Studien Details
Studien
EEG coherence effects of audio-visual stimulation (AVS)
Jon A. Frederick, Ph.D.*
DeAnna L. Timmermann, Ph.D.**
Harold L. Russell, Ph.D.***
Joel F. Lubar, Ph.D.****
Journal of Neurotherapy
*Corresponding author. Center for Computational Biomedicine, University of Texas Houston
Health Science Center, 7000 Fannin Suite 600, Houston, TX 77030. (713) 500-3464, email:
smiile@psynet.net
**Department of Psychology, Eastern Oregon University, One University Avenue, LaGrande,
OR 97850.
***P.O. Box 240, Galveston, TX, 77553.
****Department of Psychology, University of Tennessee, 307 Austin Peay, Knoxville, TN
37996.
Page 2
1
SUMMARY. The effects of a single session of audio-visual stimulation (AVS) at the
dominant alpha rhythm and twice-dominant alpha frequency on EEG coherence were
studied in 23 subjects. An eyes-closed baseline EEG determined each subject's dominant
alpha frequency. Subjects were stimulated at their dominant alpha frequency or at twice
dominant alpha frequency for twenty minutes, while EEG was recorded in 5-minute
intervals. A post-session baseline was recorded 30 minutes after each session. AVS
decreased coherence in the intrahemispheric projections from the occipital region and the
parietal midline, and generally increased coherence, with few exceptions, among all other
longitudinal pairs. Interhemispheric coherence increased posteriorily and high
frequencies, and tended to decrease frontally and low frequencies. Alpha AVS was more
effective than twice-alpha AVS at producing interhemispheric synchronization, and
tended to produce more effects overall. Although main effects of frequency and time
were observed, when individual coherence pairs changed, they almost always changed in
only one direction. Overall coherence was greater during the first ten minutes than the
last ten minutes, and greatest in the beta 1 and delta 2 bands, and lowest in the alpha and
delta 1 bands. Few, if any, significant effects persisted into the post-stimulation baseline.
A new method of assessing the effects of multiple comparisons on experimentwise error,
based on randomization theory, is proposed and implemented.
Page 3
2
INTRODUCTION
The ability of a flashing light stimulus to evoke EEG rhythms related to the
stimulus frequency has been studied since the early history of electroencephalography
(Adrian & Matthews, 1934). Known as the photic driving response (PDR), or steady state
visual evoked potential, this effect is commonly measured in routine clinical EEG
examinations, and has been proven useful for investigating neurological disorders
(Takahashi, 1987; Coull & Pedley, 1978; Duffy, Iyer, & Surwillo, 1989).
The diverse perceptual and emotional effects of photic stimulation (Walter &
Walter, 1949; Stwertka, 1993; Gizycki, Jean-Louis, Snyder, Zizi, Green et al., 1998), and
it?s ability to cause seizures in susceptible individuals (Walter, Dovey & Shipton, 1946;
Striano, Meo, Bilo, Ruosi, Soricellis et al., 1992) have led many to investigate whether
rhythmic auditory and visual stimulation (AVS) might also induce clinically beneficial
changes in brain activity. In the 1950?s and 60?s, many studies focused on the ability of
AVS to induce relaxation and hypnosis (reviewed in Morse, 1993). Others have reported
AVS to be effective for relieving a diversity of pain symptoms (Solomon, 1985;
Anderson, 1989; Shealy, Cady, Cox, Liss et al., 1990), treating dental anxiety (Morse,
1993), premenstrual syndrome (Noton, 1997), fibromyalgia (Mueller, Donaldson, Nelson
& Layman, 2001) and for alleviating the cognitive dysfunctions associated with closed
head injury (Montgomery, Ashley, Burns & Russell, 1994) and strokes (Russell, 1997;
Rozelle & Budzinski, 1995). Since the enhancement of beta (13-21 Hz) and inhibition of
theta (4-8 Hz) is a goal of EEG biofeedback for the treatment of attention deficit
hyperactivity disorder (ADHD; Lubar & Lubar, 1999; Lubar, Swartwood, Swartwood &
O?Donnell, 1995), some have proposed using AVS in neurofeedback as a ?priming
stimulus? to encourage the endogenous production of desired cortical frequencies, which
are then reinforced as the conditioned response. In a study of 25 ADHD children, Patrick
(1996) found ?photic-driven EEG neurotherapy? effective in improving cognitive,
behavioral, and clinical EEG measures in less than half the number of sessions usually
required. Meanwhile, Micheletti (1999) found AVS alone effective in improving
cognitive and behavioral measures, in a study of 99 ADHD children. Carter and Russell
(1993) reported significant improvement in cognitive and behavioral functioning, related
to the number of AVS sessions, in learning disabled boys. Joyce and Siever (2000)
reported that a 7-week audiovisual stimulation treatment in 8 reading-disabled children,
compared to a control group, normalized scores on the Test of Variables of Attention
(TOVA), improved scores on the Standardized Test for the Assessment of Reading
(STAR), and improved general behavior as noted by teachers and parents.
Mechanisms by which long-term AVS therapies may cause these behavioral
changes have been suggested by research in neuronal plasticity. A number of
investigators (van Praag, Kempermann & Gage, 2000; Rosenzweig, 2003; Mohammed,
Zhu, Darmopil, Hjerling-Leffler, Ernfors et al., 2002) are in essential agreement that
ongoing direct experience that evokes persistent neuronal activation alters brain structure
and brain functioning. Although most studies have focused on effects of an enriched
environment, persistent neuronal activation can also be evoked by trains of sensory
stimuli. Human subjects have been shown to respond to flicker frequencies from 1-100
Hz with steady-state activity at all frequencies up to at least 90 Hz with clear resonance
phenomena or harmonics at 10, 20, 40 and 80 Hz (Herrmann, 2001). A possible linkage
Page 4
3
between steady-state stimulation induced neuronal activation and neuronal plasticity is
the increasing evidence that brain electrical activity regulates the synthesis, secretion and
actions of neurotrophins (Schindler and Poo, 2000), which promote synaptogenesis.
The most commonly studied PDRs have been the effects of stimulation on alpha
(8-13 Hz) power over the occipital region (Iwahara, Noguchi, Yang & Oishi, 1974;
Aranibar & Pfurtscheller, 1978). The photic driving response is most reliable when the
stimulus approximates the subject?s peak alpha frequency (Toman, 1941; Townsend,
Lubin, & Naitoh, 1975). However, recent studies have shown that AVS activates a
diverse range of EEG frequencies, beyond the primary sensory cortices, and outside of
the frequency of stimulation. Using low-frequency theta AVS, Dieter & Weinstein (1995)
described a significant reduction in "mean activity" (an increase of delta and theta
activity) in frontal, central, and parietal regions, in addition to occipital regions. In a
study of 13 college students (Timmermann, Lubar, Rasey & Frederick, 1999), we found
that effects of AVS were widely distributed across the standard 10-20, 19-channel
montage. AVS at a subject's dominant alpha frequency had no effect in the alpha band,
but significantly increased power in the delta 1, delta 2, theta, beta 1, and beta 2 bands.
Stimulation at twice the dominant alpha frequency significantly increased theta, alpha,
beta 1, and beta 2 power.
While the amplitude and power effects of AVS have been widely studied,
relatively little is known about the effects of AVS on EEG coherence. Coherence is a
correlational measure, varying between zero and one, of the variability in phase between
two signals over time (Shaw, 1981). This frequency-specific signal correlation suggests
the extent to which two regions are cooperating on the same task. High coherence
indicates a common signal, whether it is synchronous between two locations, or delayed
by a constant conduction velocity. Coherence in the eyes-closed baseline reflects the
number of synaptic connections between recording sites, and the strength of these
connections (Thatcher, 1992). Coherence has been shown to be lower in Alzheimer
patients, comatose subjects, and in brain-injured subjects, while it is higher in mentally
retarded persons, during sleep, and during epileptic seizures. Between these extremes,
"optimal levels" of coherence for normal functioning have been described (Silberstein,
1995). Some have suggested that EEG coherence biofeedback could be used to normalize
the coherence deviations seen in dyslexic and head injured subjects (Evans & Park, 1996;
Hoffman, Stockdale, Hicks & Schwaninger, 1995).
Differences in photic driving of coherence have been described between normal
subjects and patients with Alzheimer's disease (Wada, Nanbu, Kikuchi, Koshino,
Hashimoto et al., 1998a), schizophrenia (Wada, Nanbu, Kikuchi, Koshino & Hashimoto,
1998b), and between genders (Wada, Nanbu, Kadoshima & Jiang, 1996). However, the
effects of combined auditory and visual stimulation on coherence in normal subjects have
not been previously reported.
Although AVS devices are used by many neurotherapists as an adjunct to EEG
biofeedback, the overall pattern of effects of AVS on coherence needs to be better
understood, to ensure that AVS treatment is influencing coherence in the appropriate
direction. To begin to achieve this understanding, we conducted an exploratory study of
the effects of AVS on coherence in normal college students. We hypothesized that AVS
would increase coherence at the frequency of stimulation, and assumed that effects would
be most prominent over the occipital and temporal leads, which are closest to the primary
Page 5
4
visual and auditory cortex. Given our previous findings of increased amplitude in
multiple frequency bands (Timmermann et al., 1999), we anticipated effects across the
coherence spectrum. However, since our goal was to observe the effects rather than to
verify any hypothesis about them, beyond the expected increase at the stimulus
frequencies, we did not predict directions of change.
METHODS
Participants. This study was a reanalysis, in terms of coherence, of data
previously analyzed and reported in terms of power (Timmermann et al., 1999).
However, in this study, ten additional participants were added to each stimulation
condition. Participants (11 male and 22 female) were recruited from the undergraduate
and graduate populations of the University of Tennessee, Knoxville. Participants ranged
in age from 20-45 years, with a mean of 25 years. All participants reported no previous
history of epilepsy, learning disabilities, attention deficit disorder, or mental illnesses
during personal screening interviews. Participants self-reported that they were free of
medication use during the study.
Apparatus. Audio-visual stimulation was provided by a Polysync Pro (Synetic
Systems) device. This unit consisted of headphones and a pair of "photoscopic" glasses
that were connected to a small, portable unit that was programmed to provide specified
levels of visual and auditory stimulation. The glasses had eight light emitting diodes
(LEDs), four per side, arranged in a cross pattern. The LEDs were situated approximately
1.5 cm from the eyes, and each emitted red light at .166 candle power at the frequencies
employed. Audio stimulation consisted of a cycled tone with a pitch of 185 Hz, presented
to both ears simultaneously, with a duty cycle of 50% and a loudness level of
approximately 77 dB(A scale) for the alpha condition and 81 dB(A scale) for the beta
condition. Both auditory and visual stimulation were sinusoidally modulated (Townshend
et al., 1975). The Polysync Pro equipment was tested for and did not produce any
detectable electrical interference in our EEG recordings (Timmermann et al, 1999).
Procedures. The first thirteen participants came to the laboratory on two different
occasions for AVS sessions. These sessions were at least two weeks apart to minimize
carry-over effects. The presentation of the AVS condition was counter-balanced; during
the first AVS session six of the participants experienced alpha stimulation and seven
experienced twice-dominant alpha ("beta") stimulation. For the second session, those
who experienced alpha stimulation during the first session received beta stimulation, and
those who experienced beta stimulation first received alpha stimulation. The additional
twenty participants were randomly assigned to either an alpha AVS or beta AVS group,
as part of an experiment measuring the effects of multiple sessions of AVS
(Timmermann, 1999). Only results from the first session of this experiment are included
in this analysis. The two groups were counterbalanced for age and gender, and did not
differ significantly with respect to dominant alpha frequency. The mixing of subjects
from both repeated measures and independent groups designs in the present study might
compromise the validity of some inferences about the differences between the two
stimulation conditions. However, both experimental designs have advantages and risks
for detecting such differences: repeated measures designs risk within-subject effects,
Page 6
5
while independent group designs risk between-subject effects. We decided that this
compromise was worth the increase in statistical power obtained for all other
comparisons in this study.
The procedure was the same for either stimulation condition. Participants were
seated upright in a plastic chair located in a sound-attenuated, dimly lit room for the EEG
recordings. The headphones and glasses were placed over the electrode cap. All
participants had an eyes-closed baseline EEG recorded for four minutes at the beginning
of the session. This baseline recording was analyzed to determine each subject's dominant
alpha frequency, defined as the peak power between 8 and 12 Hz at locations Pz and P4
as measured by power spectral analysis, rounded to the nearest 0.5 Hz. Participants were
then provided AVS for 20 minutes, with EEG recording occurring simultaneously. For
the alpha stimulation condition, participants were stimulated at their dominant alpha
frequency. During the beta stimulation condition each participant received AVS at twice
their dominant alpha frequency. Participants were instructed to close their eyes and relax
during the 20-minute AVS. Thirty minutes after the stimulation session, a post-session
eyes-closed EEG was recorded for 5 min.
EEG Recordings. Quantitative referential EEGs (monopolar montage) were
recorded from 19 electrodes following the International 10-20 system for electrode
placement, with linked earlobe references. All electrode impedances were below 5
KOhm. Recordings were made using an electrode cap (Electro Cap Inc.). Raw EEG was
fed through 19 matched 7P511 pre-amplifiers (Grass Instrument Co.), with bandpass
filters set to 0.5 - 100 Hz. A BMSI 12-bit A to D converter digitized the outputs of the
amplifiers. Rhythm software (Stellate Systems) was employed to record the raw EEG.
The sampling rate was set at 128 samples per second. Under Rhythm, when a sampling
rate of 128 Hz is specified, the actual rate employed is oversampled at 256 Hz and each
digitized point is replaced by a weighted sum of its neighbors. Every other point of this
filtered data is written to disk as a 16 bit value. Signal aliasing was eliminated by the use
of a 16 point FIR (finite impulse response) filter, with a sharp low pass cutoff set at 64
Hz and higher. Coherences were analyzed from the raw EEG off-line on a Pentium 233
processor using Rhythm software. Rhythm analysis employs Hanning windowing and
cosine tapering of each selected four-second epoch. Eye-blinks, large eye movements,
and all observable muscle artifacts were removed prior to analysis by a visual review of
the EEG records. Tests of normality, ANOVAs, and sign rank tests were performed with
Statistical Analysis Software (SAS Institute). Randomization tests were performed with
custom algorithms written in PERL by the first author (available by request).
RESULTS
EEG data were analyzed via Fast-Fourier transformation to derive 8
interhemispheric and 55 longitudinal coherence pairings in each of six bandpasses (delta
1, 0.75-2 Hz; delta 2, 2-4 Hz; theta, 4-8 Hz; alpha, 8-12 Hz; beta 1, 13-21 Hz; and beta 2,
21-31 Hz). Each 20-minute AVS condition was analyzed in four 5-minute blocks (0-5
min, 5-10 min, 10-15 min, and 15-20 min) to examine changes over the course of
stimulation. The post measure was a 5 minute EEG recording taken one half hour after
the AVS session. Thus, there were six coherence measures per frequency bandpass in this
Page 7
6
pilot study: baseline, 0-5 min, 5-10 min, 10-15 min, 15-20 min, and post.
These data were found to have a significantly non-normal distribution by the
Kolmogorov-Smirnov test (p<.01), so it was determined that all inferential statistics
would be performed on ranked data (Conover, 1980). Differences from the pre-
stimulation baseline condition were calculated for each stimulation condition, and sorted
and ranked by subject. A repeated measures analysis of variance on these data found
significant main effects of time (F=27.84, p<.0001), electrode pairing (F=41.28,
p<.0001), and frequency (F=43.08, p<.0001). Tukey's post-hoc procedure determined (at
p<.05) that the overall coherence was higher during the first ten minutes (intervals 1 and
2) and the post-stimulation baseline than during the last ten minutes of stimulation
(intervals 3 and 4). Tukey's procedure found three distinct frequency groupings (beta 1&
delta 2 > theta & beta 2 > alpha & delta 1). The type of stimulation had no significant
main effect (using the more conservative, independent groups ANOVA model; F=3.03,
p=.08). However, interactions between stimulation type and frequency were observed
(F=18.23, p<.0001). Alpha stimulation decreased overall coherence, compared to beta
stimulation, in the alpha (F=17.18, p<.0001) and theta (F=6.77, p<.0093) bands. Beta
stimulation decreased overall coherence, compared to alpha, in the delta 1 (F=15.69,
p<.0001) and beta 2 (F=40.43, p<.0001) bands. There were no significant differences of
stimulation type in the delta 2 and beta 1 bands.
The number of possible interactions of electrode pairing with other variables was
considered too great to be approached practically with ANOVA methods and Tukey's
procedure. Thus, the difference from baseline for each pairing location was tested for
significance with Wilcoxon's sign rank test (at p<=.01). Since this procedure was
performed on 3780 variables (2 stimtypes X 6 frequencies X 5 times X 63 pairing
locations), under the null hypothesis, 3780 x .01 = 37.8 false-positive sign rank tests
could be expected by random effects alone, whereas we actually observed 241 positive
tests. Although it is highly unlikely that this ratio of signal-to-noise would arise by
chance alone in 3780 independent experiments, these variables are highly interdependent.
A principal components analysis revealed that at most, 19 independent factors explained
100% of the variance in these data. It is possible that the large number of significant
variables is explained, not by the experimental conditions, but by random effects in only
one or a few large underlying factors. To test whether this was the case, we constructed
empirical distribution functions (EDFs) of the number of significant sign rank tests
observed when the baseline was randomized with respect to the stimulation conditions,
resampling each of 3780 sign-rank tests 1000 times with replacement. The rank of the
actual observed number among the randomized trials (divided by 1000) was thus the
probability of type I error (Edgington, 1987). To preserve the covariance relationships
among variables, a single random decision determined whether the sign of the difference
from baseline would be reversed for all variables for each subject, a total of 23 decisions
per trial. The probability of 241 false-positive sign rank tests at p<=.01 arising by chance
was thereby determined to be p<.001.
Similar comparisons demonstrated that a significant number of positive tests had
been observed in each stimulation type, frequency, and time, with the exception of the
post-stimulation baseline. When random EDFs were generated for each of the 19
electrode locations, the number of positive tests observed/expected was significant at
p<.01 for most, and at p<.05 for all 19 locations. It was then of interest to see how each
Page 8
7
of the 63 coherence pairs contributed to this pattern. Seventeen of these comparisons
were significant at p<.01 (29 at p<.05). Table 1 shows the number of coherence changes
for each variable (positive, negative, and total coherence changes), compared to the
number that would be expected from experimentwise error.
The distribution of the positive vs. negative effects in table 1 supports the general
findings of the ANOVAs: B1 and D2 had more positive changes than D1, and the first
ten minutes had a higher ratio of positive-to-negative changes than the last ten minutes.
Alpha and beta stimulation evoked roughly the same ratio of positive-to-negative
changes, although alpha stimulation evoked a greater total number of changes. The
anatomical distribution of this pattern became clear when plotted on graphical heads
(using Vbmapper software; Frederick, 2001), as shown in figures 1 and 2. With few
exceptions, the decreasing coherences were longitudinal projections from the occipital
leads and Pz. The increasing coherences were always associated with the frontal poles,
the interhemispheric pairings, and all the remaining longitudinal projections that did not
involve O1, O2, or Pz. An interesting exception was that 4/4 of the significant frontal
interhemispheric coherences (FP1FP2 and F7F8) decreased. This tendency toward
decreasing frontal interhemispheric coherences became even more apparent at p<=.05.
We found that the findings at p<=.05 showed a high degree of consistency with those at
p<=.01, so these are also represented (with lighter, less saturated lines) in figures 1 and 2.
Most remarkable was the observation that individual coherence pairing locations
generally changed in only one direction, regardless of frequency, time, or type of
stimulation. For example, the main effect of time appeared to be mediated by the
decreasing number of increased frontal intrahemispheric coherences (from 25 in the first
five minutes to 8 in the last five minutes), along with the increasing number of decreased
posterior intrahemispheric coherences (from 18 to 26).
DISCUSSION
This study has demonstrated a number of anatomical asymmetries in the
coherence effects of AVS. AVS decreased intrahemispheric coherences involving the
posterior leads Pz, O1 and O2, but increased intrahemispheric coherence frontally and
centrally. Meanwhile, the interhemispheric derivations increased posteriorly and at high
frequencies, and tended to decrease frontally and at low frequencies. The overwhelming
tendency of all the significant effects for a given electrode pair to go in the same direction
independently of frequency, time, or type of stimulation, is another striking asymmetry.
The segregation of these positive and negative changes into discrete anatomical
compartments suggests that descriptions of "coherence" that omit discussion of location
are somewhat oversimplified. The many significant decreases in coherence demonstrate
that the popular naming and marketing of AVS devices as "brain wave synchronizers"
(Morse, 1993) is also oversimplified.
Overall coherence was significantly lower during the last ten minutes of
stimulation. During the first five minutes, the ratio of positive to negative changes was
40:18, which changed to 24:32 by the last five minutes. These results suggest that
subjects were adapting or habituating to the effects of stimulation. The failure of the
effects to persist into the post-stimulation baseline suggests a limitation to the use of
Page 9
8
these devices as ?stand-alone? treatments for influencing coherence. The AVS experience
is essentially a passive one, and we would expect that more persistent effects would be
observed if concurrent biofeedback were used to reinforce desired changes in the EEG.
However, it is also possible that more persistent effects of AVS alone could be achieved
with a variable-frequency stimulus paradigm or a larger number of sessions.
Our finding of a highly consistent inhibitory effect in the occipital
intrahemispheric coherences stands in contrast with two recent studies of photic driving
of coherence. Wada et al. compared the photic driving of coherence between Alzheimer
patients and 10 normal participants (mean age 59; Wada et al., 1998a); and between
schizophrenic patients and 30 normal participants (mean age 22.3; Wada et al., 1998b).
Although comparisons with the baseline were not statistically tested, graphs comparing
the subject groups showed that 5, 10, and 15 Hz photic stimulation (with no auditory
stimulus present) increased or tended to increase all intrahemispheric coherences from
the baseline in normal subjects. These included O1P3, O2P4, O1T5, and O2T6 in the first
study, and O1C3 and O2C4 in the second study. We have, however, replicated these
findings in a new sample of 30 participants with new, custom software (coherence
algorithms by David Joffe at Lexicor; Frederick & Lubar, 2002). Interestingly, spectral
correlation, a measure which is often confused with coherence, did not decrease during
stimulation in these same recordings. Possibly, the differences between Wada et al.'s
findings and ours might be explained by their use of a more intense, square-wave
stimulus, or the lack of an auditory stimulus.
Our statistical analysis of these data has taken a new approach to the problem of
type I error from multiple comparisons. Many previous studies (e.g., Dafters, Duffy,
O'Donnell & Bouqet, 1999; Wada et al., 1998b) have avoided this problem by "not
looking" at more than a small representative subset (6 to 10) of the 171 coherences that
are possible among 19 electrodes. We believe that this approach has unacceptable
consequences for type II error. Limitations to the Stellate Rhythm software made it
impractical for us to look at more than 64 coherence pairs, but at this level of resolution,
it is apparent that groups of coherences (e.g. frontal longitudinal) can change reliably
while individual members of that group might not. Other studies have simply performed
hundreds of univariate tests and relied on the intuitive clarity of the overall magnitude
and scope of the effect (e.g., Gevins et al., 1987). We believe that an optimal balance
between type I and type II error is achieved not by avoiding or ignoring the problem of
multiple comparisons, but by measuring it. When the problem of experimentwise error is
framed as a null hypothesis, that "these significant univariate tests arise from random
effects," clearly the most direct method of testing this hypothesis is to compare the
number of observed effects to the empirical distribution of random effects.
While this study has shown that the coherence effects of discrete-frequency AVS
are distributed throughout the frequency spectrum, it should be acknowledged that our
stimulus paradigm was highly artificial compared to those used in a clinical setting.
Further research is needed to understand the effects of variable-frequency stimulation,
auditory vs. visual vs. combined AVS, binaural beat stimulation, EEG-driven AVS, and
desynchronized AVS paradigms. Nonetheless, the ability to alter patterns of coherence in
the brain is potentially a powerful tool for neurotherapy. The recruitment of alternative
pathways and circuits is often essential for recovery from neurological and perhaps
psychiatric disorders. Neurons and neural pathways have a much larger connectivity than
Page 10
9
their usual territory of functional influence, which can be "unmasked" by disinhibiting or
potentiating these connections (Mallet, 2001). Temporally paired (or coherent) inputs are
required for associative long-term potentiation to occur (Kelso, Ganong & Brown, 1986.)
Synchronous activity has been shown to be an essential signal for synaptogenesis in the
developing brain, as well as axonal sprouting after cortical lesions in the adult
(Carmichael, 2003). The changes in coherence evoked by AVS in this study suggest that
long-term AVS therapies may activate these mechanisms of neuronal plasticity to
reorganize functional linkages in the brain.
Page 11
10
REFERENCES
Adrian, E. D., & Matthews, B. H. C. (1934). The Berger rhythm: potential changes from the
occipital lobes in man. Brain, 57, 355-384.
Anderson, D. J. (1989). The treatment of migraine with variable frequency photostimulation.
Headache, 29, 154-155.
Aranibar, A., & Pfurtscheller, G. (1978). On and off effects in the background EEG activity
during one-second photic stimulation. Electroencephalography and Clinical
Neurophysiology, 44, 307-316.
Carter, J. L., & Russell, H. L. (1993). A pilot investigation of auditory and visual entrainment of
brainwave activity in learning-disabled boys. Texas Researcher: Journal of the Texas
Center for Educational Research, 4, 65-73.
Carmichael, S. T. (2003). Plasticity of cortical projections after stroke. Neuroscientist, 9 (1), 64-
75.
Conover, W. J. (1980). Practical Nonparametric Statistics. New York: John Wiley & Sons.
Coull, B. M. & Pedley, T. A. (1978). Intermittent photic stimulation. Clinical usefulness of non-
conclusive responses. Electroencephalography and Clinical Neurophysiology, 44, 353-
363.
Dafters, R. I., Duffy, F., O'Donnell, P. J. & Bouqet, C. (1999). Level of use of 3,4-
methylenedioxymethamphetamine (MDMA or Ecstasy) in humans correlates with EEG
power and coherence. Psychopharmacology, 145 (1), 82-90.
Dieter, J. N. I., & Weinstein, J. A. (1995). The effects of variable frequency photo-stimulation
goggles on EEG and subjective conscious state. Journal of Mental Imagery, 19, 77-90.
Duffy, F. H., Iyer, V. G., & Surwillo, W. W. (1989). Clinical and Electroencephalography and
Topographic Brain Mapping. New York: Spinger-Verlag.
Edgington, E. S. (1987). Randomization Tests, 2nd ed. New York: M. Dekker.
Evans, J. R. & Park, N. S. (1996). Quantitative EEG Abnormalities in a Sample of Dyslexic
Persons. Journal of Neurotherapy, 2 (1).
Frederick, J. A. (2001). Technical Note: Open source method of graphical QEEG analysis using
PERL and Visual Basic. Journal of Neurotherapy, 4 (4), 63-70.
Frederick, J. A. & Lubar, J. F. (2002). Differences between coherence and spectral correlation
during auditory and visual stimulation at dominant alpha frequency. Journal of
Neurotherapy, 6(1): 70-1.
Gevins, A. S., Morgan, N. H., Bressler, S. L., Cutillo, B. A., White, R. M., Illes, J., Greer, D. S.,
Doyle, J. C., Zeitlin, G. M. (1987). Human Neuroelectric patterns predict performance
accuracy. Science, 235, 580-585.
Gizycki, H., Jean-Louis, G., Snyder, M., Zizi, F., Green, H., Giuliano, V., Spielman, A., & Taub,
H. (1998). The effects of photic driving on mood states. Journal of Psychosomatic
Research, 44 (5): 599-604.
Herrmann, C. S. (2001). Human EEG responses to 1-100 Hz flicker: resonance phenomena in
visual cortex and their potential correlation to cognitive phenomena. Experimental Brain
Research, 137 (3-4):346-53.
Hoffman, D. A., Stockdale, S., Hicks, L. L., & Schwaninger, B. A. (1995). Diagnosis and
treatment of head injury. Journal of Neurotherapy 1 (1).
Iwahara, S., Noguchi, S., Yang, K., & Oishi, H. (1974). Frequency specific and nonspecific
effects of flickering lights upon electrical activity of the human occiput. Japanese
Psychological Research, 16, 1-7.
Joyce, M., Siever, D. (2000). Audio-visual Entrainment Program as a Treatment for Behavior
Disorders in a School Setting. Journal of Neurotherapy, 4 (2), 9-25.
Kawaguchi, T., Jijiwa, H., & Watanabe, S. (1993). The dynamics of phase relationships of alpha
waves during photic driving. Electroencephalography and Clinical Neurophysiology, 87,
Page 12
11
88-96.
Kelso, S. R., Ganong, A. H., and Brown, T. H. (1986). Hebbian synapses in hippocampus.
Proceedings of the National Academy of the Sciences USA, 83, 5326-5330.
Lubar, J. F. & Lubar, J. O. (1999). Neurofeedback assessment and treatment for attention
deficit/hyperactivity disorders. In Evans, J.R., & Abarbanel, A., (Eds). Introduction to
Quantitative EEG and Neurofeedback (pp. 103-143). San Diego, CA, US: Academic
Press, Inc.
Lubar, J. F., Swartwood, M. O., Swartwood, J. N., & O'Donnell, P. (1995). Evaluation of the
effectiveness of EEG neurofeedback training for ADHD in a clinical setting as measured
by changes in T.O.V.A. scores, behavioral ratings, and WISC-R performance.
Biofeedback and Self-Regulation, 20, 83-99.
Mallet, M. (2001). Plasticity of the human motor cortex and recovery from stroke. Brain
Research Reviews, 36 (2-3), 169-74
Micheletti, L. S., (1999). The use of auditory and visual stimulation for the treatment of Attention
Deficit Hyperactivity Disorder in children (ritalin, adderall). Dissertation Abstracts
International: Section B: The Sciences and Engineering, 60 (6-B), 2953.
Mohammed, A. H., Zhu, S. W., Darmopil, S., Hjerling-Leffler, J., Ernfors, P., Winblad, B.,
Diamond, M. C., Eriksson, P. S., & Bogdanovic, N. (2002). Environmental enrichment
and the brain. In Hofman, M. A., Boer, G. J., Holtmaat, A. J. G. D., Van Someren, E. J.
W., Verhaagen, J., & Swaab, D. F. (Eds.), Progress in Brain Research, 138, 109-133.
Montgomery, D. D., Ashley, E., Burns, W. J., & Russell, H. L. (1994). Clinical outcome of a
single case study of EEG entrainment for closed head injury. Proceedings from the
Association for Applied Psychophysiology and Biofeedback, 25, 82-83.
Morse, D. R. (1993). Brain wave synchronizers: A review of their stress reduction effects and
clinical studies assessed by questionnaire, galvanic skin resistance, pulse rate, saliva, and
electroencephalograph. Stress Medicine, 9, 111-126.
Mueller, H. H., Donaldson, C. C. S., Nelson, D. V., & Layman, M. (2001). Treatment of
fibromyalgia incorporating EEG-driven stimulation. Journal of Clinical Psychology 57
(7): 933-952.
Noton, D. (1997). PMS, EEG, and photic stimulation. Journal of Neurotherapy, 2 (2), 8-13.
Patrick, G. J. (1996). Improved neuronal regulation in ADHD: An application of 15 sessions of
photic-driven EEG neurotherapy. Journal of Neurotherapy, 1, 27-36.
Regan, D. (1966). Some characteristics of average steady-state and transient responses evoked by
modulated light. Electroencephalography and Clinical Neurophysiology 20, 238-248.
Rosenzweig, M. R. (2003). Effects of differential experience on the brain and behavior.
Developmental Neuropsychology, 24 (2-3), 523-540.
Rozelle, G.R. & Budzynski, T.H. (1995). Neurotherapy for stroke rehabilitation: A case study.
Biofeedback and Self-Regulation, 20, 211-228.
Russell, H. L. (1997). Intellectual, auditory, and photic stimulation and changes in functioning in
children and adults. Biofeedback, 25 (1), 16-17, 23-24.
Shaw, J. C. (1981). An introduction to the coherence function and its use in EEG signal analysis.
Journal of Medical Engineering Technology, 5 (6): 279-288.
Shealy, C. N., Cady, R. K., Cox, R. H., Liss, S., Clossen, W. & Culver Veehoff, D. (1990). Brain
wave synchronization (photo-stimulation) with the Shealy RelaxMate (TM). Shealy
Institute for Comprehensive Health Care: Springfield, MO.
Schindler, A. F., Poo, M. (2000). The neurotrophin hypothesis for synaptic plasticity. Trends in
Neuroscience, 23 (12), 639-45.
Silberstein, R. B. (1995). Neuromodulation of neocortical dynamics. In P. L. Nunez (Ed.),
Neocortical Dynamics and Human EEG Rhythms (pp. 591-627). New York: Oxford
University Press.
Solomon, G. D. (1985). Slow wave photic stimulation in the treatment of headache-a preliminary
Page 13
12
study. Headache, 25, 444-446.
Striano, S., Meo, R., Bilo, L., Ruosi, P., Soricellis, M., Estraneo, A., Caporella, A. (1992). The
use of EEG activating procedures in epileptology. Acta Neurologica, 14(4-6): 275-289.
Stwertka, S. A. (1993). The stroboscopic patterns as dissipative structures. Neuroscience and
Biobehavioral Reviews, 17: 69-78.
Takahashi, T. (1987). Activation methods; in Niedermeyer E., Lopez da Silva, F. (eds):
Electroencephalography, Basic Principles, Clinical Applications and Related Fields, ed
2. Baltimore, Urban & Scharzenberg 1987, pp 209-227.
Thatcher, R. W. (1992). Cyclic cortical reorganization during early childhood. Brain and
Cognition, 20, 24-50.
Timmermann, D. L. (1999). An assessment of the effects of multi-session audio-visual
stimulation on cognitive measures and the cortical EEG. Dissertation Abstracts
International: Section B: The Sciences and Engineering, 60 (3-B), 1343.
Timmermann, D., Lubar, J. F., Rasey, H. W., & Frederick, J. A. (1999). Effects of dominant and
twice-dominant alpha audiovisual stimulation on the cortical EEG. International Journal
of Psychophysiology, 32, 55-61.
Toman, J. (1941). Flicker potentials and the alpha rhythm in man. Journal of Neurophysiology, 4:
51-61.
Townsend, R. E., Lubin, A., & Naitoh, P. (1975). Stabilization of alpha frequency by sinusoidally
modulated light. Electroencephalography and Clinical Neurophysiology, 39, 515-518.
van Praag, H., Kempermann, G., & Gage, F. H. (2000). Neural consequences of environmental
enrichment. Nature Reviews Neuroscience, 1, 191-198.
Wada, Y., Nanbu, Y., Kadoshima, R., & Jiang, Z. Y. (1996). Interhemispheric EEG coherence
during photic stimulation: Sex differences in normal young adults. International Journal
of Psychophysiology, 22 (1-2), 45-51.
Wada, Y., Nanbu, Y., Kikuchi, M, Koshino, Y., Hashimoto, T., & Yamaguchi, N. (1998a).
Abnormal functional connectivity in Alzheimer's disease: Intrahemispheric EEG
coherence during rest and photic stimulation. European Archives of Psychiatry and
Clinical Neuroscience, 248 (4), 203-208.
Wada, Y., Nanbu, Y., Kikuchi, M., Koshino, Y., & Hashimoto, T. (1998b). Aberrant functional
organization in schizophrenia: Analysis of EEG coherence during rest and photic
stimulation in drug-naive patients. Neuropsychobiology, 38 (2), 63-69.
Walter, W. G., Dovey, V. J., & Shipton, H. (1946). Analysis of electrical response of human
cortex to photic stimulation. Nature, 158, 340-541.
Walter, V. J., & Walter, W. G. (1949). The central effects of rhythmic sensory stimulation.
Electroencephalography and Clinical Neurophysiology, 1, 57-86.
Page 14
13
CAPTIONS
TABLE 1. Likelihood of type I error resulting from multiple comparisons (p). The number of significant
coherence differences (obs; sign-rank test p≤0.01) was counted for each variable across all other variables,
and compared to the distribution of false-positive tests observed in 1000 randomized trials. Pos denotes the
number of increases, neg, the number of decreases, null, the average number of false-positive tests in
random data, ALPHA, the alpha stimulus condition; BETA, the twice dominant alpha stimulus condition, D1-
B2, coherence bandpasses; 0-5 min etc., recording intervals; F1, F2, etc. individual members of coherence
pairs; F1C3, F1CZ, etc., coherence pairs.
FIGURE 1. Effect of dominant alpha stimulation on EEG coherence. Dark red lines, increases at p≤.01; light
red lines, increases at p≤.05; dark blue lines, decreases at p≤.01; light blue lines, decreases at p≤.05.
FIGURE 2. Effect of twice-dominant alpha stimulation on EEG coherence. Dark red lines, increases at
p≤.01; light red lines, increases at p≤.05; dark blue lines, decreases at p≤.01; light blue lines, decreases at
p≤.05.
Testimonials
Sense your sounds!Terry G.
"I must tell you that I absolutely love the laxman system. I purchased this system without really knowing what to expect, and it has been a wonderful investment.
I have shown it to many people, all of whom were amazed at how relaxing the sessions were."
Kevin B.