RESULTS
This section covers how different baseline, training, enhancement and questionnaire scores are determined and labeled. It discusses if subjects were able to increase their training scores in the enhancement condition above the baselines and if enhancement condition EEGs were significantly higher than the stabilization condition EEGs. It also discusses the use of baselines as covariates and which covariates are significant. It covers the significant correlations between baseline scores and experiential scores and the significant correlations between enhancement scores and experiential scores as well as possible patterns of correlations. It illustrates how "left-movers" and "right-movers" differ on EEG scores for the two conditions as well as how they differ on the various correlations between experiential and EEG scores. Lastly, this section discusses reliability and if any changes were found in dependent variables over days.
hypotheses
(A) Primary Question:
Subjects trained to enhance synchronous alpha (higher synchronous alpha) will have significantly different scores on experiential questionnaire E-1 than when trained to stabilize this signal within a certain amplitude range close to their average eye-closed baseline (lower synchronous alpha).
(B) Secondary Questions:
Subjects trained to enhance synchronous alpha will be more likely to have the following basic changes in subjective experience than in the control contingency:
I. Merging With One's Experience:
A. An increased subjective sense of "merging" with their
experience.
B. A decreased subjective sense of being separate from their
experience.
II. Widening of Attentional Focus:
A. Widening of "within-sense" attentional focus
B. Widening of "between-sense" attentional focus
C. Increased "integrated" awareness"
D. An overall widening of attentional focus
scores
BASELINE average, baseline MAXIMUM and EEG training scores (see table 1)
during each daily session, three average baseline scores (average synchronous alpha amplitude levels) were taken, two before the EEG training (b1, b2) and one after (b3). Also two average EEG training scores were taken, one for the "A" condition (AC) (enhancement) and one for the "B" condition (BC) (stabilization). The three baseline scores were averaged together to produce one overall EEG average (b123). As suggested by Plotkin, the highest one minute epoch was taken for each baseline thus giving each baseline an average score and a maximum score (maxb1, maxb2, maxb3). As with the baseline scores, an overall average of all three maximum scores were taken (maxav), as well as the absolute maximum of all three (maxa) (see table 1).
determination of Enhancement scores (see table 2)
"Enhancement score" does not refer to the enhancement condition but rather to training scores from either condition, standardized by being divided by baseline scores. Standardized scores from both the stabilization condition and the enhancement condition are labeled "enhancement scores."
Four "average-based" enhancement scores were calculated for each EEG training score. this was done by dividing the daily average EEG training score for a condition, by each of the three baselines (b1, b2, b3) and by the average of all three (b123). This was done to standardize the EEG scores. The study focuses only on relative changes from daily baseline.
In addition, three "maximum-based" enhancement scores were calculated by dividing the training scores by the maximum scores of b2 (labeled "maxb2"), by the average of the maximum scores from all three baselines (labeled "maxave") and by the absolute maximum of all three baselines (labeled "maxa").
Each EEG training score therefore had a resultant seven enhancement scores for each condition: e.g., ac/b1 (EEG score for condition "A" / baseline 1), ac/b1, ac/b3, ac/b123, ac/maxb2, ac/maxave, and ac/maxa, for condition A. Subsequently, all comparisons involve all seven enhancement scores (see table 2).
why so many enhancement scores were used The decision to use all seven enhancement scores in comparisons was made because, unexpectedly, all the baseline averages were significantly different from each other, as were the maximum scores. Had the baselines not been significantly different, only the one overall average-based enhancement score and one overall max-based enhancement score would have been used. It was also subsequently found that many significant correlations would occur using one enhancement score (e.g. AC/b3) and not using the others.
experiential scores
Subjects rated their experience on 13 experiential dimensions for their experience in condition "A" and condition "B"; hence, as each enhancement score had an "A" and a "B" score, each dimension also had an "A" score and a "B" score (see table 3).
covariate scores
During the daily training, when the subjects rated their experience on the thirteen experiential dimensions, they also rated their experience on eight "covariate dimensions". As with the enhancement scores and the experiential scores, each covariate dimension had a score for condition "A" and for condition "B" (see table 4).
CORRELATIONAL scores
Since experiential scores are for conditions "A" and "B", and since all EEG scores, except baseline, are for conditions "A" and "B", four basic types of correlation scores can be seen in the results tables (see table 5). some of the correlational data tables are also divided into four quadrants for easy viewing of results.
Partialled and non-partialled correlations
Correlational scores can be classified into two additional categories: non-partialled correlations which do not control for the "effect" of the three baseline, and partialled correlations which do control for the baselines. partialled correlations were included when it was found that baseline scores were even more related to experiential scores than "enhancement scores."
comparing Eeg training scores to baselines and to baseline maximum scores
A manova comparing unstandardized EEG training scores for "A" condition (enhancement) and "B" condition (stabilization) to the three baselines showed that neither training condition exceeded grouped daily average baseline scores (see table 6). Training scores for "A" condition were significantly larger than for "B" condition. Eeg training scores for "A" condition were not significantly different from the grouped baselines (BL), however "B" condition was significantly below the grouped baselines.
comparing "A" condition training scores to individual baselines, "A" condition scores were significantly above baselines b1 and b2 but significantly below b3. B3 was taken after the daily training (Most studies have not taken post-training baselines), and it was usually the highest of the three baselines for most subjects. In addition, all three baselines scores were significantly different from each other, and daily training scores were significantly lower than the daily baseline maximum scores.
tests involving experiential measures
Differences in experiential scores between "A" & "B"
A manova was done comparing overall scores on experiential questionnaire E-1 between conditions ("A" and "B") and between groups ("left-movers" and "right-movers"). The Manova used a repeated measure within-subjects design, controlling for the eight covariates (mood, arousal, abundance of feedback, oculomotor strategy, sensory deprivation, expectation and motivation, perceived success, and perceived task difficulty). The manova included univariate f-tests of the "A" and "B" values of the thirteen experiential sub-dimensions of the E-1 questionnaire (table 7).
No overall differences were found between conditions (see table 7). However, there was a significant interaction between conditions and groups (eye-movement preference) (see table 8).
When "right-movers" and "left-movers" were looked at as groups, only "left-movers" significantly differed between enhancement and stabilization conditions (see table 9). however, no significant between-condition effects were found for any of the univariate F-tests for experiential dimensional subscales for "left-movers" or "right-movers".
"Right-movers" and "left-movers" had significantly different experiential scores from each other, overall, and on several experiential subscales (see table 10).
testing "standard covariates"
Covariates that affected experiential questionnaire scores
Univariate F-tests of significance of the covariates were included in the manova. Significant covariates were: "perceived success", "perceived task difficulty", "expectation-motivation", "perceived abundance of feedback", "arousal", and "sensory deprivation". table 17 shows the covariates that accounted for a significant amount of the variance in the experiential dimensions for questionnaire "e-1".
between-condition differences of covariates
Covariates, as a group, differed significantly between conditions. the experience of "mood", "arousal", "expectations and motivation", "task success", and "task difficulty" were significantly different between groups (see table 11). This shows that "task difficulty" was not equated between enhancement and stabilization conditions, as had been attempted through the pilot study. the results also show that the "abundance of feedback" was not experienced differently between conditions. In addition, there was no significant correlation between any experienced use of an "oculomotor strategy" and enhancement scores.
correlation of covariates with enhancement scores
Without using baselines as covariates, a number of covariates correlated significantly with the enhancement scores. Perceived "arousal" during condition "A" correlated positively with enhancement scores for condition A; perceived "abundance of feedback" correlated positively with enhancement scores under all conditions. "Expectation and motivation" during condition "B", unexpectedly correlated positively with enhancement scores for condition "A". "Perceived success" of conditions "A" and "B" correlated negatively with enhancement scores only from condition "B" (see table 18). This last correlation shows that, the more subjects perceived their performance as successful in either condition, the lower their enhancement scores were for condition B. When baselines were controlled for in the correlations for the whole group, the only correlation that was found was with arousal for "A" enhancement scores (table 22).
tests involving enhancement scores
Differences in enhancement scores between conditions and between groups
Enhancement scores were significantly different between conditions "A" and "B". A manova, testing all enhancement scores (standardized daily EEG training scores) for differences between conditions ("A" and "B") and between groups ("left-movers" and "right-movers"), showed a significant overall difference between conditions. also, the univariate F-tests showed that the seven pairs of enhancement scores for "A" and "B" conditions (e.g. ac/b1 compared to bc/b1) were significantly different (see table 12).
The above-mentioned overall effects, however, appear to be more do to "left-mover" than "right mover" scores. Overall differences in enhancement scores between conditions were found for "left-movers" but not for "right-movers" (see table 13). Looking at individual enhancement score comparisons, all were significant for "left-movers." for "right-movers", all individual enhancement score comparisons were significant except ac/b123 vs. bc/b123 (see table 13.
differences in correlations between conditions -- correlation of enhancement with experiential scores (not controlling for baselines)
Enhancement scores for condition "A" correlated negatively with questionnaire "E-1" scales for "within-sense field size, "between-sense field size," "multisensory perception," "incorporation of opposite states," "exclusion," "concentration," and positively with "attention to weaker stimuli" and " between-sense simultaneous" (see table 23).
Enhancement scores for condition "B" correlated negatively with questionnaire "E-1" scales for "within-sense field size," "multisensory perception," "between-sense field size," "incorporation of opposite states" and positively with "attention to weaker stimuli" (see table 23).
there were twice as many significant correlations with the array of seven "A" enhancement scores than the array of "B" enhancement scores. The largest correlation was between the subject's experience of the "incorporation of opposite states" and enhancement scores for condition "A" (negative). (see table 23)
"left" and "right-movers" both "left" and "right-movers" had relatively strong correlations between enhancement scores and "incorporation of opposite states." "Left-movers" show a strong correlation to "attention to weaker stimuli." for "left-movers", correlations also occurred for "merging" and "separating" with "A" enhancement scores.
why baselines were included as covariates
Correlations of baseline and experiential scores
Unexpectedly, experiential scores frequently correlated with baseline scores as well as with enhancement scores (see table 21). Table 21 also includes correlations with experiential dimensions from the PCI. the dimensions from the pci will not be focused on here, but are included to show that the baselines correlate with quite a few other subjective dimensions besides the ones in questionnaire E-1. Baseline three ("b3") tended to have the highest correlations, especially on the PCI. The majority of the time, correlations increased progressively from "b1" to "b3". "B3" was the baseline given after the training, and it usually had the highest correlations. significant correlations with maximum scores (see table 24) were basically the same as with baseline scores except they tended to be higher. Also, most of the correlations were negative here and throughout the study. overall, experiential dimensions had much higher correlations with baselines than with enhancement scores. Many correlations of experiential scores with
baselines were between -.2 and -.3, while correlations with enhancement scores tended to have fewer correlations in this range.
When comparing correlations of experiential scores with baselines, with the correlations experiential scores with unstandardized training scores, it was found that both baseline and training score correlations were close in terms of which dimensions they correlate with and the degree of correlation (see tables 19 and 27). In other words, there was essentially little difference between training and baseline scores in terms of correlations with experiential scores. This strongly suggests that enhancement score correlations without controlling for baselines were more a function of baselines rather than training.
Because of these results, the decision was made to run many of the correlations again using baselines as covariates. any significant results from these partial correlations could be assumed to be due to EEG changes in the training condition and not from changes in baselines. These partial correlations were done, first, correlating experiential scores with the unstandardized training scores and second, correlating experiential scores with the standardized enhancement scores. The significant results of these two types of correlations were, for the most part, identical. The correlations using the standardized enhancement scores were chosen to use for the presentation of the results.
using baselines as covariates in correlations
training and enhancement scores
when baselines were used as covariates when correlating training and enhancement scores with experiential scores, most of the original correlations were no longer present, (see tables 14 and 25) and one new one appeared, indicating that many of the previous significant correlations were not due to training influence but were simply related to the amplitudes of the subjects baselines. Three significant correlations with training scores and with enhancement scores remained, and they were all negative and all with training (or enhancement) scores for condition "A". these correlations were with "separation" for "A", "multisensory perception" for "B", and "incorporation of opposite states" for "A" (and "B" for enhancement).
major contrasts
When baselines were not controlled for, some experiential scores correlated positively with enhancement scores in one condition, and negatively in another, showing a large possible contrast between conditions. The strongest contrast was between "concentration" and enhancement scores. Enhancement scores for condition "A" significantly correlated negatively with "concentration" scores for condition "A", and non-significantly correlated positively with "concentration" scores for condition "B". In other words, as a subject's alpha increased for condition "A" (higher alpha), his experience of rigid concentration was less. The reverse was true for his experience of concentration for condition "B" (lower alpha), but only in relationship to "A" enhancement scores.
Other contrasts were with "within-sense simultaneous" for condition "A" and with "fragmentation- integration" for condition A" both of which had significant negative correlations with "A" enhancement scores and non-significant positive correlations with "B" enhancement
scores (see table 15). Since these contrasts only occurred when baselines were not controlled for, it cannot be definitely said that these correlations are due to the effects of the training.
possible correlational patterns for entire group, right and "left-movers"
correlations for the entire group, "left-movers" and "right-movers", tended to be entirely in certain halves or quarters of the correlational charts and tended to be positive or negative in that section (see table 16). For the entire group, all significant correlations were negative and were only for enhancement scores for condition "A" (left half of table). For "right movers" all significant correlations were for "A" experiential scores (upper half of table). The negative correlations were for "A" enhancement scores (upper left quadrant) and the positive were for "B" enhancement scores ( upper right quadrant). For "left-movers", all significant correlations were for "A" enhancement scores and "B" experiential scores (lower left quadrant). Because of the limited number of significant correlations, these tendencies were not conclusive.
Reliability and validity of experiential questionnaire E-1 and of the "pci"
The "PCI" (Phenomenology of consciousness inventory) scores were looked at because they measure experiential dimensions similar to those on the "E-1" questionnaire. In addition, the "PCI" includes five near-identical pairs of questions to measure reliability. Also, the "PCI" has had past validity studies significantly correlating its scores with hypnotic susceptibility (pekala 1975).
Reliability analyses were done for the five pairs of reliability items on the "PCI". All alpha reliability values were greater than seven showing a high internal consistency reliability.
Similar experiential dimensions on experiential questionnaire "E-1" had relatively high and significant correlations with the "PCI" experiential dimensions (see table 29), and some of these and other "PCI" significantly correlated with baseline and enhancement scores (see
tables 21 and 30). Since the "E-1" has no direct reliability or validity tests or history, its reliability and validity can only be inferred from its high correlations with similar measures on the "PCI".
"Constant (pre-baseline) factors"
"constant factors" were three dimensions from the pre-baseline questionnaire where subjects were asked to rate their overall expectations and motivation, arousal and mood prior to the baselines and to the training. These questions did not pertain to "A" or "B" condition; they just assessed the subjects' overall condition prior to the training hence they were "constant" between conditions. When these three dimensions were correlated with enhancement scores, (controlling for baselines) "expectation-motivation" correlated positively with one enhancement score for condition "B" ("maxave") (see table 31).
days
An array of tests were done to determine if there were any changes in dependent variables over days. No significant change in EEG scores, enhancement scores or experiential questionnaire scores was found over days, overall, for conditions or for groups.