Hemi-Sync® Journal 1990-2 (Spring)

theoretically provide the subject a measure of the success with which he or she is allowing the binaural beat to entrain the EEG. Subjects Sixty volunteer undergraduate and graduate students from Memphis State University and Christian Brother’s College participated in the study. The students from Christian Brother’s College were volunteers from Jane Davis’ introductory psychology classes. The Memphis State students were from Burl Gilliland’s, Bob Davis’, and Fleetis Hannah’s counseling classes. Participants were screened for known neurological damage and abnormalities. Instrumentation The binaural beats were provided by a model 201B Hemi-Sync Synthesizer (“Instruction Manual,” undated). The alpha-frequency binaural beats were created by presenting two pure tones, one to each ear, through a set of headphones, which differed in frequency by 10.5 Hz. The instrument was tested for validity and reliability on an oscilloscope and found to meet adequate standards for both. EEGs were recorded bipolarly from occipital and temporal sites of both hemispheres (T3, T4, 01, 02 sites as per Jasper, 1958) by a Mind Mirror II EEG (Blundell, undated; Cade & Coxhead, 1979). After recording the EEGs on magnetic tape, the information was converted to digital form and computer analyzed. Design and Procedure Sixty subjects received brief relaxation response training based on a handout they were given, and randomly assigned to one of four groups: (a) alpha frequency binaural-beat stimulation, (b) visual, eyes-open alpha brain-wave biofeedback, (c) both alpha-frequency binaural beats and alpha biofeedback, or (d) artificially produced surf sounds. The ratio of males to females was kept constant for all groups. The procedure for each subject consisted of the following steps: (a) the subject completed a pre- test of subjective mental and physical relaxation, (b) the subject was introduced to the four components of the “relaxation response” (Benson, 1975), (c) the subject was introduced to the stimulus which served as the mental device to theoretically elicit the relaxation response (either alpha binaural beats, alpha biofeedback, both, or phased white noise), (d) the subject was connected to the EEG, (e) the subject was instructed to become comfortable, relax, and breathe slowly and deeply for three to four minutes, (f) a two-minute eyes-open EEG baseline was recorded, (g) the subject was provided with the appropriate stimulus and allowed to become oriented to the situation, (h) the subject engaged in a ten minute eyes-open session of attempting to passively allow the stimulus to serve as the mental device to elicit the relaxation response, (i) the subject was briefly interviewed concerning strategies being used and subjective experience of the procedure and possibly reminded of previously mentioned strategies, (j) the subject engaged in a second ten minute session identical to the first, (k) the subject was disconnected from the EEG, (l) the subject completed the Self-Report Form and was interviewed concerning subjective experience of the procedure. Data Analysis Hypotheses H(1), H(2), and H(3) were tested by utilizing a 2 X 4 mixed analysis of variance and appropriate follow-up procedures. The between subjects independent variable was be the specific stimulus used by the subject as a mental device to elicit the relaxation response. The within- subjects variable was the time of the sampling of the alpha production; baseline or treatment sample. The dependent variable was the alpha production of the subject. Hypothesis H(4) was tested by utilizing a 2 X 4 mixed analysis of variance with appropriate follow- up procedures. The between subjects independent variable was the stimulus used as the mental device and the within subjects independent variable was the time of testing; pre- or post- procedure. The dependent variable was the level of relaxation reported. Summary This study attempts to examine the effects of alpha-frequency binaural-beat stimulation combined with alpha-frequency brain-wave biofeedback on alpha production and subjective report of relaxation through the utilization of a 2 X 4 mixed ANOVA design. It seems plausible that the combination of visual alpha feedback and alpha binaural beats will enhance the frequency- following response and assist the subjects voluntarily entrain their cortical rhythms to the stimulus. Analysis of the Data Demographics of Subject Sample Sixty volunteer subjects, forty females and twenty males, from various Memphis State counseling classes and from two Christian Brother’s College introductory psychology classes participated in the study. Volunteers from Christian Brother’s College were offered extra credit for their participation. Subjects were solicited by the author to participate in a study of the relaxation response. Age of subjects ranged from eighteen to forty-five with a mean of 27.7, a mode of 19, and a standard deviation of 7.64. Data was gathered between October 6, 1989 and October 21, 1989. Data Analysis Techniques Subjects were randomly assigned to four treatment groups of fifteen, each with ten females and five males. Each of the four groups received brief relaxation training followed by one of four treatments, a) alpha-frequency binaural-beat stimulation, b) alpha-frequency brain-wave feedback, c) alpha-frequency binaural beats with alpha-frequency brain-wave feedback, or d) artificially produced ocean surf sounds. Baseline and treatment alpha production ratios were obtained as well as pre- and posttreatment measures of subjective experience of mental and physical relaxation. The data was analyzed using the Statistical Package for the Social Sciences X (SPSSX) analysis of variance and follow-up procedures (Norusis, 1988). Since the form of the alpha production scores was proportional, arcsine transformations were performed on the alpha ratios prior to analysis in order to promote homogeneity of error variance and normality of error effects and to obtain additivity of effects (Kirk, 1982). For the experimental effects which achieved significance, the omega squared statistic was computed to indicate the strength of the associations (Kirk, 1982). Assumptions The mathematical model upon which the SPSSX analysis of variance procedures rest assumes that the error effects are distributed normally in the treatment population, independently determined and distributed in the treatment population, and vary homogeneously in the treatment population. The degree to which these assumptions were met affects the validity of the findings. Homogeneity of variance. Homogeneity of variance is a major assumption underlying the SPSSX analysis of variance procedures. The Bartlett-Box F test for univariate homogeneity of variance was used as a starting point for testing this assumption. The results of this procedure are reported in Table 1. Table 1 Bartlett-Box F Test for Homogeneity of Variance Measure F P Alpha Production Baseline 1.109 .344 Treatment 2.305 .075 Relaxation Scores Pre-test 0.121 .948 Post-test 0.735 .531 The significance levels indicate that there is no reason to reject the hypothesis that the variances in the two groups are equal. However, an additional test which examines the variances and covariances simultaneously is necessary in order to sufficiently test for homogeneity of dispersion (Norusis, 1988). Homogeneity of dispersion. Homogeneity of dispersion matrices must be considered when using multivariate analysis of variance (Norusis, 1988). Box’s M test is based on the determinants of the variance- covariance matrices in each cell as well as the pooled variance-covariance matrices, thus providing a multivariate test for the homogeneity of the matrices. The results of this procedure are presented in Table 2. As indicated, there appears to be no reason to reject the hypothesis that the variance-covariance matrices are equal across all levels of the between-subjects factors. We can conclude, therefore, that the assumption of homogeneity of variance of the error effects is not violated in this data set. Table 2 Box’s M Test for Homogeneity of Dispersion Measure F P Alpha Production 1.128 .338 Total Relaxation 0.317 .970 Hypothesis 1 It was hypothesized that alpha-frequency binaural- beat-stimulation would increase alpha brain wave production above eyes-open baseline levels. Table 3 shows the results of the 2 X 4 SPSSX repeated measures ANOVA of alpha production. Table 3 ANOVA Summary for Alpha Production Ratios Source SS df MS F Between .01 3 .003 1.07 Error .20 56 .004 Within .04 1 .04 101.84* Interaction .01 3 .003 4.16** Error .02 56 .0004 *p < .01 **p < .05 Between effect showed no significant differences among the groups, indicating that all groups were essentially equal in their baseline alpha production. However, the within effect, the difference between baseline and treatment alpha ratios, was significant (F(1,56) = 101.84; p<.01). Table 4 displays the group means for baseline and treatment alpha production. Table 4 Mean Alpha Production Ratios | Baseline | Treatment | Marginal* Group| Mean SD | Mean SD | Mean A | .081 (.033) | .114 (.044) | .098 B | .073 (.028) | .092 (.033) | .083 C | .084 (.040) | .134 (.052) | .109 D | .075 (.045) | .127 (.068) | .101 -----+------------------+--------------------+------------ * | .078 (.036) | .117 (.052) | .098 -----+------------------+--------------------+------------ *row and/or column averages Additionally, a significant interaction effect was found (F(3,56) = 4.16; p<.05), indicating that significant differences were present in cell group means. Post-hoc analysis was accomplished by the SPSSX one-way analysis of variance follow-up procedure. As demonstrated by Table 5, the treatment alpha production ratio of Group A was found to be significantly higher than the baseline alpha production ratio (F(1,56) = 93.34; p<.01). Thus Hypothesis 1 was not rejected. Omega squared for the effect ( = .613) indicates that we can conclude that the treatment for group A accounts for about 61% of the variance in the alpha production scores. Table 5 ANOVA Summary for Followup on Group A Source SS df MS F Between groups .0327 1 .0327 93.3429* Error .0196 56 .0004 *p < .01 Hypothesis 2 It was hypothesized that visual eyes-open alpha biofeedback training will increase alpha production above eyes-open baseline levels. As demonstrated by Table 6, post-hoc analysis reveals that the treatment alpha production ratio of Group B was found to be significantly higher than the baseline alpha production ratio of Group B (F(1,56) = 30.94; p<.01). Thus Hypothesis 2 was not rejected. Omega squared ( = .346) indicates that the treatment accounts for about 35% of the variance in the alpha production scores of Group B. Table 6 ANOVA Summary for Followup on Group B Source SS df MS F Between Groups .0108 1 .0108 30.9429* Error .0196 56 .0004 *p < .01 Hypothesis 3 It was hypothesized that the combination of visual eyes-open alpha biofeedback training with alpha frequency binaural beats stimulation will interact to increase alpha production more than either technique alone. As demonstrated by Table 7, post-hoc analysis reveals that the treatment alpha production ratio of Group C is significantly higher than the baseline alpha production ratio (F(1,56) = 214.29; p < .01). Omega squared ( = .785) indicates that the treatment effects account for about 79% of the variance in the alpha production of Group C. Table 7 ANOVA Summary for Followup on Group C Source SS df MS F Between Groups .0750 1 .0750 214.286* Error .0196 56 .0004 *p < .01 Follow-up analysis of the interactions between groups is displayed in Table 8. As indicated, a significant interaction was found (F(3,112)=2.6914; p < .05). Table 8 ANOVA Summary for Followup on Interaction Source SS df MS F Between Groups .0153 3 .0051 2.6914* Error .2128 112 .0019 *p < .05