slow, negative-going waveform associated with cognitive energy in anticipation of task performance, is considered indicative of dysfunctional regulation of energetical resources in ADHD [234]. ...
... Rather than simply improving neuropsychological deficits, it is thought that NF may be used as a tool for enhancing or optimising specific cognitive or attentional states [246,247]. This model recognises the bio-psycho-social model of neurodevelopmental disorders, characterising ADHD as impairments in attention, executive functions and self-regulation [229,230]. In this model, self-regulation, or neuro-regulation, is defined as explicit learning of controlled cognitive processes of cortical regulation evidenced by normalised shifts in EEG amplitudes [242,248,249]. ...
Article
Full-text available
Psychological theory and interpretation of research are key elements influencing clinical treatment development and design in Attention Deficit Hyperactivity Disorder (ADHD). Research-based treatment recommendations primarily support Cognitive Behavioural Therapy (CBT), an extension of the cognitive behavioural theory, which promotes a deficit-focused characterisation of ADHD and prioritises symptom reduction and cognitive control of self-regulation as treatment outcomes. A wide variety of approaches have developed to improve ADHD outcomes in adults, and this review aimed to map the theoretical foundations of treatment design to understand their impact. A scoping review and analysis were performed on 221 documents to compare the theoretical influences in research, treatment approach, and theoretical citations. Results showed that despite variation in the application, current treatments characterise ADHD from a single paradigm of cognitive behavioural theory. A single theoretical perspective is limiting research for effective treatments for ADHD to address ongoing issues such as accommodating context variability and heterogeneity. Research into alternative theoretical characterisations of ADHD is recommended to provide treatment design opportunities to better understand and address symptoms.
... Attention-deficit/hyperactivity disorder (ADHD) is a highly heritable condition with symptoms which begin in childhood and often continue into adulthood, and is currently estimated to affect 2.5-3.4% of the adult population (Kessler et al., 2006). Here, we report exploratory analyses of a randomised controlled clinical trial of neurofeedback treatment in adults with ADHD (registered trial ISRCTN13915109; Cowley et al., 2016), focusing on the understudied area of neurofeedback learning, and its implications for treatment. ...
... The clinical trial (Cowley et al., 2016) analysed in this study followed one of the most established approaches: patients must train to regulate the power (decibels) in specific bands of the electroencephalograph (EEG) frequency spectrum-e.g., thetabeta ratio (TBR) and sensorimotor rhythm (SMR) regimes target theta (4-8 Hz) and beta (13-30 Hz) bands. Other common approaches focus not on frequency bands but, e.g., on "slow" potentials of the EEG waveform (Strehl et al., 2017), or peripheral nervous system signals (Calderon and Thompson, 2004). ...
... We here report a novel, exploratory analysis of a clinical trial of NFB for adults with ADHD, focused on their learning and its predictors, correlates, and outcome effects. Our original research (e.g., sample size based on statistical power analysis, as detailed in section 2.5 of Cowley et al., 2016) was designed to test the efficacy of NFB training in the entire training group. We found no such general effect; however, (unlike data in most RCTs), our data includes detailed measurements from each training session, allowing us to distinguish between two groups of patients based on training progress. ...
Article
Full-text available
Neurofeedback for attention deficit/hyperactivity disorder (ADHD) has long been studied as an alternative to medication, promising non-invasive treatment with minimal side-effects and sustained outcome. However, debate continues over the efficacy of neurofeedback, partly because existing evidence for efficacy is mixed and often non-specific, with unclear relationships between prognostic variables, patient performance when learning to self-regulate, and treatment outcomes. We report an extensive analysis on the understudied area of neurofeedback learning. Our data comes from a randomised controlled clinical trial in adults with ADHD (registered trial ISRCTN13915109; N = 23; 13:10 female:male; age 25–57). Patients were treated with either theta-beta ratio or sensorimotor-rhythm regimes for 40 one-hour sessions. We classify 11 learners vs 12 non-learners by the significance of random slopes in a linear mixed growth-curve model. We then analyse the predictors, outcomes, and processes of learners vs non-learners, using these groups as mutual controls. Significant predictive relationships were found in anxiety disorder (GAD), dissociative experience (DES), and behavioural inhibition (BIS) scores obtained during screening. Low DES, but high GAD and BIS, predicted positive learning. Patterns of behavioural outcomes from Test Of Variables of Attention, and symptoms from adult ADHD Self-Report Scale, suggested that learning itself is not required for positive outcomes. Finally, the learning process was analysed using structural-equations modelling with continuous-time data, estimating the short-term and sustained impact of each session on learning. A key finding is that our results support the conceptualisation of neurofeedback learning as skill acquisition, and not merely operant conditioning as originally proposed in the literature.
... Twelve documents explored Neurofeedback as an intervention for ADHD. These include randomised controlled trials [217][218][219], individual interventions [220,221], case-control studies [222,223], a single case study [224], and treatment guidance [104,[225][226][227]. Neurofeedback (NF) treatment models focus heavily on neurocognitive deficits as being the origin of ADHD behaviours. The research uses Electroencephalography (EEG) measures to study the correspondences between intracranial electrical currents and responding voltages on the scalp. ...
... Theta/beta and theta/alpha waveform ratios (TBR) are considered a measure of differences in excess, slow-wave activity and epileptiform spike and wave activity [228], interpreted as abnormal brain processes indicating cortical under arousal, insufficient inhibitory control, and maturational delay in ADHD [229]; however recent studies have challenged TBR as a marker for ADHD diagnosis [223]. Sensory-motor rhythm (SMR) or low beta waveform ratios are thought to indicate cortical hypo-arousal, interpreted as deficiencies in the early stages of information processing [218]. Decreased contingent negative variation (CNV), a steady, slow, negative-going waveform associated with cognitive energy in anticipation of task performance, is considered indicative of dysfunctional regulation of energetical resources in ADHD [222]. ...
... Rather than simply improving neuropsychological deficits, it is thought that NF may be used as a tool for enhancing or optimising specific cognitive or attentional states [234,235]. This model recognises the bio-psycho-social model of neurodevelopmental disorders, characterising ADHD as impairments in attention, executive functions and self-regulation [217,218]. In this model, self-regulation, or neuroregulation, is defined as explicit learning of controlled cognitive processes of cortical regulation evidenced by normalised shifts in EEG amplitudes [230,236,237]. ...
Preprint
Psychological theory and interpretation of research are key elements influencing clinical treatment development and design in Attention Deficit Hyperactivity Disorder (ADHD). Research-based treatment recommendations primarily support Cognitive Behavioural Therapy (CBT), an extension of the cognitive behavioural theory, which promotes a deficit-focused characterisation of ADHD and prioritises symptom reduction and cognitive control of self-regulation as treatment outcomes. A wide variety of approaches have developed to improve ADHD outcomes in adults, and this review aimed to map the theoretical foundations of treatment design to understand their impact. A scoping review and analysis were performed on 221 documents to compare the theoretical influences in research, treatment approach, and theoretical citations. Results showed that despite variation in the application, current treatments characterise ADHD from a single paradigm of cognitive behavioural theory. A single theoretical perspective is limiting research for effective treatments for ADHD to address ongoing issues such as accommodating context variability and heterogeneity. Research into alternative theoretical characterisations of ADHD is recommended to provide treatment design opportunities to better understand and address symptoms.
... 15 RCTs used non-pharmacological interventions. Five of which were targeting ADHD symptoms [47][48][49][50][51], two targeting sleep symptoms [52,53], and eight targeting both ADHD and sleep symptoms [35,36,[54][55][56][57][58][59]. ...
... The most commonly employed non-pharmacological interventions were behavioural therapy/parent training. For the treatment of ADHD, several novel interventions, including trigeminal nerve stimulation [50], slow oscillating transcranial direct current [51], and neurofeedback [47,49] were used. All non-pharmacological interventions are listed in Table 3. [43] translated the SDSC into Persian. ...
Article
Full-text available
Sleep disturbances are highly prevalent among children with ADHD. Yet, diagnostic and treatment regimens are primarily focused on daytime symptomatology. The goals of this scoping review are to 1) identify interventional ADHD RCTs that have used sleep as an outcome measure, 2) describe and assess the validity of tools utilized to measure sleep-specific outcomes. 40/71 RCTs used sleep as a primary outcome. Actigraphy (n=18) and sleep log/diary (n=16) were the most common tools to measure sleep, followed by Children’s Sleep Habits Questionnaire (n=13), and polysomnography (n=10). Sleep was a secondary outcome in 31 RCTs. Polysomnography and actigraphy used a heterogeneous spectrum of sleep-related variables and technical algorithms, respectively. 19/23 sleep questionnaires were validated covering a spectrum of sleep-related domains. Despite the intrinsic nature of sleep disturbances in ADHD, the number of RCTs measuring sleep-specific outcomes is limited and tools to measure outcomes are not standardized. Given the potential adverse effects of ADHD medications on sleep, sleep should be included as a core outcome measure in future clinical trials.
... Most studies are based on electroencephalogram (EEG) recordings and apply neurofeedback in clinical contexts, exploring its potential as a treatment for psychopathological syndromes. Cowley et al. (2016) carried out a controlled clinical trial of neurofeedback therapeutic intervention in ADHD for adults. The EEG device used was Enobio from the company Neuroelectrics SL. ...
... The measurement consisted of visual TOVA CPT with EEG recording, gathered as part of a larger project detailed in Cowley et al. (44). An ethical approval was obtained from the Ethical Committee of the Hospital District of Helsinki and Uusimaa. ...
Preprint
Full-text available
Attention-deficit hyperactivity disorder (ADHD) in adults is understudied, especially regarding neural mechanisms such as oscillatory control of attention sampling. We report an EEG study of such cortical oscillations, in ADHD-diagnosed adults taking a continuous performance test that measures the ability to sustain attention and inhibit impulsivity for a prolonged period of time. We recorded 53 adults (28f, 25m, aged 18-60), and 18 matched healthy controls, using 128-channel EEG. We analysed features with established links to neural correlates of attention: event-related (de)synchronisation (ERS), alpha and theta frequency band activation, and stimulus-locking indices; in frontal and parietal scalp regions. Test performance distinguished healthy controls from ADHD adults. The ADHD group manifested significantly less parietal 4 Hz theta ERS during correct inhibition trials, in addition to having greater sensitivity to targets in stimulus-locking measures. Our results suggest that ADHD adults have impaired attention sampling in relational categorisation tasks.
... In terms of the psychological approach reported in the studies included here; the most common was CBT (including both group and individual sessions) with 23 studies (Anastopoulos et al. 2018;Bramham et al. 2009;Cherkasova et al. 2016;Cole et al. 2016;Corbisero et al. 2018;Dittner et al. 2018;Emilsson et al. 2011;Hiltunen et al. 2014;Hirvikoski et al. 2015;LaCount et al. 2015;Nasri et al. 2017;Pettersson et al. 2017;Ramsay and Rostain 2011;Rostain and Ramsay 2006;Safren et al. 2005aSafren et al. , 2005bSafren et al. , 2010Salakari et al. 2010;Stern et al. 2014;Vidal et al. 2013;Virta et al. 2008;Weiss et al. 2012;Young et al. 2015Young et al. , 2017 followed by Mindfulness approaches with 9 studies (Bachmann et al. 2018;Bueno et al. 2015;Edel et al. 2017;Gu et al. 2018;Hepark et al. 2015;Hoxhaj et al. 2018;Janssen et al. 2018;Mitchell et al. 2017;Schoenberg et al. 2014) and Dialectical Behavioral Therapy (DBT) with 8 studies (Cole et al. 2016;Edel et al. 2017;Fleming et al. 2015;Hesslinger et al. 2002;Hirvikoski et al. 2011;Morgensterns et al. 2016;Nasri et al. 2017;Philipsen et al. 2007). Neurofeedback (NFB) with 5 studies (Cowley et al. 2016;Mayer et al. 2016;Mayer et al. 2012;Schönenberg et al. 2017;Zilverstand et al. 2017), Psychoeducation approaches with 3 studies Salomone et al. 2015;Vidal et al. 2013), Hypnotherapy with 2 (Hiltunen et al. 2014;Virta et al. 2015), Metacognitive therapy with 2 (Solanto et al. 2010;Solanto, Solanto et al. 2008), CogMed training with 2 (Dentz et al. 2017;Mawjee et al. 2015), Cognitive remediation therapy with 1 (Stevenson et al. 2002), Goal management training with 1 (In de Braek et al. 2017), Psychotherapy with 1 (Philipsen et al. 2007), Self-directed psychosocial intervention with 1 (Stevenson et al. 2003) and, Stress management training with 1 study (Langer et al. 2013). 1 ...
Article
Full-text available
Attention deficit hyperactivity disorder (ADHD) is a common neurodevelopmental disorder, characterized by symptoms of inattention, hyperactivity and or impulsivity. First line treatment is medication; however, medication alone may not provide sufficient functional improvement for some patients, or be universally tolerated. A recent surge in research to treat ADHD using non-pharmacological interventions demands a comprehensive, systematic review of the literature. The aim of this review was to examine the evidence base for psychological treatments for ADHD management in adulthood. A systematic search of PsycINFO, MEDLINE, CINAHL, AMED, PubMed, and EMBASE was undertaken until January 2019 for peer-reviewed articles exploring psychological interventions for adults (18 years with no upper limit) diagnosed with ADHD. A total of 53 papers were identified for inclusion. Collectively, 92% of studies (employing various non-pharmacological interventions) found a variant of significant positive effect on either primary or secondary outcomes associated with ADHD. The strongest empirical support derived from Cognitive Behavioral Therapy interventions. In addition, findings indicated support for the effectiveness of Mindfulness, Dialectical Behavior Therapy and Neurofeedback. Other types of interventions also demonstrated effectiveness; however, support was limited due to lack of available research and methodological rigor. Psychological interventions should be considered a valid and useful addition to clinical practice. Implications and areas for future research are discussed.
Article
Full-text available
Attention-deficit/hyperactivity disorder (ADHD) in adults is understudied, especially regarding neural mechanisms such as oscillatory control of attention sampling. We report an electroencephalography (EEG) study of such cortical mechanisms, in ADHD-diagnosed adults during administration of Test of Variables of Attention (TOVA), a gold-standard continuous performance test for ADHD that measures the ability to sustain attention and inhibit impulsivity. We recorded 53 adults (28 female, 25 male, aged 18-60), and 18 matched healthy controls, using 128-channel EEG. We analyzed sensor-space features established as neural correlates of attention: timing-sensitivity and phase-synchrony of response activations, and event-related (de)synchronization (ERS/D) of α and θ frequency band activity; in frontal and parietal scalp regions. TOVA test performance significantly distinguished ADHD adults from neurotypical controls, in commission errors, response time variability (RTV) and d' (response sensitivity). The ADHD group showed significantly weaker target-locked and response-locked amplitudes, that were strongly right-lateralized at the N2 wave, and weaker phase synchrony (longer reset poststimulus). They also manifested significantly less parietal prestimulus 8-Hz θ ERS, less frontal and parietal poststimulus 4-Hz θ ERS, and more frontal and parietal prestimulus α ERS during correct trials. These differences may reflect excessive modulation of endogenous activity by strong entrainment to stimulus (α), combined with deficient modulation by neural entrainment to task (θ), which in TOVA involves monitoring stimulus spatial location (not predicted occurrence onset which is regular and task-irrelevant). Building on the hypotheses of θ coding for relational structure and rhythmic attention sampling, our results suggest that ADHD adults have impaired attention sampling in relational categorization tasks.
Article
Attention deficit hyperactivity disorder (ADHD) is a neurobiological condition that appears during an individual's childhood and may follow her/him for life. The research objective was to understand better how and which computer technologies have been applied to support ADHD diagnosis and treatment. The research used the systematic literature review method: a rigorous, verifiable, and repeatable approach that follows well-defined steps. Six well-known academic data sources have been consulted, including search engines and bibliographic databases, from technology and health care areas. After a rigorous research protocol, 1,239 articles were analyzed. For the diagnosis, the use of machine learning techniques was verified in 61 percent of the articles. Neurofeedback was ranked second with 9.3 percent participation, followed by serious games and eye tracking with 5.6 percent each. For the treatment, neurofeedback was present in 50 percent of the articles, whereas some studies combined both approaches, accounting for 31 percent of the total. Nine percent of the articles reported remote assistance technology, whereas another 9 percent have used virtual reality. By highlighting the leading computer technologies used, their applications, results, and challenges, this literature review breaks ground for further investigations. Moreover, the study highlighted the lack of consensus on ADHD biomarkers. The approaches using machine learning call attention to the probable occurrence of overfitting in several studies, thus demonstrating limitations of this technology on small-sized bases. This research also presented the convergence of evidence from different studies on the persistence of long-term effects of using neurofeedback in treating ADHD.
Article
Background: Adult ADHD is common, highly comorbid, and restricts daily functioning. However, only a minority of patients receive appropriate treatment. Aims: Primary objective: To identify psychological interventions that diminish inattention symptoms in adults with ADHD. Secondary objective: To evaluate feasibility aspects. Methods: A search in PubMed, Cochrane Library, PsycInfo, PSYNDEX, and Eric was conducted. Interventions were grouped in: ADHD coaching, neuro feedback, cognitive training, psychoeducation, and behavioral therapy. Inattention symptoms were evaluated using standardized mean differences. Study quality was rated with the Cochrane Risk of Bias tool. Feasibility was assessed by number and time of sessions, setting, and qualification of the provider. Results: A total of N = 2229 results were identified, 19 randomized controlled studies were included in the analysis. Behavioral therapy showed effects on inattention symptoms [SMD: 0.44-1.71] when compared to inactive controls. In terms of feasibility, longer interventions did not outperform shorter ones and individual sessions were not superior to group sessions. No effects were given for neuro feedback, cognitive training, and psychoeducation in comparison to controls. Conclusion: For adults with ADHD behavioral therapy seems an effective intervention to reduce inattention symptoms. In terms of feasibility, brief interventions may be valuable for a primary care setting.
Article
Full-text available
While issues of efficacy and specificity are crucial for the future of neurofeedback training, there may be alternative designs and control analyses to circumvent the methodological and ethical problems associated with double-blind placebo studies. Surprisingly, most NF studies do not report the most immediate result of their NF training, i.e., whether or not children with ADHD gain control over their brain activity during the training sessions. For the investigation of specificity, however, it seems essential to analyze the learning and adaptation processes that take place in the course of the training and to relate improvements in self-regulated brain activity across training sessions to behavioral, neuropsychological and electrophysiological outcomes. To this aim, a review of studies on neurofeedback training with ADHD patients which include the analysis of learning across training sessions or relate training performance to outcome is presented. Methods on how to evaluate and quantify learning of EEG regulation over time are discussed. "Non-learning" has been reported in a small number of ADHD-studies, but has not been a focus of general methodological discussion so far. For this reason, selected results from the brain-computer interface (BCI) research on the so-called "brain-computer illiteracy", the inability to gain control over one's brain activity, are also included. It is concluded that in the discussion on specificity, more attention should be devoted to the analysis of EEG regulation performance in the course of the training and its impact on clinical outcome. It is necessary to improve the knowledge on characteristic cross-session and within-session learning trajectories in ADHD and to provide the best conditions for learning.
Article
Full-text available
In children and adults with attention deficit/hyperactivity disorder (ADHD), a general slowing of spontaneous electroencephalographic (EEG) brain activity and a decrease of event-related potential amplitudes such as the contingent negative variation (CNV) are observed. Additionally, some studies have reported decreased skin conductance level (SCL) in this clinical population leading to the hypothesis of a peripheral hypoarousal, which may be a target of biofeedback treatment in addition to or instead of neurofeedback. To our knowledge, the relationship between SCL and CNV has not been simultaneously investigated in one experiment. Using the theoretical background of the hypoarousal model, this article aims to gain more insight into the differences and correlations of cortical (CNV) and peripheral (SCL) arousal in adults with ADHD. A sample of 23 adults with ADHD and 22 healthy controls underwent an auditory Go-NoGo task with simultaneous 22-channel EEG and SCL recordings. Reaction time (RT) and reaction time variability (RTV) were also measured to assess task performance. Significantly decreased CNV amplitude and significantly higher RTV were observed in the ADHD group, reflecting cortical underarousal and problems with sustained attention. No significant correlation between peripheral underarousal and cortical underarousal was observed in the ADHD group or the control group. The observed cortical underarousal reflected in the decreased CNV supports the notion of a reduced CNV amplitude as a possible biomarker for ADHD. However, the connection between cortical and peripheral arousal is not as clear as is suggested in previous research investigating both separately. Implications of these results for new treatment options for ADHD such as biofeedback are discussed. © EEG and Clinical Neuroscience Society (ECNS) 2015.
Article
Full-text available
The main purpose of randomisation is to avoid bias by distributing the characteristics of patients that may influence outcome randomly between treatment groups so that any difference in outcome can be explained only by treatment. These characteristics might be demographic ones such age or prognostic factors such as clinical history or disease severity. For example menopausal status may influence outcome of treatment for breast cancer.The most elementary form of randomisation is, in the case of two treatments, equivalent to allocating treatment by tossing a coin. Lists for allocating patients by simple randomisation may be constructed with tables of random numbers or random functions on pocket calculators or statistical software. Treatments may then be allocated to patients in sequence using numbered opaque envelopes containing treatment allocations or remotely by phone.While such simple randomisation will on average allocate equal numbers to each arm, even in quite large trials simple randomisation can result in groups of different sizes. In small trials there may be substantial differences in group sizes that will reduce the precision of estimates of the difference in treatment effect and hence efficiency of the study.One method to prevent unequal treatment group sizes is block randomisation. This guarantees that at no time will the imbalance be large and at certain points the numbers of participants in each group will be equal. If, for example, we choose blocks of four, there are six sequences to which we can allocate treatments A and B: AABB, ABAB, ABBA, BAAB, BABA, and BBAA. One of the six arrangements is selected randomly and then four participants assigned accordingly. The process is then repeated as many times as is needed for the required sample size.With simple randomisation or block randomisation substantial imbalance in prognostic characteristics can, nevertheless, arise by chance and can bias the analysis of outcome. One method to achieve balance between groups for a prognostic variable is stratified randomisation, in which separate randomisation lists are used for each prognostic subgroup. For example, in a study of alternative treatments for breast cancer it would be advantageous to stratify on menopausal status. Separate randomisation lists would be prepared for each stratum using a block randomisation. It should be noted that using simple randomisation with each stratum would defeat the purpose of stratification as the resulting randomisation would be no different from simple randomisation. A standard practice in multicentre trials is to stratify randomisation by treatment centre.Stratification may be extended to two or more factors, although the number of separate randomisation lists rapidly becomes very large. For example, if one was to stratify on three prognostic variables, with each having just two levels, eight separate randomisation lists would be required for each combination of factors. In practice therefore it is rarely feasible to go beyond two factors.Stratified randomisation makes the process more elaborate and brings with it the risk of mistakes that the simpler methods might prevent. If there is uncertainty about which patient characteristics may influence the outcome of treatment it may be prudent to proceed without stratification.1An alternative method of obtaining treatment groups that are comparable in prognostic variables is minimisation. This achieves balance on a set of prognostic factors, although not for each combination. Even in small trials it will provide groups that are very similar on several prognostic factors. For all levels of each prognostic factor on which the investigator wishes to maintain balance, a running total is kept of how many patients have been assigned to each treatment. At the start of the trial treatment is randomly allocated to the first patient. Subsequent patients are assigned using a randomisation weighted towards the group to which assignment would minimise the imbalance. After each patient is entered the relevant totals for each factor are updated ready for the patient. Details with examples of the minimisation method are discussed by Pocock1 and Altman.2Whatever method of allocation is used, the process of allocation needs to be done in such as way that the randomisation cannot be deciphered, a topic discussed in a forthcoming note.References↵Pocock SJ. Clinical trials: a practical approach. Chicester: John Wiley, 1983.
Article
Full-text available
Recent studies suggest a role for sleep and sleep problems in the etiology of attention deficit hyperactivity disorder (ADHD) and a recent model about the working mechanism of sensori-motor rhythm (SMR) neurofeedback, proposed that this intervention normalizes sleep and thus improves ADHD symptoms such as inattention and hyperactivity/impulsivity. In this study we compared adult ADHD patients (N = 19) to a control group (N = 28) and investigated if differences existed in sleep parameters such as Sleep Onset Latency (SOL), Sleep Duration (DUR) and overall reported sleep problems (PSQI) and if there is an association between sleep-parameters and ADHD symptoms. Secondly, in 37 ADHD patients we investigated the effects of SMR and Theta/Beta (TBR) neurofeedback on ADHD symptoms and sleep parameters and if these sleep parameters may mediate treatment outcome to SMR and TBR neurofeedback. In this study we found a clear continuous relationship between self-reported sleep problems (PSQI) and inattention in adults with- and without-ADHD. TBR neurofeedback resulted in a small reduction of SOL, this change in SOL did not correlate with the change in ADHD symptoms and the reduction in SOL only happened in the last half of treatment, suggesting this is an effect of symptom improvement not specifically related to TBR neurofeedback. SMR neurofeedback specifically reduced the SOL and PSQI score, and the change in SOL and change in PSQI correlated strongly with the change in inattention, and the reduction in SOL was achieved in the first half of treatment, suggesting the reduction in SOL mediated treatment response to SMR neurofeedback. Clinically, TBR and SMR neurofeedback had similar effects on symptom reduction in ADHD (inattention and hyperactivity/impulsivity). These results suggest differential effects and different working mechanisms for TBR and SMR neurofeedback in the treatment of ADHD.
Book
This edited text comprehensively describes the multiple ways in which other psychiatric and learning disorders complicate ADHD in both children and adults. More than 30 leading clinician-researchers provide information on ADHD and its full range of comorbidities including anxiety disorders, mood disorders, learning disorders, substance use disorders, sleep disorders, OCD, autism spectrum disorders, oppositionality and aggression, Tourette syndrome and developmental coordination disorder.