Addressing Efficacy in a Clinical Setting

Evaluating 19-Channel Z-score Neurofeedback:

Addressing Efficacy in a Clinical Setting

Submitted by

Nancy L. Wigton

A Dissertation Presented in Partial Fulfillment

of the Requirements for the Degree

Doctorate of Philosophy

Grand Canyon University

Phoenix, Arizona

May 15, 2014

All rights reserved

INFORMATION TO ALL USERS The quality of this reproduction is dependent upon the quality of the copy submitted.

In the unlikely event that the author did not send a complete manuscript and there are missing pages, these will be noted. Also, if material had to be removed,

a note will indicate the deletion.

Microform Edition © ProQuest LLC. All rights reserved. This work is protected against

unauthorized copying under Title 17, United States Code

ProQuest LLC. 789 East Eisenhower Parkway

P.O. Box 1346 Ann Arbor, MI 48106 – 1346

UMI 3625170 Published by ProQuest LLC (2014). Copyright in the Dissertation held by the Author.

UMI Number: 3625170

© by Nancy L. Wigton, 2014

All rights reserved.

Abstract

Neurofeedback (NF) is gaining recognition as an evidence-based intervention grounded

in learning theory, and 19-channel z-score neurofeedback (19ZNF) is a new NF model.

Peer-reviewed literature is lacking regarding empirical-based evaluation of 19ZNF. The

purpose of this quantitative research study was to evaluate the efficacy of 19ZNF, in a

clinical setting, using archival data from a Southwest NF practice, with a retrospective

one-group pretest-posttest design. Each of the outcome measures framed a group such

that 19ZNF was evaluated, as it relates to the particular neuropsychological constructs of

attention (n = 10), behavior (n = 14), executive function (n = 12), as well as

electrocortical functioning (n = 21). The research questions asked if 19ZNF improves

these constructs. One-tailed t tests performed, compared pre-post scores for included

clinical assessment scales, and selected quantitative electroencephalographic (QEEG)

metrics. For all pre-post comparisons, the direction of change was in the predicted

direction. Moreover, for all outcome measures, the group means were beyond the

clinically significant threshold before 19ZNF, and no longer clinically significant after

19ZNF. All differences were statistically significant, with results ranging from p = .000

to p = .008; and effect sizes ranging from 1.29 to 3.42. Results suggest 19ZNF improved

attention, behavior, executive function, and electrocortical function. This study provides

beginning evidence of 19ZNF’s efficacy, adds to what is known about 19ZNF, and offers

an innovative approach for using QEEG metrics as outcome measures. These results may

lead to a greater acceptance of 19ZNF, as well as foster needed additional scientific

research.

Keywords: Neurofeedback, QEEG, z-score neurofeedback, 19ZNF, EEG biofeedback

v

Dedication

This dissertation is dedicated to my Lord and Savior, Jesus. From my first

thoughts of considering a doctoral program being divinely inspired and directed, through

to the last step I will take across a graduation stage, the Father, Son, and Holy Spirit are

always the center point, the anchor. To that end, three Bible passages capture the

experience of my journey.

The way of God is perfect, the Lord’s word has stood the test; He is the shield of

all who take refuge in Him. What god is there but the Lord? What rock but our

God? – the God who girds me with strength and makes my way blameless, who

makes me swift as the deer and sets me secure on the mountains (Psalms 18:30-

33, New English Bible).

“Commit your life to the Lord; trust in Him and He will act. He will make your

righteousness shine clear as the day and the justice of your cause like the sun at noon”

(Psalms 37:5-6).

“Not to us, O Lord, not to us, but to thy name ascribe the glory, for thy true love

and for thy constancy” (Psalms 115:1).

vi

Acknowledgments

It is only through the Lord’s strength and wisdom that this dissertation came to

fruition. Next, I acknowledge the man with whom the Lord has made me one, my

husband. You are truly the wind beneath my wings, and without you I would not have

had the wherewithal to complete this endeavor. Thank you for all your support and

sharing your perseverance for my good. I also wish to acknowledge, with unbounded

gratitude, the most perfect dissertation committee possible for this journey.

To my chair, Dr. Genomary Krigbaum, words are insufficient to fully express the

depth and breadth of my appreciation for your support, guidance, and direction. When I

first read descriptions of what the ideal chair would be, with characteristics inclusive of

mentor, advocate, role model, teacher, defender, guide, supervisor, coach, encourager,

and friend, I wondered if it would ever be possible to find all those elements in one

person. Yet in you, I found them all, and more. Por siempre agradecida. Moreover, thank

you for encouraging me to build on the methodology you started. To Dr. Daniel Smith, I

am grateful that you joined my dissertation team. I knew I could count on you for your

statistical expertise, and you did not disappoint. Thank you for the many conversations

prior to my dissertation journey, and in helping to pave the way for the best committee

possible. To Dr. Genie Bodenhamer-Davis, as a most respected neurofeedback

practitioner and educator, I am humbled and honored that you were willing to assist me in

my dissertation journey. Thank you, so much, for your counsel over the last 3 years. To

Dr. Ron Bonnstetter, thank you for your support in being my adjunct dissertation reader.

Thank you for your compliments on my writing and your assurance I have what it takes

to succeed as a scholar.

vii

Table of Contents

List of Tables ……………………………………………………………………………………………………… xi

List of Figures ……………………………………………………………………………………………………. xii

Chapter 1: Introduction to the Study …………………………………………………………………………1

Introduction ……………………………………………………………………………………………………..1

Background of the Study …………………………………………………………………………………..2

Problem Statement ……………………………………………………………………………………………4

Purpose of the Study …………………………………………………………………………………………5

Research Questions and Hypotheses …………………………………………………………………..6

Advancing Scientific Knowledge ……………………………………………………………………….8

Significance of the Study …………………………………………………………………………………..9

Rationale for Methodology ………………………………………………………………………………10

Nature of the Research Design for the Study ………………………………………………………11

Definition of Terms…………………………………………………………………………………………13

Assumptions, Limitations, Delimitations …………………………………………………………..19

Summary and Organization of the Remainder of the Study ………………………………….22

Chapter 2: Literature Review …………………………………………………………………………………23

Introduction and Background to the Problem ……………………………………………………..23

Historical overview of EEG and QEEG. ……………………………………………….24

Historical overview of NF …………………………………………………………………..25

How problem/gap of 19ZNF research evolved into current form ……………..28

Theoretical Foundations and/or Conceptual Framework ………………………………………31

Foundations of EEG and QEEG …………………………………………………………..31

viii

Learning theory as applied to NF………………………………………………………….31

Traditional/amplitude-based models of NF ……………………………………………33

QNF model of NF ………………………………………………………………………………35

ZNF model of NF……………………………………………………………………………….38

Review of the Literature – Key Themes …………………………………………………………….39

QNF in the literature …………………………………………………………………………..39

4ZNF in the literature………………………………………………………………………….47

19ZNF in the literature………………………………………………………………………..50

Outcome measures for ZNF research ……………………………………………………53

Summary ……………………………………………………………………………………………………….59

Chapter 3: Methodology ……………………………………………………………………………………….61

Introduction ……………………………………………………………………………………………………61

Statement of the Problem …………………………………………………………………………………61

Research Questions and Hypotheses …………………………………………………………………62

Research Methodology ……………………………………………………………………………………64

Research Design……………………………………………………………………………………………..65

Population and Sample Selection………………………………………………………………………66

Instrumentation ………………………………………………………………………………………………68

Validity …………………………………………………………………………………………………………72

Reliability ………………………………………………………………………………………………………74

Data Collection Procedures ………………………………………………………………………………76

Data Analysis Procedures ………………………………………………………………………………..78

Ethical Considerations …………………………………………………………………………………….81

ix

Limitations …………………………………………………………………………………………………….82

Summary ……………………………………………………………………………………………………….84

Chapter 4: Data Analysis and Results ……………………………………………………………………..86

Introduction ……………………………………………………………………………………………………86

Descriptive Data……………………………………………………………………………………………..86

Data Analysis Procedures ………………………………………………………………………………..93

Results …………………………………………………………………………………………………………..96

Summary ……………………………………………………………………………………………………..103

Chapter 5: Summary, Conclusions, and Recommendations ……………………………………..105

Introduction ………………………………………………………………………………………………….105

Summary of the Study …………………………………………………………………………………..106

Summary of Findings and Conclusion ……………………………………………………………..107

Implications………………………………………………………………………………………………….113

Theoretical implications…………………………………………………………………….114

Practical implications ………………………………………………………………………..115

Future implications. ………………………………………………………………………….116

Recommendations …………………………………………………………………………………………117

Recommendations for future research. ………………………………………………..117

Recommendations for practice. ………………………………………………………….118

References …………………………………………………………………………………………………………120

Appendix A ……………………………………………………………………………………………………….136

Appendix B ……………………………………………………………………………………………………….137

x

Appendix C ……………………………………………………………………………………………………….138

Appendix D ……………………………………………………………………………………………………….139

xi

List of Tables

Table 1.1. Research Questions and Variables …………………………………………………………….8

Table 4.1. Descriptive Data for All Groups ……………………………………………………………. 91

Table 4.2. Shapiro-wilk Results for Difference Scores ……………………………………………. 95

Table 4.3. Summary of Results – All Groups………………………………………………………….104

xii

List of Figures

Figure 1.1. Formation of Sample Groups ………………………………………………………………. 13

Figure 4.1. IVA Group Pre-Post Scores…………………………………………………………………. 97

Figure 4.2. DSMD Group Pre-Post Scores …………………………………………………………….. 99

Figure 4.3. BRIEF Group Pre-Post Scores …………………………………………………………… 101

Figure 4.4. QEEG Group Pre-Post Scores …………………………………………………………… 102

1

Chapter 1: Introduction to the Study

Introduction

Neurofeedback (NF) is an operant conditioning brainwave biofeedback technique,

which is also referred to as electroencephalographic (EEG) biofeedback. This modality,

dating back to the 1970s (Lubar & Shouse, 1976; Sterman, LoPresti, & Fairchild, 2010),

trains electrical signals of targeted frequencies and involves recording EEG data from

scalp sensors with an amplifier, which is subsequently processed by computer software.

The software provides visual and sound display feedback to the trainee, thereby

providing a reward stimulus when the brain is functioning in the target range. This

reward process generates learning such that the brain’s functioning is conditioned in the

intended manner.

Over the years, new models of NF have been developed, and the most current

iteration is a style of NF which is termed z-score NF (ZNF). ZNF is different from more

traditional NF models in that it incorporates into the NF session real-time quantitative

EEG (QEEG) z-score metrics making it possible to combine operant conditioning with

real-time assessment using a normative database (Collura, Thatcher, Smith, Lambos, &

Stark 2009; Thatcher, 2012). In 2006, a 4-channel ZNF (4ZNF) technique was

introduced, which in 2009 was expanded to include all 19 sites of the International 10-20

System (of electrode placement) to allow for a 19-channel ZNF (19ZNF). To date, case

study and anecdotal clinical reports within the field indicate this new 19ZNF approach is

an improvement over traditional NF models (J. L. Koberda, Moses, Koberda & Koberda,

2012a; Wigton, 2013). However the efficacy of this new model has not yet been

established from empirical studies. This research is different from prior qualitative

2

studies; it has been completed as a quantitative analysis of pre-post outcome measures

with group data, and thus, it is a beginning in establishing empirical evidence regarding

19ZNF.

The remainder of this chapter formulates this dissertation through a review of the

study background, problem statement, purpose and significance, and how this research

advances the scientific knowledge. Moreover the research questions and hypotheses are

presented, together with the methodology rationale and the nature of the research design.

An extended Definition section is included to review the many technical terms germane

to this research. Readers unfamiliar with NF or QEEGs may find it helpful to review the

definitions first. Finally, to establish the scope of the study, a list of assumptions,

limitations, and delimitations are included.

Background of the Study

In recent years NF has seen increasing acceptance as a therapeutic technique.

Current literature includes reviews and meta-analyses which establish a recognition of

NF as effective for the specific condition of attention deficit hyperactivity disorder

(ADHD) (Arns, de Ridder, Strehl, Breteler, & Coenen 2009; Brandeis, 2011;

Gevensleben, Rothenberger, Moll, & Heinrich, 2012; Lofthouse, Arnold, Hersch, Hurt, &

DeBeus, 2012; Niv, 2013; Pigott, De Biase, Bodenhamer-Davis, & Davis, 2013).

However, the type of NF covered in these reviews is limited to the oldest NF model

(theta/beta ratio) and/or slow cortical potential NF. Yet of note are reports in the literature

of a different NF model which is informed by QEEG data. This QEEG-guided NF (QNF)

is reported to be used for a much wider range of conditions; not only ADHD, but also

behavior disorders, cognitive dysfunction, various mood disorders, epilepsy,

3

posttraumatic stress disorder, head injuries, autism spectrum disorders, migraines,

learning disorders, schizophrenia, and mental retardation (Arns, Drinkenburg, &

Kenemans, 2012; Breteler, Arns, Peters, Giepmans, & Verhoeven, 2010; Coben &

Myers, 2010; J. L. Koberda, Hillier, Jones, Moses, & Koberda 2012; Surmeli, Ertem,

Eralp, & Kos, 2012; Surmeli & Ertem, 2009, 2010, 2011; Walker, 2009, 2010b, 2011,

2012b).

Yet, all the aforementioned models are limited in their use of only one or two

electrodes and they also require many sessions to achieve good clinical outcomes. For the

above-cited studies the reported average number of sessions was 40.5. Moreover,

Thatcher (2012, 2013) reports 40 to 80 sessions to be the accepted norm for these older

style models; thus leading to a sizeable cost to access this treatment. However, one of the

newest ZNF models shows promise to bring about positive clinical outcomes in

significantly fewer sessions (Thatcher, 2013). With 4ZNF there have been reports of

successful clinical outcomes with less than 25 sessions (Collura, Guan, Tarrant, Bailey, &

Starr, 2010; Hammer, Colbert, Brown, & Ilioi, 2011; Wigton, 2008); whereas clinical

reviews and recent conference reports (J. L. Koberda, Moses, Koberda, & Koberda,

2012b; Rutter, 2011; Wigton, 2009, 2010a, 2010b, 2013; Wigton & Krigbaum, 2012)

suggest 19ZNF can result in positive clinical outcomes, as well as QEEG normalization,

in as few as 5 to15 sessions. Therefore a NF technique which shows promise to bring

clinical improvement in fewer sessions – thereby reducing treatment cost – deserves

empirical study.

Currently in the peer-reviewed published literature, there are a couple of

descriptive and clinical review articles about the 19ZNF model (Thatcher, 2013; Wigton,

4

2013) and two single case study reports (Hallman, 2012; J. L. Koberda et al., 2012a);

however rigorous scientific studies evaluating 19ZNF have not been found, which poses

a gap in the literature. Therefore, before the question of efficiency and number of

sessions is examined, first its efficacy should be established. NF and ZNF efficacy has

been discussed in the literature as having the desired effect in terms of improved clinical

outcomes (La Vaque et al., 2002; Thatcher, 2013; Wigton, 2013), a definition that fits

well within the scope of this research. In this study, there are two types of clinical

outcome measures; one type (clinical assessments) is a set of psychometric tests designed

to measure symptom severity and/or improvement, the other type (QEEG z-scores)

provides a representative measure of electrocortical dysfunction and/or improvement.

Thus, this dissertation is intended to address efficacy of 19ZNF in a clinical setting,

through a retrospective evaluation of clinical outcomes, as measured by clinical

assessments and QEEG z-scores.

Problem Statement

It is not known, by way of statistical evaluation of either clinical assessments or

QEEG z-scores, if 19ZNF is an effective NF technique. This is an important problem

because 19ZNF is a new NF model currently in use by a growing number of practitioners,

yet scientific research investigating its efficacy is lacking. According to an Efficacy Task

Force, established by the two primary professional organizations for NF and biofeedback

professionals, 1 anecdotal reports (regardless of how many) are insufficient as a basis for

1 The primary professional societies for neurofeedback and biofeedback are the International

Society for Neurofeedback and Research (ISNR; www.isnr.org) and the Association for Applied

Psychophysiology and Biofeedback (AAPB; www.aapb.org).

5

determining treatment efficacy, and uncontrolled case studies are scientifically weak (La

Vaque et al., 2002). Therefore, scientific evidence of efficacy for 19ZNF is needed.

The identified population for this study is made up of those seeking NF services

(both adults and children), and those who become NF clients. These individuals may

have an array of symptoms, which adversely affect their daily functioning; they may also

have previously diagnosed mental health disorders. When seeking NF services these

individuals must choose among a variety of NF models. However the dearth of scientific

literature regarding 19ZNF limits the information available to inform that decision-

making process. Therefore, it is vital that both NF clinicians and clients have empirically

derived information regarding the clinical value and efficacy of this new NF technique.

Consequently, the problem of this empirical gap impacts the NF clinician and client alike.

The goal of this research is to contribute in providing a first step towards addressing this

research gap.

Purpose of the Study

The purpose of this quantitative, retrospective, one-group, pretest-posttest study

research was to compare the difference between pre and post clinical assessments and

QEEG z-scores data, before and after 19ZNF sessions, from archived data of a private

neurofeedback practice in the Southwest region of the United States. The comparisons

were accomplished via statistical analysis appropriate to the data (i.e. paired t tests), and

will be further discussed in the Data Analysis section of Chapter 3. The independent

variable is defined as the 19ZNF, and the dependent variables are defined as the standard

scaled scores of three clinical assessments and QEEG z-score data. The clinical

assessments measure symptoms of attention, behavior, and executive function, whereas

6

the z-scores provide a representative measure of electrocortical function. The full scopes

of the assessments are further outlined in the Instrumentation section of Chapter 3.

Given the retrospective nature of this study, there were no individuals, as subjects,

with which to interact. However the target population group is considered to be adults

and children with clinical symptoms of compromised attention, behavior, or executive

function, who are interested in NF as an intervention for improvement of those

symptoms. This pretest-posttest comparison research contributes to the NF field by

conducting a scientific study, using quantitative group methods, to address the efficacy of

the new 19ZNF model.

Research Questions and Hypotheses

If the problem to be addressed is a lack of scientific evidence demonstrating

efficacy of 19ZNF, the solution lies in evaluating its potential for improving clinical

outcomes as measured by clinical assessments and electrocortical metrics. Therefore

research questions posed in terms of clinical symptomology and cortical function

measures is a reasonable approach. For this research the independent variable is the

19ZNF and the dependent variables are clinical outcomes, as measured by the scaled

scores from three clinical assessments and z-scores from QEEG data. The clinical

assessments are designed to measure symptom severity of attention, behavior, and

executive functioning, and the z-scores are a representational measure of electrocortical

function. The data gathering, scores calculation, and, data analysis were conducted by the

researcher.

7

The following research questions guided this study:

R1a. Does 19ZNF improve attention as measured by the Integrated Visual and

Auditory continuous performance test (IVA; BrainTrain, Incorporated,

Chesterfield, VA)?

Ha1a: The post scores will be higher than the pre scores for the IVA

assessment.

H01a: The post scores will be lower than, or not significantly different

from, the pre scores of the IVA assessment.

R1b. Does 19ZNF improve behavior as measured by the Devereux Scale of

Mental Disorders (DSMD; Pearson Education, Incorporated, San Antonio, TX)?

Ha1b: The post scores will be lower than the pre scores for the DSMD

assessment.

H01b: The post scores will be higher than, or not significantly different

from, the pre scores of the DSMD assessment.

R1c. Does 19ZNF improve executive function as measured by the Behavior

Rating Inventory of Executive Functioning (BRIEF; Western Psychological

Services, Incorporated, Torrance, CA)?

Ha1c: The post scores will be lower than the pre scores for the BRIEF

assessment.

H01c: The post scores will be higher than, or not significantly different

from, the pre scores of the BRIEF assessment.

R2. Does 19ZNF improve electrocortical function as measured by QEEG z-scores

(using the Neuroguide Deluxe software, Applied Neuroscience Incorporated, St.

8

Petersburg, FL), such that the post z-scores are closer to the mean than pre z-

scores?

Ha2: The post z-scores will be closer to the mean than the pre z-scores.

H02: The post z-scores will be farther from the mean, or not significantly

different from, the pre z-scores.

See as follows Table 1.1, outlining the research questions and variables.

Table 1.1

Research Questions and Variables

Research Questions Hypotheses Variables Instrument(s) 2. 1a. Does 19ZNF improve

attention as measured by

the IVA?

The post scores will be

higher than the pre scores

for the IVA assessment.

IV: 19ZNF

DV: IVA standard scale

scores

IVA

computerized

performance test

1b. Does 19ZNF

improve behavior as

measured by the DSMD?

The post scores will be

lower than the pre scores

for the DSMD

assessment.

IV: 19ZNF

DV: DSMD standard

scale scores

DSMD

rating scale

1. 1c. Does 19ZNF improve executive function as

measured by the BRIEF?

The post scores will be

lower than the pre scores

for the BRIEF

assessment.

IV: 19ZNF

DV: BRIEF standard

scale scores

BRIEF

rating scale

2. 2. Does 19ZNF improve electrocortical function

as measured by QEEG z-

scores such that the post

z-scores are closer to the

mean than pre z-scores?

The post QEEG z-scores

will be closer to the mean

than the pre z-scores.

IV: 19ZNF

DV: QEEG

z-scores

QEEG

z-score data generated

from Neuroguide

software

Advancing Scientific Knowledge

The theoretical framework of NF is the application of operant conditioning upon

the EEG, which leads to electrocortical changes, and in turn, better brain function and

clinical symptom improvement; moreover, studies evaluating traditional NF have

9

demonstrated its efficacy (Arns et al., 2009; Pigott et al., 2013). The 19ZNF model is

new, and experiencing increased use in the NF field, yet efficacy has not been established

via empirical investigation. There is a gap in the literature in that the only peer-reviewed

information available to date, regarding 19ZNF, are reviews, clinical report presentations,

and single case studies. Also noted as typically absent from traditional NF studies are

analyses of pre-post QEEG data (Arns et al., 2009); this lack of pre-post QEEG data

continues in the QNF literature as well. This, then, poses a secondary gap, in terms of

methodology, which this study has the potential to fill.

The clinical condition most researched for demonstrating traditional NF efficacy

is ADHD (Pigott et al., 2013), which includes cognitive functions of attention and

executive function. These issues also lead to some associated behavioral problems with

adverse impacts in instructional settings that are also treated with 19ZNF. Therefore,

addressing efficacy of 19ZNF with clinical assessments designed to measure these

constructs, will contribute to filling the gap of what is not known about this new NF

model, within a framework related to cognition and instruction. If efficacy is

demonstrated, the theory of operant conditioning, upon which NF is founded, may be

expanded to include 19ZNF.

Significance of the Study

The 19ZNF model is theoretically distinctly different from traditional NF in that it

targets real-time QEEG z-scores with a goal of normalizing QEEG metrics (as indicated

by clinical symptom presentation) rather than only increasing or decreasing targeted brain

frequencies. This model has been in existence for five years and its use by NF clinicians

is rapidly growing. Thus far, other than two qualitatively-oriented, single case study

10

reports (Hallman, 2012; J. L. Koberda et al., 2012a), there are no empirical group studies,

with a quantitative methodology, studying the efficacy of 19ZNF in peer-reviewed

literature. The significance of this study is that it aims to fill this significant gap manifest

as a dearth of 19ZNF efficacy studies.

Moreover, few NF studies include analysis of EEG measures as an outcome

measure (Arns et al., 2009). Therefore demonstrating how z-scores from QEEG data can

be used for group comparison studies, in a way not previously explored, will benefit the

scientific community. Thus, this research has the potential for opening doors for further

research.

It was expected the findings would demonstrate 19ZNF results in improved

clinical outcomes, as measured by clinical and QEEG assessments; thus demonstrating

efficacy. Potential NF clients will benefit from this contribution of what is known about

19ZNF by having more information upon which to base decisions for what type of NF

they wish to pursue. The potential effect of these results may provide the start of an

evidence-based foundation for its use. This foundation may lead to a greater acceptance

of what may be a more efficient (and thereby more economical) NF model, as well as

foster the needed additional scientific research of 19ZNF.

Rationale for Methodology

The field of clinical psychophysiology makes use of quantifiable variables and the

associated research should include specific independent variables, as well as dependent

variables that relate to treatment response (e.g. clinical assessments) and the measured

physiological component (e.g. EEG metrics) (La Vaque et al., 2002). Yet, many NF

studies do not use the EEG metric as a psychophysiologic measure, but rather provide

11

reports, which are more qualitative in nature. Therefore, there is a need for NF research,

with sound quantitative methodologies, using QEEG data as an outcome measure.

Currently, the available 19ZNF studies are in the form of qualitative research

(Hallman, 2012; J. L. Koberda et al., 2012a). This literature entails presenting data, from

single case studies, in the form of unstructured subjective reports of symptom

improvement and graphical images of before and after QEEG findings, where the

improvement is represented by a change in color on the picture (without statistical

analysis of data). However, for this dissertation, the goal is to explore statistical

relationships between the variables under investigation. The strength of quantitative

methodologies, including quasi-experimental research, is that they provide sufficient

information, regarding the relationship of the investigation variables, to enable the study

of the effects of the independent variable upon the dependent variable (Carr, 1994); this

is suitable in the evaluation of a quantitative technology such as 19ZNF.

As previously stated, for this research the independent variable is specified as

19ZNF. The dependent variables in this study are continuous variables in the form of

standard scores from clinical assessments (IVA, DSMD, and BRIEF) and z-scores from

QEEG data. The alternative hypotheses for all research questions predict a directional

significant difference between the means of the pre and the post values for all dependent

variables. Therefore, a quantitative methodology is appropriate for this dissertation.

Nature of the Research Design for the Study

This quasi-experimental research used a retrospective one-group, pretest-posttest

design. When the goal of research is to measure a modification of a behavior pattern, or

internal process that is stable and likely unchangeable on its own, the one-group pretest-

12

posttest design is appropriate (Kerlinger, 1986). In this type of design the dependent

variable pretest measures are compared to the posttest values for each subject, thus

comparing the members of the group to themselves rather than to a control or comparison

group (Kerlinger, 1986). Consequently, the group members become their own control,

hence reducing the potential for extraneous variation due to individual-to-individual

differences (Kerlinger & Lee, 2000). Moreover, the size of the treatment effect can be

estimated by analyzing the difference between the pretest to the posttest measures

(Reichardt, 2009). Therefore, this design as well as a quantitative methodology, is well

suited to evaluate the pre-post outcome measures from a clinical setting.

The rationale for this being a retrospective study is based on the fact that data

available for analysis came from pre-existing archived records, which frequently provides

a rich source of readily accessible data (Gearing, Mian, Barber, & Ickowicz, 2006).

Within the pool of available data, a sample group was gathered for which various pre and

post assessments were performed during the course of 19ZNF treatment. As depicted in

Figure 1.1, an initial group was formed for which pre-post QEEG assessments and z-

scores were available, and for which either the IVA, DSMD, or BRIEF pre-post

assessment data was also available (n = 21). From this collection three additional groups

were formed: One group for the IVA data (n = 10), a second group for the DSMD data (n

= 14), and a third group for the BRIEF data (n = 12). Therefore, using a one-group

pretest-posttest design with these identified groups is fitting. The independent variable is

the 19ZNF and the dependent variables are the data from the clinical assessments and

QEEG files (IVA, DMSD, BRIEF, and z-scores).

13

Formation of Sample Groups

Figure 1.1. Illustration of how the sample groups were formed. The

total number of subjects in the sample is 21. However, out of those

21, some may have multiple assessments, therefore subjects may be

in more than one clinical assessment group.

Definition of Terms

The following terms were used operationally in this study.

19ZNF. 19-channel z-score NF is a style of NF using all 19 sites of the

International 10-20 system, where real-time QEEG metrics are incorporated into the NF

session in the form of z-scores (Collura, 2014). The goal is for the targeted excessive z-

score metrics (whether high or low) to normalize (move towards the mean). The 19ZNF

cases included in this study are those for which the assessed clinical symptoms

corresponded with the z-score deviations of the QEEG findings, such that a treatment

goal of overall QEEG normalization was clinically appropriate. While the 19ZNF

protocols are individually tailored to the clinical and QEEG findings, the same treatment

goal always applies, that is the overall QEEG normalization. Therefore, the underlying

19ZNF protocol of overall QEEG normalization is consistent for all cases.

14

Absolute power. A QEEG metric which is a measure of total energy, at each

electrode site, for a defined frequency band (Machado et al., 2007); may be expressed in

terms of microvolts, microvolts squared, or z-scores when compared to a normative

database (Collura, 2014).

Amplifier. The equipment that detects, amplifies, and digitizes the brainwave

signal (Collura, 2014). The term is more correctly referred to as a differential amplifier

because the electrical equipment measures the difference between two signal inputs

(brainwaves from electrode locations) (Collura, Kaiser, Lubar, & Evans, 2011).

Amplitude. A measure of the magnitude or size of the EEG signal; and is

typically expressed in terms of microvolts (uV) (Collura et al., 2011). This can be thought

of as how much energy is in the EEG frequency.

Biofeedback. A process of learning how to change physiological activity with the

goal of improving health and/or performance (AAPB, 2011). A simple example of

biofeedback is the act of stepping on a scale to measure one’s weight.

Behavior Rating Inventory of Executive Functioning (BRIEF). The BRIEF,

published by Western Psychological Services, Incorporated (Torrance, CA), is a rating

scale. It has forms for both children and adults, and is designed to assess behavioral,

emotional, and metacognitive skills, which broadly encompass executive skills, rather

than measure behavior problems or psychopathology (Donders, 2002). The test results

are expressed as T scores for various scales and sub-scales (with clinically significant

scores ≥ 65), and lower scores indicate improvement upon re-assessment. The composite

and global scales of Behavior Regulation Index, Metacognition Index, and Global

Executive Composite were included in this study.

15

Coherence. A measure of similarity between two EEG signals, which also

reflects the degree of shared information between the sites; computed in terms of a

correlation coefficient, which varies between .00 to 1.00 (Collura et al., 2011).

Devereux Scale of Mental Disorders (DSMD). The DSMD, published by

Pearson Education, Incorporated (San Antonio, TX), is a rating scale. It is designed to

assess behavior problems and psychopathology in children and adolescents (Cooper,

2001). The test results are reported in the form of T scores for various scales and sub-

scales (with clinically significant scores ≥ 60), and lower scores indicate improvement

upon re-assessment. The composite and global scales of Externalizing, Internalizing, and

Total were included in this study.

Electrode. Central to NF is the detection and analysis of the EEG signal from the

scalp. In order to record brainwaves it is necessary to attached metallic sensors

(electrodes) to the scalp and/or ears (with a paste or gel) to facilitate this process (Collura,

2014).

Electroencephalography (EEG). A recording of brain electrical activity (i.e.

brainwaves) using differential amplifiers, measured from the scalp (Collura et al., 2011).

The information from each site or channel is digitized to be viewed as an oscillating line,

such that all channels can be viewed on a computer screen at one time.

Fast Fourier transform (FFT). The conversion of a series of digital EEG

readings into frequency ranges/bands, which can be viewed in a spectral display. Just as

different frequencies of light can be seen when filtered through a prism, so too can EEG

elements be isolated when filtered through a FFT process into different frequency bands

(Collura, 2014).

16

Frequency / frequency bands. The representation of how fast the signal is

moving, expressed in terms of Hertz (Hz) (Collura, 2014) and commonly arranged in

bandwidths, also referred to as bands. Generally accepted frequency bands are delta (1-4

Hz), theta (4-7 Hz), alpha (8-12 Hz), beta (12-25 Hz), and high beta (25-30 Hz); the beta

band may be broken down into smaller bands of beta1 (12-15 Hz), beta2 (15-18 Hz), and

beta3 (18-25 Hz), and the alpha band may be divided into alpha1 (8-10 Hz) and alpha2

(10-12 Hz).

Gaussian. Referring to the normal distribution and/or normal curve (Thatcher,

2012).

Hedges’ d. An effect size, belonging to the d family indices (along with Hedges’

g), which use the standard score form of the difference between the means; therefore it is

similar to the Cohen’s d, with the same interpretation (Hunter & Schmidt, 2004).

However, when used with small sample sizes, both the Cohen’s d and Hedges’ g, can

have an upward bias and be somewhat over-inflated; however the Hedges’ d includes a

correction for this bias (Hunter & Schmidt, 2004). Therefore, in studies with smaller

sample sizes, the use of the Hedges’ d provides a more conservative, and likely more

accurate effect size. Also, complicating this issue is confusion in the literature regarding

the use of the designator g or d for which particular Hedges index, and/or which

calculation does or does not include the correction factor (Hunter & Schmidt, 2004). For

example, frequently Hedges’ g is described as adjusting for small samples sizes;

however, this is only true if the calculation used includes the correction factor. Moreover,

there are even variations in the literature of the correction equation which is applied. As a

result, the only way to know which calculation is actually being used is for the Hedges’

17

index equation to be explicitly reported. To that end, for this study, the Hedges’ d

definition/calculation will be that used in the Metawin 2.1 meta-analysis software

(Rosenberg, Adams, & Gurevitch, 2000). In this context the Hedges’ d is calculated by

multiplying the Hedges’ g by the correction, which is sometimes referred to as J.

Where and therefore .

Hertz (Hz). The number of times an EEG wave oscillates (moves up and down)

within a second; commonly expressed as cycles per second (Collura, 2014).

International 10-20 System. A standardized and internationally accepted method

of EEG electrode placement locations (also referred to as sites) on the scalp. The

nomenclature of 10-20 derives from electrode locations being spaced a distance of either

10% or 20% of the measured distance from certain landmarks on the head. The system

consists of a total of 19 sites, with eight locations on the left, eight on the right, and three

central sites found on the midline between the right and left side of the head (Collura,

2014).

Visual and Auditory + Plus Continuous Performance Test (IVA). The IVA,

developed and published by BrainTrain Incorporated (Chesterfield, VA), is a

computerized interactive assessment. It is normed for individuals over the age of 5, and it

is designed to assess both auditory and visual attention and impulse control with the aim

to aid in the quantification of symptoms and diagnosis of ADHD (Sanford & Turner,

2009). The test results are reported in the form of quotient scores for various scales and

sub-scales (with clinically significant scores ≤ 85), and higher scores indicate

improvement upon re-assessment. The global and composite scales of Full Scale

18

Attention Quotient, Auditory Attention Quotient, and the Visual Attention Quotient were

included in this study.

Joint time frequency analysis (JTFA). A method of digitizing the EEG signal

which allows for moment-to-moment (i.e. real time) measures of EEG signal changes

(Collura, 2014).

Montage. The configuration of the electrodes and software defining the reference

point and electrode linkages, for the differential recording of the EEG signals (Thatcher,

2012). For example, in a linked-ears montage, the signal for each electrode site is

referenced to the signal of the ear electrodes linked together. In a Laplacian montage, the

signal for each electrode site is referenced to the signal of the weighted average of the

surrounding electrode sites.

Neurofeedback. An oversimplified, yet accurate, definition of neurofeedback is

that it is simply biofeedback with brainwaves. Generally, it is an implicit learning process

(involving both operant and classical conditioning) where changes in brainwave

signal/patterns, in a targeted direction, generates a reward (a pleasant tone and change in

a video animation) such that the desired brainwave events occur more often (Collura,

2014; Thatcher, 2012).

Normalization. In the context of NF, refers to the progression of excessive z-

scores towards the mean (i.e. z = 0), meaning the NF trainee’s EEG is moving closer to

the EEG range of normal (i.e. typical) individuals of his/her age (Collura, 2014). Thus,

the concept of normalization is generally accepted to be when the z-scores of the QEEG

move towards the mean (i.e. in the direction of z = 0).

19

Power spectrum. The distribution of EEG energy across the frequency bands,

typically from 1 Hz to 30 Hz and frequently displayed as a line graph, histogram, or color

topographic (i.e. visual representations of the numerical data) images (Collura, 2014).

Phase. The temporal relationship between two EEG signals, reflecting the speed

of shared information (Collura et al., 2011).

Protocol. The settings designated in NF software, informed by a treatment plan,

which determines how the NF proceeds. This establishes parameters such as metrics (e.g.

absolute vs. relative power), direction of training (i.e. targeting more or less), length of

session, and other decision points in the NF process (Collura, et al., 2011).

Quantitative EEG (QEEG). The numerical analysis of the EEG such that it is

transformed into a range of frequencies as well as various metrics such as absolute

power, relative power, power ratios, asymmetry, coherence, and phase (Collura, 2014;

Thatcher, 2012). The data is typically made up of raw numbers, statistical transforms into

z-scores, and/or topographic images (Collura, 2014). As a dependent variable in this

study, QEEG z-scores are considered a representational measure of electrocortical

function. The metrics of absolute power, relative power, and coherence were included.

Relative power. A QEEG metric representing the amount of energy, divided by

the total energy, at each electrode site, for a defined frequency band. It reflects how much

energy is present compared to all other frequencies (Collura, 2014).

Assumptions, Limitations, Delimitations

This section identifies the assumptions and specifies the limitations, together with

the delimitations of the study. The following assumptions were present in this study:

20

1. It was assumed that traditional neurofeedback is deemed efficacious as

discussed and demonstrated in the literature (Arns et al., 2009; Pigott et

al., 2013).

2. It was assumed that the subjects are representative of the population of those

who seek NF treatment for various mental health disorders; thus allowing

for results to be generalized to that population (Gravetter & Wallnau,

2010).

3. It was assumed the sample is homogeneous and selected from a population

that fits the normal distribution such that the sample means distribution are

also likely to fit a normal distribution (Gravetter & Wallnau, 2010).

4. It was assumed that responses provided on rating scale instruments accurately

reflect perceived or remembered observations, thus minimizing bias for

over or under-reporting of observations (Kerlinger & Lee, 2000).

The following limitations were present in this study:

1. Research design elements. A general limitation of designs that incorporate a

pretest-posttest formulation is primarily related to the passage of time

between administering the pre and post assessments (Kerlinger & Lee,

2000). Factors such as history and maturation cannot be controlled for;

therefore it is not possible to know whether or not they have impacted the

dependent variable measures (Hunter & Schmidt, 2004). However, for this

study the time between the pre and post assessment is relatively short, and

can be measured in terms of weeks. Therefore, the impact of time-related

confounds were anticipated to be minimal. Further limitations which also

21

must be recognized are a lack of comparison to a traditional NF group, and

a lack of a randomized control group.

2. Small sample size. Larger sample sizes are preferred in order to allow for

stronger statistical analysis and more generalizability (Gravetter &

Wallnau, 2010). Given this study used pre-existing archived data, the

number of samples were restricted to what was found in the files; thus

there was no option to increase sample size. Though, as detailed in

Chapter 3, the sample sizes for each group provided sufficient power to

allow for adequate statistical analysis.

The following delimitations will be present in this study:

1. This study was delimited to the scope of the surface formulation of 19ZNF.

Therefore, it did not include in its scope other variations of 19-channel NF

models, founded in inverse solution theories, such as low-resolution brain

electromagnetic tomography (LORETA) ZNF or functional magnetic

resonance imaging (fMRI) tomography NF models.

2. This study was delimited to a scope of NF research data collected primarily

from clinical settings, as opposed to laboratory-based experimental

research.

3. The academic quality standards for this dissertation delimit the literature

reviewed for this study to exclude certain non-peer-reviewed sources (i.e.

NF industry newsletters).

In spite of the above stated assumptions, limitations, and delimitations, this study

has potential to be of value to the scientific and neurofeedback community. Given the

22

data for this research comes from a real-world clinical setting, the findings of this study

still contribute to advancing the scientific knowledge of 19ZNF.

Summary and Organization of the Remainder of the Study

In summary, while NF has a history spanning over 40 years, it is only now

gaining acceptance as an evidence-based mental health intervention (Pigott et al., 2013).

Various models of NF have been developed over the years, with one of the newest

iterations including 19ZNF, which is reported to lead to improved clinical outcomes in

fewer sessions than other models (Thatcher, 2013; Wigton, 2013). However, there are

significant gaps in terms of peer-reviewed literature and research, such that efficacy of

19ZNF has yet to be established. This dissertation intends to fill these gaps by addressing

efficacy of 19ZNF, in a clinical setting, using a comparison of pretest-posttest measures

of clinical assessments and QEEG z-scores.

The following chapters include the literature review in Chapter 2 and a

description of the methodology, research design, and the procedures for the study in

Chapter 3. The literature review first explores the background and history of the problem,

then discusses theoretical foundations and conceptual frameworks, and finally reviews

the literature pertaining to the NF models relevant to this study. Of note is the necessity

of a significantly expanded theoretical/conceptual section. The methodological

foundations of a treatment intervention based in EEG/QEEG technology, combined with

the need to explore the theoretical foundations of three different NF models (traditional,

QNF, and ZNF), require more in depth coverage of the topics involved in that section.

23

Chapter 2: Literature Review

Introduction and Background to the Problem

The focus of this study was to explore the efficacy of 19ZNF in a clinical setting,

through the use of clinical assessments and QEEG z-scores as outcome measures. Yet, a

review of the literature is necessary to place this research into context of NF theory and

the various models that have come before 19ZNF. This literature review consists of three

sections.

The first section addresses the history and background of NF in general and

specifically introduces ZNF, as well as comments on how the gap in research for 19ZNF

evolved into its current form. The second section focuses on the theoretical foundations

and conceptual frameworks of NF and QEEG. First, an overview of the foundations of

EEG and QEEG is presented. Next, an overview of learning theory as applied to NF is

discussed. Then, the theoretical frameworks supporting the different models of NF

(traditional, QNF, and ZNF) are reviewed. Last, key themes of NF concepts relevant to

this dissertation including applications of QNF, the development of 4ZNF, and finally the

emergence of 19ZNF are examined. Also included in this section is a review of suitable

outcome measures for use in ZNF research, with special attention paid to prior NF

research regarding performance tests, rating scale assessments, and QEEG z-scores, as

outcome measures.

Of note for this literature review is the necessity to include reviews of conference

oral and poster presentations (which are subject to a peer-review acceptance process).

While inclusion of these sources may be an unusual dissertation strategy, it is necessary

due to the scarcity of sources in the peer-reviewed published literature regarding ZNF

24

models. To exclude these sources would be to limit the coverage of the available

literature regarding the NF model which is the focus of this dissertation (19ZNF).

The literature for this review was surveyed through a variety of means. The

researcher’s personal library (from nearly fifteen years of practicing in the NF field)

served as the foundation for the literature search. Then, this was expanded through online

searches of various university libraries via academic databases such as Academic Search

Complete, PsycINFO, PsycARTICLES, and MEDLINE, with search strings of

combinations of terms such as NF, QEEG, EEG biofeedback, z-score(s). Additionally,

the databases of various industry specific journals, such as the Journal of Neurotherapy,

Clinical EEG and Neuroscience, as well as the Applied Psychophysiology and

Biofeedback journal were queried with similar search terms. Moreover, with the specified

journals, names of leading authors in the QNF and ZNF field (e.g. Koberda, Surmeli,

Walker) were used for search terms.

Historical overview of EEG and QEEG. A review of NF literature reveals a

common theme that the deepest roots of NF go back only as far as Hans Berger’s (1929)

discovery of EEG applications in humans. However, the antecedents of EEG technology

can actually be traced back as far as the 1790s with the work of Luigi Galvani and the

discovery of excitatory and inhibitory electrical forces in frog legs, leading to the

recognition of living tissue having significant electrical properties (Bresadola, 2008;

Collura, 1993). The next notable application occurred when Richard Caton (1875) was

the first to discover electrical activity in the brains of monkeys, rabbits, and cats, and to

make observations regarding the relationship of this activity to physiological functions

(Collura, 1993). Yet for applications of EEG in humans, Berger is generally recognized

25

as the first to record and report on the phenomenon. Thus, it would be most correct to

consider Caton as the first electroencephalographer, and Berger as the first human

electroencephalographer (Collura, 1993). Moreover, Berger’s contributions were

significant as they spurred a plethora of research and technological advancements in EEG

technology in the 1930s and 1940s worldwide. Of note is that Berger not only identified

both alpha and beta waves, but he was also the first to recognize the EEG signal as being

a mixture of various frequencies which could be quantitatively estimated, and spectrally

analyzed through the use of a Fourier transform, thus paving the way for QEEG

technology as well (Collura, 2014; Thatcher, 2013; Thatcher & Lubar, 2009).

Even while there was an understanding of multiple components to the EEG signal

as early as the 1930s, the advent of computer technology was necessary to make possible

QEEG advances (Collura, 1995); for example, the incorporation of normative databases

in conjunction with QEEG analysis. Therefore, the historical landmarks of EEG

developments can trace the modern start of normative database applications of QEEG

back to the 1970s with the work of Matousek and Petersen (1973) as well as John (1977)

(Pizzagalli, 2007; Thatcher & Lubar, 2009). However, while work exploring NF

applications with QEEG technology began in the 1970s, its wider acceptance and use in

the NF field was not until closer to the mid-1990s (Hughes & John, 1999; Thatcher &

Lubar, 2009). Here too, advances in computer technology, whereby personal computers

were able to process more data in less time, made way for advances in the clinical

applications of NF.

Historical overview of NF. The historical development of neurofeedback dates

back to the 1960s and early 1970s when researchers were studying the EEG activity in

26

both animals and humans. In these early days, Kamiya (1968, 1969) was studying how

humans could modify alpha waves, and Sterman and colleagues (Sterman et al., 2010;

Wyricka & Sterman, 1968) were able to demonstrate that cats could generate sensory

motor rhythm, which led to the discovery that this process could make the brain more

resistant to seizure activity; this eventually carried over to work in humans (Budzynski,

1999). Later, Lubar (Lubar & Shouse, 1976), expanded on Sterman’s work, and began

studies applying NF technology to the condition of attention disorders. This work led to

an expansion of clinical applications of neurofeedback to mental health issues such as

ADHD, depression and anxiety, using a training protocol generally designed to increase

one frequency (low beta or beta, depending on the hemisphere) and decrease two other

frequencies (theta and high beta) (S. Othmer, Othmer, & Kaiser, 1999).

Then, in the 1990s QEEG technology began gaining wider acceptance in the NF

community, for the purpose of guiding the development of protocols for NF (Johnstone,

& Gunkelman, 2003). The use of normative referenced databases has been an accepted

practice in the medical and scientific community and the advantage it brings to

neurofeedback is the allowance for the comparison of an individual to a norm-referenced

population, in terms of z-scores, to identify measures of aberrant EEG activity (Thatcher

& Lubar, 2009). This made possible the development of models, which focused more on

the individualized and unique needs of the client rather than a one-size-fits-all model.

Consequently, during the ensuing decade, the QNF model began taking hold in the NF

industry. However, the primary number of channels incorporated in the amplifiers of the

time was still limited to only two.

27

In 2006, the 4-channel – 4ZNF – technique was introduced. ZNF incorporates the

application of an age matched normative database to instantaneously compute z-scores,

via Joint Time Frequency Analysis (as opposed to the fast Fourier transform), making

possible a dynamic mix of both real-time assessment and operant conditioning

simultaneously (Collura et al., 2009; Thatcher, 2012). While the QNF of the 1990s held

as a common goal movement of the z-scores in the QEEG towards the mean, the advent

of ZNF brought with it the more frequent use of the term normalizing the QEEG or

normalization to refer to this process. It is now generally acknowledged that the term

normalization, when used to describe the process of ZNF, refers to the progression of the

z-scores towards the mean (i.e. z = 0), meaning that the NF trainee’s EEG is moving

closer to the EEG range of normal (i.e. typical) individuals of his/her age. But by 2009

the 4ZNF model was further enhanced to include the availability of up to all 19 electrode

sites in the International 10-20 system.

This surface potential 19ZNF greatly expands the number of scalp locations and

measures, including the ability to train real-time z-scores using various montages such as

linked-ears, averaged reference, and Laplacian, as well as simultaneous inclusion of all

connectivity measures such as coherence and phase lag. This, then, makes possible the

inclusion of all values from the database metrics for any given montage (as many as a

total of 5700 variables) in any protocol (Collura, et al., 2009). But the advent of 19ZNF

not only increases the number and types of metrics available to target, it also brought two

major changes to the landscape of NF. First, it established a new model wherein the

target of interest for the NF is the QEEG calculated z-scores of the various metrics

(frequency/power, coherence, etc.), rather than the amplitude of particular frequency

28

bands (theta, beta, etc.). Second, it changed the makeup of a typical NF session. In either

the conventional QNF model, or 4ZNF, the clinician will develop a protocol guided by

the QEEG findings, but will generally employ the same protocol settings repeatedly for

multiple NF sessions until the next assessment QEEG is scheduled. However with

19ZNF, in every session the clinician can acquire and process QEEG data, compare the

pre-session data to past session data, then design an individualized z-score normalization

protocol based on that day’s QEEG profile, and then perform a 19ZNF session, all within

an hour (Wigton, 2013). Thus, each 19ZNF session uses a protocol unique to the client’s

brainwave activity of that day, providing further tailoring of the NF to the individual

needs of the client, on a session-by-session basis. This, then, brought a new dynamic to

the normalization model of NF such that z-scores (rather than amplitude of frequencies)

could be targeted, on a global basis, so as to make possible a goal of normalizing all the

QEEG z-scores (when clinically appropriate) in the direction of z = 0.

How problem/gap of 19ZNF research evolved into current form. Over its

more than 40-year history NF has frequently been criticized as lacking credible research,

as evident by Loo and Barkley’s (2005) critique. Nevertheless, even Loo and Makeig

(2012) concede recently the research has improved. For example, Arns et al. (2009)

conducted the first comprehensive meta-analysis of NF, covering 1194 subjects,

concluding that it was both efficacious and specific as a treatment for ADHD, with large

to medium effect sizes for inattention and impulsivity, respectively. Then, in a research

review sponsored by the International Society for Neurofeedback and Research (ISNR),

in what is a comprehensive review of controlled studies of NF, Pigott et al. (2013)

evaluated 22 studies to conclude that NF meets the criteria of an evidence-based

29

treatment for ADHD. This review further documents that NF has been found to be

superior to various experimental group controls, shows equivalent effectiveness to

stimulant medication, and leads to sustained gains even after termination of treatment.

However, as encouraging as this body of research is, it is limited in that the model

covered by these studies is largely limited to one of the most traditional models of NF

(theta/beta ratio NF) and only addresses a single condition of ADHD. Missing from these

comprehensive reviews and meta-analyses are newer QNF models, which have been in

use since the 1990s, and are frequently employed for a wider range of disorders in

addition to ADHD. Yet, that is not to say that QNF is devoid of research. In fact, from

2002 to 2013 there are at least 20 studies in peer-reviewed literature covering the QNF

model, yet there is great diversity in the different conditions treated in these studies, as

well as a greater use of individualized, custom-designed protocols; hence making meta-

analysis of this collection of research less feasible. Nonetheless, these studies do

represent a body of research pointing to the efficacy of the QNF model.

Yet, when it comes to the newest models of surface ZNF, there is no such

collection of research in the literature. There exist only two studies (Collura et al., 2010;

Hammer et al., 2011) which evaluate sample groups of the 4ZNF model, and the Collura

et al. report is mostly descriptive in nature. This, then, leaves only one experimental

study. There is one dissertation on 4ZNF (Lucido, 2012), but it too is a single case study.

Regarding 19ZNF, as of this writing, there are only two peer-reviewed published

empirical reports specifically evaluating surface potential 19ZNF (Hallman, 2012; J. L.

Koberda et al., 2012b) and those are only case study in nature.

30

Yet, this is not to say the peer-reviewed literature landscape is entirely devoid of

any mention of surface ZNF models. Nevertheless, what does exist is mostly information

about the technique in the form of review articles (Collura, 2008; Stoller, 2011; Thatcher,

2013; Wigton, 2013), chapters in edited books (Collura et al. 2009; Wigton, 2009), as

well as numerous qualitative oral and poster conference presentations since 2008. Of note

is a recent poster presentation (Wigton & Krigbaum, 2012), with a later expansion of that

work (Krigbaum & Wigton, 2013), which was a multicase empirical investigation of

19ZNF; however it primarily focused on a proposed research methodology for assessing

the degree of z-scores progression towards the mean. There also exist anecdotal

observations in the form of case reports in non-peer-reviewed publications and internet

website postings. Yet, while anecdotal observations and information from review and

case study reports are helpful for initial appraisals of a new model, quantitative statistical

analysis is needed to validate theories born of early qualitative evaluations, to counter a

lack of acceptance from the wider neuroscience community.

Much of the focus of discussions of 19ZNF is on the potential for good clinical

outcomes in fewer sessions than traditional NF (J. L. Koberda et al., 2012a; Rutter, 2011;

Thatcher, 2013; Wigton, 2009; Wigton, 2013). Though, before the question of number of

sessions is examined, first there should be an establishment of the efficacy of this

emerging model; because empirical studies evaluating the efficacy of 19ZNF are absent

from the literature. This dissertation was intended to fill this gap of knowledge, by

analyzing the efficacy of 19ZNF in a clinical setting.

31

Theoretical Foundations

Foundations of EEG and QEEG. Hughes and John (1999) discussed a decade-

long history, inclusive of over 500 EEG and QEEG related reports, the findings of which

indicate that cortical homeostatic systems underlie the regulation of the EEG power

spectrum, that there is a stable characteristic in healthy humans (both for age and cross-

culturally), and that the EEG/QEEG measures are sensitive to psychiatric disorders.

These factors made possible the application of Gaussian-derived normative data to the

QEEG metrics such that these measures are independent of ethnic or cultural factors,

which allow objective brain function assessment in humans of any background, origin, or

age. As a result, Hughes and John assert when using artifact-free QEEG data, the

probability of false positive findings are below that which would be expected by chance

at a p value of .0025. Thus, changes in the QEEG values would not be expected to occur

by chance, nor is there a likelihood of a regression to the mean of QEEG derived z-scores

because EEG measures, and the corresponding QEEG values, are not random. Since the

work of Hughes and John, well over a decade ago, there have been numerous studies

published in the literature further demonstrating the reliability and validity of QEEGs

(Cannon et al., 2012; Corsi-Cabrera, Galindo-Vilchis, del-Río-Portilla, Arce, & Ramos-

Loyo, 2007; Hammond, 2010; Thatcher, 2012; Thatcher & Lubar, 2009).

Learning theory as applied to NF. As has been stated, NF is also frequently

referred to as EEG biofeedback, and biofeedback has been defined simply as the process

whereby an individual learns how to change physiological activity (AAPB, 2011). As

Demos (2005) asserted, biofeedback is a two-way model such that 1) the physiologic

activity of interest is recorded, and 2) reinforcement is provided each time the activity

32

occurs; as a result, voluntary control of the targeted physiologic activity becomes

possible. On the surface this is a basic descriptor of operant conditioning. As a result, a

common practice in the literature is for NF to be referred to only as an operant

conditioning technique. However, the theoretical frameworks of NF are more correctly

framed as encompassing both classical and operant conditioning mechanisms (Collura,

2014; Sherlin, Arns, Lubar, Heinrich, Kerson, Strehl, & Sterman., 2011; Thatcher, 2012;

M. Thompson & Thompson, 2003). Operant conditioning – as first conceptualized by

Edward Thorndike (1911) with the Law of Effect, which holds that satisfying rewards

strengthens behavior, and as further advanced by B. F. Skinner (1953) – has as its

primary principle when an event is reinforced/rewarded it is likely to reoccur

(Hergenhahn, 2009); and for Skinner the reinforcer is anything that has contingency to a

response. It is important to note that operant conditioning relates to the learning of

volitionally controlled responses, motivation is necessary, and rewards need to be desired

or meaningful (M. Thompson & Thompson, 2003).

In contrast, classical conditioning, established by Ivan Pavlov (1928), differs in

that it deals with learning of reflexive or autonomic nervous responses. The primary

mechanism is based in the associative principles of contiguity and frequency such that the

presence of a dog’s food, which naturally elicits a salivation reflex, when paired

(contiguity) with a bell, repeatedly (frequency), will lead to the dog salivating upon the

presentation of only the bell (Hergenhahn, 2009). Thus, the pairing of two previously

unpaired events results in automatic learning in the form of classical conditioning. Yet, it

is important to note that while operant conditioning involves volitionally oriented

behavior modification, NF is a learning process which occurs largely outside of

33

conscious awareness; in essence, an implicit learning process (Collura, 2014). As applied

to NF, the change in the EEG, as reflected in brainwave frequencies, patterns, or z-scores,

is the behavior which is modified as a result of the combined classical and operant

conditioning occurring in the NF session (Collura, 2014).

In this context then, successful NF involves a motivated trainee experiencing the

repeated pairing of meaningful auditory and/or visual reward signals, when the recorded

brainwaves fall in a targeted range. The reward signal is typically in the form of an

auditory tone (beep, chime, music) in combination with an animated visual display

(simple game-like displays or movies), which when aesthetically pleasant to the trainee

enhances and promotes the process. Some have noted the importance of additional

learning theory components such as shaping (Collura, 2014; Sherlin et al., 2011; M.

Thompson & Thompson, 2003), anticipation of future rewards (Thatcher, 2012), and

secondary reinforcers (Sherlin et al., 2011; M. Thompson & Thompson, 2003) to further

inform NF to varying degrees. These variations as applied to NF have served to generate

a range of NF models over the years; however the basic foundations of classical/operant

conditioning remain constant in all the models.

Traditional/amplitude-based models of NF. In NF, when the EEG is divided

into different frequency bands (alpha, beta, etc) the amplitude measures how much of that

frequency is present within the total EEG spectrum recording. The basic goal of

amplitude NF treatment models is to either increase or decrease the amplitude of a

particular frequency. These models are the longest-standing conceptualization of NF

techniques and for that reason, for the purposes herein, the term traditional will be used

to refer to these models of NF. The earliest traditional model of NF started with Kamiya’s

34

(1968) discovery in the early 1960s that human alpha waves could be increased and

trained to occur for increased periods of time. Next, Sterman and Fiar (1972) followed up

on this work by expanding the training Sterman had been conducting with cats to include

humans, with the first known case of resolving a seizure disorder in a person using NF. In

this model the goal was to increase the beta frequency of 12-15 Hz, also referred to as

sensorimotor rhythm (SMR), along the sensorimotor cortex of the brain. Others then

expanded on this model. For example, Lubar believed the model Sterman developed

would be applicable to children with attention disorders (Robbins, 2000). After a year-

long academic fellowship with Sterman, he moved on to develop his own model which

incorporated decreasing the theta frequency in addition to increasing beta (Robbins,

2000). Lubar and Shouse (1976) reported on the first use of this approach, which was the

foundation for what would become one of the most commonly reported and researched

protocols (for use with attention disorders) in the literature since the early 1990s; that of

the theta/beta ratio model.

Another example of a traditional NF model with roots to Sterman’s efforts is the

Othmer model (S. Othmer, Othmer, & Kaiser, 1999), employing a combination of

increasing beta (either 12-15 Hz or 15-18Hz) together with decreasing theta (4-7 Hz), and

a higher beta band (22-30 Hz); again with electrode placements primarily along the

sensorimotor cortex locations of the scalp. In the years since its introduction, there have

been different modifications and variations of the Othmer approach (S. F. Othmer &

Othmer, 2007). Nevertheless, consistent with traditional NF, this model makes use of

targeting the amplitudes of frequency bands in particular directions (i.e. make more or

less of targeted frequencies).

35

While some built models based in the original findings of Sterman, others

expanded on Kamiya’s work, by developing models which targeted the increase of alpha

and/or theta frequencies (in parietal brain regions) to enhance relaxation and creative

states (Budzynski, 1999). Peniston and Kullkosky (1990, 1991) developed applications of

these approaches, which led to treatment models for alcoholism and posttraumatic stress

disorders. Yet still others, such as Baehr, Rosenfeld, and Baehr (1997), established

protocols targeted to balance alpha in the frontal regions as a treatment for depression.

While each of the above models targeted different frequencies with a variety of

protocols, consistent was a focus on changing the amount of the brainwave of interest;

the desired outcome is either greater or lesser amplitude of a target frequency. Moreover,

pre-treatment assessment of EEG activity to inform NF protocols is limited to nonexistent

in the majority of these models, with a typical one-size-fits-all approach. While selecting

the particular NF model for a treatment approach (i.e. theta-beta ratio versus alpha-theta

training) is informed by the presenting symptoms of each case, personalizing a NF

protocol to address the individual brainwave patterns of the client is not the focus of these

approaches.

QNF model of NF. A key focus of QNF is precisely tailoring the NF protocol,

based on the individual EEG baseline and symptom status of the client, as determined by

the QEEG, in conjunction with clinical history and presenting symptoms (Arns et al.,

2012). The primary premise of this approach is that localized cortical dysfunctions, or

dysfunctional connectivity between localized cortical areas, correspond with a variety of

mental disorders and presenting symptoms (Coben & Myers, 2010; Collura, 2010;

Walker, 2010a). When the EEG record of an individual is then compared to a normative

36

database representing a sample of healthy individuals, the resulting outlier data

(deviations of z-scores from the mean) help link clinical symptoms to brain dysregulation

(Thatcher, 2013). For example, when an excess of higher beta frequencies are found, the

typical associated symptoms include irritability, anxiety, and a lowered frustration/stress

tolerance (Walker, 2010a).

The conceptual framework of the stability of QEEG, as noted above, applies to

QNF in that a stable EEG is not expected to change without any intervention, thus the

changes seen as a result of QNF is not occurring by chance, but due to the operant

conditioning of the brainwaves as a result of the NF process (Thatcher, 2012). Therefore,

in the example of excess beta frequencies, when the symptoms of anxiety and irritability

are resolved after QNF, and the post QEEG shows the beta frequencies to be reduced

(closer to the mean), it is assumed the improvement in symptoms is due to the change in

the QEEG; thus representing improved electrocortical functioning (Arns et al., 2012;

Walker, 2010a). The term for this process, which has arisen secondary to QNF, is

generally referred to as normalization of the QEEG, or simply normalization (Collura,

2008; Surmeli & Ertem, 2009; Walker, 2010a). Consequently, the concept of

normalization is generally accepted to be when the z-scores of the QEEG move towards

the mean (i.e. z = 0).

It is also important to note that the QNF model, with its reliance on the QEEG to

guide the NF protocol, embraces the heterogeneity of QEEG patterns as discussed by

Hammond (2010). In understanding that a particular clinical symptom presentation may

be related to varied deviations in the QEEG, it quickly becomes apparent that each NF

protocol needs to be personalized to the client; as well as monitored and modified for

37

maximum treatment effect (Surmeli et al., 2012). This, then, results in different

electrophysiological presentations being treated differently, even if the overarching

diagnosis is the same. This clinical approach is supported through multiple reports in the

literature discussing how training the deviant z-scores towards the mean (i.e. normalize

the QEEG) in QNF results in the greatest clinical benefit (Arns et al., 2012; Breteler et

al., 2010; Collura, 2008; Orgim & Kestad, 2013; Surmeli et al., 2013; Surmeli & Ertem,

2009, 2010; Walker, 2009. 2010a, 2011, 2012a).

However, while the personalization of NF protocols aids in greater specificity in

client treatment, it creates methodological challenges for researching QEEG based NF

models; which will be discussed further below. When boiling down the elements of study

to a lowest common denominator, overall normalization of the QEEG is the only

common point of measurement. Therefore a reasonable tool, as a measure of change in

the QEEG, would be a value reflecting the change of targeted z-scores for a particular

metric.

In summary then, in the normalization model of QNF, when the QEEG data show

excessive deviations of z-scores, and those deviations correspond to the clinical picture,

the NF protocol is targeted to train the amplitude of the frequency in the direction of the

mean (i.e. create more or less energy within a specified frequency band). In other words,

if the QEEG indicates an excess of a beta frequency (i.e. high z-scores), and the

presenting symptoms are expected with that pattern (i.e. anxiety), the protocol would be

designed to decrease the amplitude of that beta frequency. Conversely, if the QEEG

indicates a deficit of an alpha frequency, with corresponding symptoms, the protocol

would be designed to increase the amplitude of the alpha frequency. The QNF model

38

then, is simply traditional amplitude based NF using the QEEG to guide the protocol

development for the NF sessions.

ZNF model of NF. The ZNF model leverages the statistical underpinnings of a

normal distribution, where a value converted to a z-score is a measure of the distance

from the mean of a population, such that the mean represents a range considered to be

normal (or typical) (Collura, 2014). With ZNF the real-time QEEG metrics are

incorporated into the NF session using a joint time frequency analysis (rather than fast

Fourier transform) to produce instantaneous z-scores, which allows for real-time QEEG

assessment to be paired with operant conditioning (Collura, 2014; Thatcher, 2013).

Therefore, where the QNF model has amplitude (as guided by the QEEG) as its targeted

metric, in its most basic form, the ZNF model targets the calculated real-time z-scores.

Yet, that being said, it is important to note that the z-scores can be considered a meta-

component of EEG metrics (i.e. amplitude or connectivity) and ultimately, even when z-

scores are targeted, the underlying EEG components are still being trained.

Nevertheless, directly targeting z-scores results in a different dynamic in the NF

training protocol. The goal is no longer to simply make more or less frequency amplitude,

but for the targeted excessive z-score metrics (whether high or low) to move towards the

mean, that is to normalize. Thus, there is a greater focus on the construct of

normalization. A second change is the inclusion of many more metrics to target. ZNF

makes available simultaneously, for up to ten frequency bands, both absolute and relative

power, ratios between frequencies (i.e. theta/beta ratio or alpha/beta ratio), as well as the

inclusion of connectivity metrics such as asymmetry, coherence, or phase lag, all as

active training metrics. Therefore, when applied to 4ZNF, the maximum number of

39

metrics to train is 248 (Collura, 2014) and, within the scope of the 19ZNF the maximum

number of metrics is 5700 (Collura, et al., 2009). These changes make the entire range of

all QEEG metrics, or a subset of selected metrics, available for targeting with ZNF

models. Moreover, the increased number of metrics targeted by 19ZNF may allow for an

increase in regulation and synchronization of neural activity simply by the greater

number of training variables. Nonetheless, one consistent theme remains aligned with the

QNF model, in that the decision to target normalization of QEEG metrics is determined

by the presenting clinical symptoms; thus when QEEG deviations correspond to

presenting symptoms, normalization is a reasonable treatment goal.

In asking if the 19ZNF improves attention, behavior, executive function, or

electrocortical function, the research questions for this study add to what is known

regarding whether operant conditioning with 19ZNF, produces clinical results that are

comparable to those reported in the literature for traditional or QNF models. Moreover,

this study also evaluates questions regarding 19ZNF and normalization of QEEG metrics.

This research fits within the overarching NF model with a specific focus on evaluating

efficacy of the ZNF model. As has been demonstrated in the literature, traditional NF is

well researched (Arns et al., 2009; Pigott et al., 2013), and as will be discussed in the next

section, the QNF model is well addressed in the literature. Conversely, as will be seen,

the ZNF models (4ZNF and 19ZNF) are still minimally represented in the literature.

Therefore, this study addresses an area which calls for further research.

Review of the Literature – Key Themes

QNF in the literature. Beginning with QNF models in reviewing the NF

literature is applicable in that the QNF model laid the ground-work for the ZNF models

40

that followed. Both QNF and ZNF models hold the generalized goal of normalizing the

QEEG, and for that reason, QNF is chosen as the first key theme in reviewing NF in the

literature. With few exceptions, literature presented on the QNF model comes from

research conducted in clinical settings. As a result, given the ethical constraints of

conducting research in clinical settings (e.g. asking clients to accept sham or placebo

conditions) (Gevensleben et al., 2012) few are blinded and/or randomized-controlled

studies.

Arns et al. (2012) conducted a well-designed open-label study of 21 ADHD

subjects using the QNF model, incorporating pre-post outcome measures and QEEG data.

The purpose was to investigate if the personalized medicine approach of QNF was more

efficacious (as defined by effect size) for ADHD than the traditional theta/beta or slow

cortical potential models, as reported in his meta-analysis three years earlier (Arns, et al.,

2009). The outcome measures incorporated were a self-report scale based on the

Diagnostic and Statistical Manual-IV list of symptoms and the Beck Depression

Inventory. The findings of this study were statistically significant improvements (p ≤

.003) in both the attention (ATT) and hyperactivity (HI) subtypes of ADHD symptoms as

well as depression symptoms. In this study, the mean number of sessions was 33.6, and

the effect size was 1.8 for the ATT subtype, and 1.2 for the HI subtype; this was a

substantial increase over the traditional model effect sizes of 1.0 (ATT) and 0.7 (HI)

respectively. This suggests the QNF model is more efficacious (i.e. effect size of clinical

improvements) than the older traditional theta/beta or slow cortical potential models.

Furthermore, in this study, non-z-score EEG microvolt data was reported for only nine

frontal and central region electrode sites, and three frequency bands, on a pre-post basis.

41

In addition to that the protocols employed are described as a selection of one of five

standard protocols, with QEEG informed modifications. The limitations of this study

were few but include a lack of a control group, a fairly small sample size, and that some

outcome measures were collected on only a sub-group of participants (thus reducing net

sample size). Moreover the pre-post QEEG data analysis was limited.

J. L. Koberda, Hillier, et al., (2012) reported on the use of QNF in a clinical

setting of a neurology private practice. All 25 participants were treated with at least 20

sessions of a single-channel traditional NF protocol, which was guided by QEEG data

and symptoms, with a goal to improve symptoms and normalize the QEEG. Clinical

improvement was measured by subjective reports from the participants in the categories

of not sure (n = 4), mild if any (n = 1), mild improvement (n = 3), improved/improvement

(n = 13), much improved (n = 2), and major improvement (n = 2); with a total of 84% (n

= 21) reporting some degree of improvement. QEEG change was reported as a clinical

subjective estimation (based on visual inspection of the QEEG topographic images) of

change in the targeted frequencies, in the categories of no major change/no improvement

(n = 6), mild improvement (n = 9), improvement (n = 8), or marked improvement (n = 1),

and one participant not interested in post-QEEG; with a total of 75% (n = 18) showing

estimation of improvement in the QEEG. Of note with this study was the heterogeneous

collection of symptoms treated which included ADD/ADHD, anxiety, autism spectrum,

behavior symptoms, cognitive symptoms, depression, fibromyalgia, headaches, major

traumatic brain injury, pain, seizures, stroke, and tremor, in varying degrees of

comorbidity per case. However, the primary limitation of this study was the loosely

42

defined subjective estimations of improvement for both clinical symptoms and QEEG

outcomes.

In their randomized control study, Breteler et al. (2010) evaluated QNF as an

additional treatment with a linguistic education program. From the total sample of 19, ten

participants were in the NF group and nine were in the control group. Individual NF

protocols were based on QEEG results and four rules, with a generally (though not

strictly adhered to) 1.5 z-score cutoff; which resulted in the use of eight personalized

protocols. Improvement was determined by results of outcome measures of various

reading and spelling tests, as well as computerized neuropsychological tests. Paired t tests

were applied for analysis of the difference values between the pre and post scores. The

reported findings showed the NF group improved spelling scores with a very large

Cohen’s d effect size of 3; however no improvement in reading or neuropsychological

scores. QEEG data was reported, in terms of pre-post z-scores, on an individual basis (i.e.

per each case) for a limited number of targeted sites, frequencies, and coherence pairs;

with most showing statistically significant normalization.

In a retrospective study using archived clinical case files, Huang-Storms,

Bodenhamer-Davis, Davis, and Dunn (2006) evaluated the efficacy of QNF for 20

adopted children with a history of abuse who also had behavioral, emotional, social, and

cognitive problems. The children all received 30 sessions of NF (from a private practice

setting) with QNF protocols, which were individualized based on the QEEG profiles.

Data from the files of 20 subjects were collected to include pre and post scores for

outcome measures from a behavioral rating scale (Child Behavior Checklist; CBCL), and

a computerized performance test (Test of Variables of Attention; TOVA). The findings

43

for the CBCL were statistically significant (p < .05) for most scales and the TOVA

findings were statistically significant (p < .05) for three scales, thus demonstrating QNF

efficacy for the subjects in this study. There was no quantified QEEG reported; only

observations of general trends in the pretreatment QEEG findings, such as excess slow

waves in frontal and/or central areas.

Two researchers are most notable for several published studies evaluating the

QNF model, that being Walker and then Surmeli and colleagues. Each has a particular

consistent style in structuring their studies; and both have reported on the use of QNF

with a wide variety of clinical conditions. Therefore their works will be reviewed in a

grouping format. Walker has reported on mild closed head injury (Walker, Norman, &

Weber, 2002), anxiety associated with posttraumatic stress (Walker, 2009), migraine

headaches (Walker, 2011), enuresis (Walker, 2012a), dysgraphia (Walker, 2012b), and

anger control disorder (Walker, 2013). His QNF protocol development centers on

tailoring the protocol to the individual clinical QEEG data, with some restrictions of

either increasing or decreasing the amplitude of certain frequency ranges. For example,

the protocols for the anger outburst study restricted the target range to decrease only

excess z-scores of beta frequencies, combined with decreasing excess z-scores of 1-10 Hz

frequencies. For the migraine and anxiety/posttraumatic stress studies both were based on

individual excess z-score values found in the beta frequencies in a range of 21-30 Hz (to

decrease) with an addition of increasing 10 Hz. For all studies the electrode sites selected

were ones where the deviant z-scores in the targeted range were found. In the mild closed

head injury article, the protocol was different because the study was meant to evaluate

coherence training with a stated goal to normalize coherence z-scores. Thus, the most

44

deviant coherence pair was selected first (for five sessions each) and, then progressed to

lesser deviant pairs until the symptoms resolved or until 40 sessions were completed.

None of Walker’s reports declare a particular research design; still all involve pretest-

posttest comparisons of various outcome measures.

The outcome measures that Walker typically employs are primarily Likert or

percentage-based self-reports, except in the anger control disorder study where the

DeFoore Anger Scale self-report instrument was used to track the number of anger

outbursts. However, while all protocols are personalized, and based on QEEG findings,

there are no quantified pre-post QEEG data used as an outcome measure, and none are

reported in his studies. Overall the findings of all of Walker’s studies show improvements

in the targeted clinical conditions. In the mild closed head injury study, with an n = 26,

84% of the participants reported greater than 50% improvement in symptoms. For the

anxiety/post-traumatic stress article, with an n = 19, all improved on a Likert scale (1 –

10; 10 being worst) from an average rating of 6 before NF treatment to an average rating

of 1 after NF treatment. With the migraine study, where 46 NF participants were

compared to 25 patients who chose to remain on medication, 54% had complete

remission of headaches, 39% had a greater than 50% reduction, and 4% experienced less

than 50% reduction in migraines, all in the NF group, while in the medication group, 84%

had no change in migraines and only 8% had a greater than 50% reduction in headaches.

In three of his more recent studies, for the enuresis (n = 11), dysgraphia (n = 24), and

anger control research (n = 46), Walker reported all findings for all participants (in all

three studies) showed statistically significant improvement at p < .001.

45

Surmeli and colleagues reported on Down syndrome (Surmeli & Ertem, 2007),

personality disorders (Surmeli & Ertem, 2009), mental retardation (Surmeli & Ertem,

2010), obsessive compulsive disorder (Surmeli & Ertem, 2011), and schizophrenia

(Surmeli et al., 2012). Notable in this collection of work are conditions previously not

known to respond to NF, such as personality disorders, mental retardation, Down

syndrome, and schizophrenia. All of these studies report the QNF protocol as being

individualized, as informed by a combination of the QEEG findings and clinical

judgment; with an overall goal to normalize the QEEG patterns. Notable for most of

Surmeli et al. studies are a high number of sessions reported for the cases; ranging from

an average of 50 to an average of 120 sessions. No particular research design is declared

in the Surmeli et al. studies, but here too, comparisons of pretest-posttest outcome

measures are reported.

The outcome measures in the studies mentioned above generally make use of

clinical assessment instruments designed to measure the symptoms targeted for the QNF

treatment. For example, the schizophrenia study employed the Positive and Negative

Syndrome Scale (PANSS), and for the obsessive compulsive disorder research they

incorporated the Yale-Brown Obsessive-Compulsive Scale (Y-BOCS). For many studies,

the computerized performance Test of Variable Attention (TOVA) was used. Yet, as with

Walker’s work, in spite of all protocols being individually QEEG-guided, QEEG data is

not used or reported as an outcome measure; only observations of general trends of the

changes in QEEGs are discussed. However, the targeted clinical symptoms, as measured

by the clinical assessments, were reported as having statistically significant improvement

in all studies. For the personality disorder study, with an n = 13, twelve were significantly

46

improved on all outcome measures; with the Symptom Assessment 45 Questionnaire at p

= .002, the Minnesota Multiphasic Personality Inventory (MMPI) Psychopathy scale at p

= .000, and the TOVA at p < .05 on the visual and auditory impulsivity scales. With the

article reporting the study with mentally retarded participants, including an n = 23, for 19

there was improvement on the Wechsler Intelligence Scale for Children-Revised (Verbal

scale, p = .034; Performance scale, p = .000; Total scale, p = .000) and the TOVA

(Auditory and Visual Omission scale, p < .02; Auditory and Visual Commission scale, p

< .03; Auditory and Visual Response Time Variability scale, p < .03). In the Down

syndrome study, while the outcome measure was not a commercialized assessment, they

did develop a questionnaire formulated to evaluate symptoms associated with Down

syndrome. The findings were that all subjects in the study (n = 7) showed improvement at

p < .02 on all questionnaire scales. With QNF for obsessive compulsive disorder, with an

n = 36, 33 showed improvement on the Y-BOCS (Obsession subscale, Compulsion

subscale, and Total score all p < .01). Finally, in the schizophrenia study, with an n = 51,

47 out of 48 patients who completed pre and post PANSS improved on all scales at p <

.01. Moreover of the 33 who were able to complete the MMPI, findings showed

significant improvements (p < .01) on the scales of Schizophrenia, Paranoia,

Psychopathic Deviation, and Depression.

This review of QNF research fits within this dissertation topic as examples of how

prior studies with QEEG data have been addressed in the literature. As can be seen,

studies evaluating QNF are typically found in clinical settings, with a wide variety of

clinical symptoms and/or mental health diagnoses, and frequently have relatively small

sample sizes. Moreover the NF protocols employed typically are tailored to the

47

individual, informed by QEEG, with a goal to normalize the QEEG. The overwhelming

majority of clinical QNF research employs retrospective pre-post comparison research

designs and the outcome measures used are tied to the symptoms of investigation. Yet

few, if any, report pre-post QEEG metrics, and only one (Arns et al., 2012) incorporated

statistical analysis of QEEG metrics as an outcome measure (and that was to a limited

degree). Therefore, in the QNF literature, it has become an accepted practice to define

efficacy in terms of measuring symptom improvement with various clinical assessments

(both commercially and informally developed). Nevertheless, clearly there is a gap in the

reporting of group QEEG z-score mean data in the present QNF research.

4ZNF in the literature. Given that 4ZNF is the forerunner to 19ZNF, this topic is

explored to provide historical context on both its development and its coverage in the

literature. While there are numerous studies in the literature for QNF, when it comes to

ZNF studies, such is not the case. However, for the 4ZNF model there are four

representations of 4ZNF clinical results in the literature.

In a first poster presentation on the topic, Wigton (2008) presented a single case

study where 4ZNF was used with an adult to address a diagnostic history of ADHD,

Bipolar disorder, and anxiety symptoms. The primary pre-post outcome measure was the

IVA. Also included were topographic images of pre and post QEEG assessments. After

25 sessions of 4ZNF, in addition to multiple subjective reports of symptom improvement

from the participant, the scaled scores for the IVA showed marked improvement. The full

scale Response Control scale improved from 29 to 94, and the full scale Attention scale

from 0 to 96. The QEEG findings (as reported by visual presentation of QEEG

topographic images) showed improvements in terms of normalization in the QEEG, most

48

noticeably in the left frontal delta and theta frequencies, as well as coherence and phase

lag normalization. However, a limitation of this study was a lack of statistical analysis of

pre-post QEEG data and the use of only one clinical assessment for outcome measures.

Collura et al. (2010) was the first peer-review publication addressing 4ZNF

although its organization was a loosely structured collection of clinical reports from six

clinicians covering 24 successful cases. Nonetheless, for a model with little scientific

evidence, it does stand as the only representation in the literature of a multiple-clinician

report of clinical results with 4ZNF. All cases reported clinical improvement, with no

abreactions, and the average number of sessions for all cases presented were 21.1. The

limitations of this case study are the lack of a structured methodology, no statistical

analysis, and limited pre-post outcome measures and/or QEEG data.

The study conducted by Hammer et al. (2011) represents, to-date, the only

quantitative analysis of 4ZNF. Its strength is a sound methodology with a randomized,

parallel group, single-blind design, together with QEEG z-scores as an outcome measure.

Though, the setting for this research was not in a clinical setting, but rather a university

psychophysiology laboratory wherein participants were recruited specifically for the

study. The purpose was to both explore 4ZNF as a new NF model, and to evaluate the

efficacy of two different 4ZNF protocols for insomnia. The primary findings suggest that

4ZNF may be a beneficial treatment for insomnia. While this study had very small group

sample sizes (n = 5 and n = 3) all insomnia related outcome measures resulted in pre-post

treatment improvement in symptoms, and normal (or near normal) sleep was achieved by

all participants. Moreover, at follow-up 6 to 9 months after treatment, over half sustained

the treatment response. The findings of this study included QEEG measures showing

49

statistically significant electrocortical change occurred for the delta frequency (p < .001)

and beta frequency (p < .01), but not high beta (p < .11). However, a limitation is that the

reported findings only included three frequencies, and the absolute and relative power z-

scores were combined in the analysis; therefore a more discrete picture of overall QEEG

normalization was not available. Further limitations of this study were the small sample

size and the lack of control group. Yet this study does stand alone, being a peer-reviewed

publication, as an example of a quantitative methodology for measuring normalization of

QEEG z-scores with the binomial test of significance, with the 4ZNF model.

A dissertation conducted by Lucido (2012) was a single case study to evaluate the

use of 4ZNF for an adult with Autism spectrum condition (ASC). This study used a

multiple baseline design, with five rounds of assessment data gathered before the 4ZNF

sessions, and a round of assessments at five incremental points during/after the NF

treatment. The outcome measures employed were the Neuropsych Questionnaire, the

CNS Vital Signs computerized neurocognitive assessment, and the Test of Nonverbal

Intelligence. While QEEG data was gathered and purported as an outcome measure, only

limited pre-post colorized topographic images were provided as a means to demonstrate

generalized changes in QEEG metrics. The results were that, with only one exception

(cognitive processing speed), all symptoms assessed with the outcome measures

improved. These included ASC symptoms, executive function, depression, anxiety, mood

stability, attention, and intelligence. To the study’s credit, this was a well-designed, well-

controlled case study; however still a representation of a single case, nonetheless.

Overall, the 4ZNF model is poorly represented in the NF literature. However,

there are still themes relevant to this dissertation. Of the studies reported, most are from

50

clinical settings. Moreover, clinical assessments, as outcome measures, are used in all

studies. A particular stand out, though, is the Hammer et al. (2011) research, wherein

statistical analysis of QEEG metrics was used as an outcome measure.

19ZNF in the literature. With 19ZNF being the focus of this study, reviewing

what literature is available is necessary. Yet, there is an even greater dearth of published

literature for 19ZNF than 4ZNF. Therefore a review of conference oral and poster

presentations is necessary to sufficiently address what is known regarding 19ZNF.

Moreover, the literature reviewed herein is restricted to evaluative and/or case study

research reports regarding clinical applications of 19ZNF (rather than technical reviews

of 19ZNF).

In a first published clinical review of 19ZNF, Wigton (2009) reported initial

findings in which substantial QEEG normalization and clinical improvement was

achieved in as little as three sessions. While research into this technique was clearly

needed, the degree of success achieved in just a few sessions was a novel finding for

previously known NF models. Later in a conference presentation, Wigton (2010a)

reported on a series of case reviews that employed the Laplacian montage with 19ZNF.

There were 10 cases which included conditions such as anger issues, anxiety, ADHD, and

impaired cognition. The findings were that 19ZNF led to clinical improvements and

QEEG normalization, in less than 10 sessions, in seven out of the 10 cases. In this

presentation outcome measures included the IVA, the DSMD, and Likert scale reports. A

year later Rutter (2011) described, in a conference presentation, her use of 19ZNF and

how she was able to see initial indications of QEEG normalization in as little as five

sessions.

51

In their conference oral presentation, J. L. Koberda, et al. (2012a) reported on a

comparison between 25 cases using traditional 1-channel NF and a mixed pool of 15

cases using either surface 19ZNF or LORETA ZNF. However, it is not clear how many

were 19ZNF and how many were LORETA ZNF cases. In this presentation the clinical

symptoms addressed in the 15 cases was varied and included anxiety, headaches, chronic

pain, cognitive and behavioral disorders, as well as focal neurological disorders. The

essential finding of this presentation was that both the traditional single-channel NF and

the 19ZNF/LORETA ZNF lead to improvement in clinical symptoms and improvements

in QEEG measures, but the 19ZNF/LORETA ZNF did so in fewer sessions. The

traditional NF group showed subjective self-report improvements of 84% and an

improvement of 75% of QEEG improvements, whereas the 19ZNF/LORETA ZNF group

showed 95% subjective improvement and 62.5% improvement in QEEG measures.

However an operationalized definition of these improvements was not clearly described

or quantified; nor were there any follow-up data reported. Nevertheless, the number of

sessions for the traditional NF was at least 20, whereas the number for the

19ZNF/LORETA ZNF group was an average of nine sessions.

Hallman (2012) presents a qualitative style clinical review of a single case study,

of a child with fetal alcohol syndrome. The purpose of the article was to describe the case

wherein 80 sessions of 19ZNF resulted in unexpectedly remarkable symptom and

behavior improvements. Moreover, the topographic images of pre-post QEEG data also

showed almost complete normalization; still there was no quantified measurement or

statistical analysis of QEEG data. There also were only subjective parental reports and no

outcome measures to quantify degree of symptom improvement.

52

J. L. Koberda et al. (2012b) also conducted a single case study, of a 23 year-old

male, for the purpose of reporting clinical outcomes using two types of 19ZNF (surface

and LORETA). After only 15 sessions, improvements in a cognitive assessment outcome

measure were achieved, still there were no inferential statistical analysis reported for the

pre-post outcome measures. Moreover, the use of two distinctly different 19ZNF

modalities (surface and LORETA ZNF) makes it hard to know if one better accounted for

the improvement over the other. Finally, while improvements in QEEG data were

reported, again no inferential statistical analyses of these improvements were presented.

Krigbaum and Wigton (2013) present findings for 10 cases with 19ZNF. This

study is notable in that it introduced a proposed methodology for statistically

demonstrating z-score progression towards the mean (i.e. z = 0), and an approach for

plotting individual learning curves as a result of 19ZNF. Additionally, cases in the study

included outcome measures such as the IVA, DSMD, BRIEF and Likert scale (reported

on a supplementary basis, with only an indication of improvement or not), and all

outcome measures showed improvement at case completion. Repeated measures analysis

of variance (rANOVA) and paired t tests supported all three research questions such that

the z-scores progressed towards the mean (rANOVA absolute power, p < .001; relative

power, p < .04; coherence, p < .001); the post z-scores were closer to the mean than the

pre z-scores (paired t test absolute power, p < .007; relative power, p < .05; coherence, p

< .03); and clinical improvement was reported in all cases. However, no follow-up data

was reported.

Clearly, the research evaluating 19ZNF is in its infancy and there is a great need

for scientifically sound investigations. More so, the research needs to move beyond

53

clinical reviews and case studies. As is incorporated in QNF research, use of clinical

assessments as outcome measures are important elements; additionally, finding ways to

include QEEG metrics as outcome measures would benefit 19ZNF research.

Outcome measures for ZNF research. This topic is included to explore outcome

measures that are suitable for ZNF research. A good deal of NF research occurs in

clinical settings, where assessment instruments are employed as part of the case workup.

As such, the use of those same measures after treatment is a natural fit for what are

frequently pretest-posttest research frameworks. Other than informal self-reports (i.e.

Likert scales) two types of popular outcome measures found in the NF literature are

rating-scale type assessments and computerized performance tests. Moreover, commonly

found in NF studies is the use of multiple outcome measures. Further, while the use of

EEG metrics as outcome measures of electrocortical change are infrequently incorporated

in NF research, there are a few reports in the literature which will be reviewed.

Computerized performance tests. Computerized performance tests are common

outcome measures in NF research, usually as a means to evaluate attention-related

symptoms associated with ADHD. One of those instruments is the IVA. While the IVA

was designed as a diagnostic aid for ADHD, the manual provides usage indications to

include assessing self-control and attention problems related to other disorders such as

depression, anxiety, head injuries, dementia, and other medical problems (Sanford &

Turner, 2009). Several NF studies have incorporated the IVA as an outcome measure to

assess attention related symptoms.

In their study to evaluate NF in a nonclinical group of college students’ cognitive

abilities, Fritson, Wadkins, Gerdes, and Hof (2008) used the IVA as one of their outcome

54

measures; each group (experimental and control) had an n = 16. The stated objective was

to determine effects of NF on attention, impulsivity, mood, intellectual functioning,

emotional intelligence, and general self-efficacy. The IVA was one of several outcome

measures and was included to assess response control (i.e. impulsivity) and attention. The

researchers reported results in terms of the means and standard deviations of pre-post

values of eight of the primary scales of the instrument. The statistical analysis performed

were multivariate analysis of variance (MANOVA) between the control and experimental

groups.

In evaluating the utility of the Tower of London test, as a suitable assessment

instrument for clients with Asperger’s who undergo NF, Knezevic, Thompson, and

Thompson (2010) employed the IVA as one of the outcome measures. They included six

scales of the IVA (Auditory and Visual Prudence, Auditory and Visual Vigilance, and

Auditory and Visual Speed) to assess the efficacy of NF, and evaluate the measure of

impulse control as compared to the Tower of London test. The number of subjects

reported for the IVA varied for the different scales used from a low of n = 6 to a high of n

= 12, because they only included for analysis cases where pre-test scores needed to

improve. The researchers reported the means and standard deviations of the pre-post

values of the included scales, and performed paired t tests for statistical analysis.

Steiner, Sheldrick, Gotthelf, and Perrin (2011) conducted a randomized controlled

study with 41 children, comparing NF to a standardized computer attention training

program and used four outcome measures including the IVA. However, they only

included for analysis the two most broadly defined full-scale components of Response

Control and Attention, and only reported on an n = 6 for the NF group, and an n = 10 for

55

the computerized training group. Repeated measures ANOVAs were performed to

analyze the pre-post outcome measures in this study.

Rating scales. Rating scale instruments are one of the most common assessment

tools found in NF literature for measuring clinical outcomes. Rating scales are

instruments which require rated objects to be assigned to categories or numerical

continua, by the rater or observer, based on their perception or remembrance of the

behavior being rated (Kerlinger & Lee, 2000). Rating scales frequently employed in NF

literature include the BRIEF, the Conner’s’ Rating Scale-revised (CRS-R), the Behavior

Assessment Scale for Children (BASC), and the Child Behavior Checklist (CBCL). The

following are examples from the literature of their use in NF studies.

In a randomized study, Orgim and Kestad (2013) compared NF to medication for

a heterogeneous ADHD group with various comorbidities; each group had an n = 16, and

the NF group was administered 30 NF sessions. The outcome measures included the

rating scales of CRS-R and BRIEF. They conducted analysis of covariance (ANCOVA)

statistical tests, using baseline measurement (Time-1) as the covariate; and they analyzed

group differences at Time-2 for selected scaled scores.

The study of Huang-Storms et al. (2006) provided an example of the use of rating

scales, in a retrospective clinical study, in the form of the CBCL together with a

computerized performance test. The total number of valid CBCLs reported on was an n =

18, and all aforementioned scales were included in the analysis. The statistics employed

were two-tailed paired t test analysis.

Drechsler et al. (2007) conducted a study with an experimental design to assess

the efficacy of slow cortical potential NF with ADHD using multiple outcome measures;

56

where the experimental group had an n = 17 and the control group had an n = 13. Here

they employed two rating scales: The CPS-R and the BRIEF. Moreover, they only

included the composite or global scales from these instruments and performed repeated

measures MANOVAs for analysis.

In a randomized control study, Steiner et al. (2011) compared traditional NF to

computerized attention training to a waitlist control group; the group sizes were n = 9, n =

11, and n = 15. In this study, they used three rating scales: the CRS-R, the BASC, and the

BRIEF. Here too, they included selected scales from the assessments for analysis. The

statistics applied were rANOVAs, in an effort to detect if the experimental conditions

resulted in greater effects for the post NF assessment over the control group.

QEEG z-scores. As has been stated, with the QNF studies, by far, the vast

majority did not use pre-post EEG metrics or z-scores as an outcome measure. Though,

equally so, few traditional NF studies included EEG values as an outcome measure. Yet,

in one study purported to evaluate EEG effects of NF, Gevensleben et al. (2009) reported

values, as grouped together for nine regions across the scalp, and four frequency bands.

The averages of the microvolt values (raw, non z-score EEG values) were computed for

each region and frequency band, and post values minus pre values were used as a

measure of change. Since this was a study for traditional/amplitude NF, no z-score

metrics were used. Further, there were no goals of normalization in the NF protocols.

Two QNF studies do stand out for reporting, to some degree, pre-post EEG

metrics as part of the research. With Arns et al. (2012), non z-score pre-post EEG

microvolt data was analyzed, but for only nine sites, exclusive to frontal and central

areas, and for just three power frequencies. The group data was averaged, and presented

57

in a graph, for each site and frequency combination. Statistically significant pre-post

differences were noted for this data. The second QNF study (Breteler et al., 2010), did

report some pre-post z-scores information, but it was lacking in depth. The QEEG data

were reported for a limited number of sites and frequencies, as well as coherence pairs,

presumably as identified from the personalized training protocols.

Hammer et al. (2011) presented a unique offering in performing the binomial test

of significance to evaluate z-scores as an outcome measure of normalization. While the

results did show a statistically significant number of z-scores normalized after 4ZNF, the

findings were for only three frequencies (delta, beta, and high beta), and combined values

for absolute and relative power. Moreover, this methodology is limited in that it only

provides a yes/no level of analysis for normalization, not a discrete measure of change

towards the mean. Nonetheless, it is a useful offering in an effort to present a measure of

normalization of the QEEG in response to 4ZNF.

One reason for the lack of reporting of z-scores as outcome measures may be due

to the nature of z-scores encompassing both positive and negative values, which, when

averaged, tend to cancel out a magnitude of effect. This was noted in Ramezani’s (2008)

dissertation, which was a study comparing pre and post z-scores of coherence and phase

lag as a result of traditional NF. He noted that mean comparisons of z-scores, with both

positive and negative values being cancelled in the averaging process, had the potential of

masking true differences. In an effort to account for this, he chose to transform the values

by computing the absolute value of the z-score. He then used a score of z ≥ 1.0 as

inclusion criteria for analysis. This approach allowed for statistical analysis, (i.e.

58

averaging, ANOVAs, t tests) to be performed on the resulting z-scores transformed to

absolute values.

Krigbaum and Wigton (2013) presented a methodological approach to account for

positive and negative z-scores, by splitting the positive from negative z-scores, outside of

a cut-off score of ± z = 1.0, to calculate what is termed Sites of Interest (SoI). The

averaged SoI values were then plotted to display a learning curve for each participant,

and statistical analysis (i.e. t tests and rANOVAs) performed on the mean SoI z-score

values. While this methodology fits well for a single-subject design, and in quantifying

the progression of the z-scores towards the mean, its limitation lies in that (in the form

presented) it is not well suited for comparisons of group mean QEEG data. For example,

the split of positive and negative z-scores does not provide a single overall measure of

change for the z-scores. However, there is room to build on this research to develop a

methodology for comparing group data of QEEG z-scores.

Therefore, while few NF studies include EEG or QEEG z-score metrics as

outcome measures, when they do, frequently they only analyze selective components (i.e.

selected sites and/or frequencies). As a result, to date, no proposed methodology for

quantifying overall normalization has been published. Averaging non-transformed z-

scores is less than optimal due to the cancelling factor of the positive and negative values;

and the binomial test of significance provides only limited categorical analysis of the

data, without a measure of distance from the mean. The Krigbaum and Wigton (2013)

study appeared the closest to providing a model for measuring overall normalization of

the QEEG at this time. Still, building on this approach, by taking the absolute value of the

59

z-scores, to provide a single value as a measure of the distance from the mean, could

prove advantageous.

In summary, common themes in the literature present suitable outcome measures

for NF research to consist of computerized performance tests, rating scale instruments,

and QEEG metrics. Examples such as the IVA, the BRIEF, and z-scores were discussed.

These findings are relevant to this research in that the same or similar instruments were

used for the present study.

Summary

In reviewing the 40 year history of NF, a discussion of the historical context of

EEG, QEEG, and NF was presented. NF is grounded in learning theory and through the

years various models, such as traditional NF, QNF, ZNF, have emerged. While 19ZNF is

one of the newest NF models, it does not enjoy a demonstration of efficacy by evidence-

based research, which exists for the traditional models. In fact, there are significant gaps

in the literature in that no scientifically rigorous studies of 19ZNF have been found. This

study aims to address this empirical gap by analyzing the question of efficacy of 19ZNF

in a clinical setting, thus contributing to the field in terms of beginning to fill this

empirical gap. Thus this study aims to contribute to the body of scholarly knowledge

regarding 19ZNF.

Prior QNF and ZNF research is commonly found in clinical settings. These

research studies typically employ pretest-posttest designs using relatively small s

Place this order or similar order and get an amazing discount. USE Discount code “GET20” for 20% discount