Neural Network and Coherence Study

EEG data from 17 sites across the cortex were collected at initial baseline and at specific points following a standard hypnotic induction for 18 individuals (11 high, 7 low). Our results show that classifications employing multi-channel spectral theta values yielded high accuracies of 94% for cascade-correlation, 93% for backpropagation, and 72% for discriminant analysis.  Classifications employing multi-channel fractal dimensionality values were less accurate showing 87% for backpropagation, 84% for cascade-correlation, and 79% for discriminant analysis.

Higher coherency magnitude was found at nine electrode pairs for the high-hypnotizable group, and at two electrode pairs for the low-hypnotizable group.  Direction differences were found in nine electrode pairs with long-distance pairs showing an anterior to posterior direction for the high-hypnotizable group, and the reverse for the low-hypnotizable group.  Time delay differences were found five electrode pairs, with all showing shorter propagation times for the high-hypnotizable group.

Overall, these results demonstrate that neural networks can be used to accurately classify hypnotic susceptibility and that the high and low susceptible individuals show differential cortical activity as reflected by coherence measures.

Neuropsychological Study

We administered four neuropsychological tests (WICS, Controlled Oral Word Association Test [or the FAS], Stroop, and the Towers of Hanoi) to high or low hypnotically suggestible individuals.  These tests were selected to reflect a diversity along verbal and visual- spatial modalities to test the range of potential differences in frontal tasks which might reflect executive functioning.

Participants were chosen by their performance during a group hypnotic screening, however, participants were not informed of the selection criteria when contacted to for the study.  Therefore, the present study was able to address executive function in hypnosis as an issue of individual differences.  To examine the relationships between tests, indices from each were correlated.  A significant positive correlation was found between the total number of moves on the Towers of Hanoi and the total correct moves on the Wisconsin Card Sort, r(15)=0.64, p<.008, perhaps reflecting similar visual spatial processing abilities that both tests measure.  However, both tests show weaker, non-significant correlations to the FAS and Stroop, potentially suggesting those test's function as verbally mediated frontal tasks (Stroop reading time had a small positive correlation to FAS totals, however, it was also non-significant) .

Overall, high suggestible participants performed better than the low suggestible participants.  The only significant finding was a faster performance for high suggestible participants with the WICS, t(14) = 2.21, p<.04, which would suggest a greater ability to detect relevant abstract information, shift cognitive set, and to decrease perseverative tendencies. However, while not significant, the high suggestible participants also performed better on both the FAS and Stroop, tests which should reflect verbally-mediated abilities.  This may indicate that the high suggestible participants differed from the low suggestible participants on general intelligence as well. Therefore, the results would indicate that the high suggestive participants were more flexible in their ability to shift set, plan complex tasks, and perform effectively. 

Our current research focused on two levels of analysis, that of cortical functioning and that of executive function. In terms of the former, our research suggest that high and low susceptible individuals can be distinguished using neural networks in terms of their EEG. Further, as shown by coherence measures, high and low susceptible individuals also display different patterns of cortical activity. In terms of the later level of analysis, the use of neuropsychological measures presented a less certain pattern that clearly distinguished high and low susceptible individuals although high susceptible individuals did appear to be more flexible in their ability to shift set. Future research might test out this suggestion by having individuals perform these tasks during an hypnotic procedure. Additional research from other labs examining hypnosis and source memory indicate that hypnotized individuals have less detail in their recall and remember the ‘gist' of information rather than details.  In regards to previous theory, this may indicate that easily hypnotizable people process information in a more fluid manner and that any differences in executive functions may  be relative to those more primary cognitive abilities.  Future research should continue to explore hypnosis as an execute function separate from general intelligence to better articulate the manner in which hypnotic processes can produce the inability to monitor or govern basic motor and perceptual functions.