Neural Bases of Hypnosis

Helen Crawford

Dear Pierre:

Thanks for your comments regarding our theoretical model that proposes  that hypnotic analgesia is an active inhibitory process that operates to reallocate thalamocortical activities.  Most certainly, other inhibitory pain systems are actively interacting with the frontal attentional system (as we discussed in the IJCEH 98 paper).  You wrote (see below) that in our 1993 rCBF study (ref below): "you found an INCREASE in somatosensory cortex activity during hypnotic analgesia using an ischemic model" and question the findings fitting into our model.  The increased rCBF during hypnotic analgesia we observed in the general region of S1/S2 cannot be compared directly with the SERP results.      

rCBF, fMRI, and PET activation does not mean "activity" as we think you may have inferred.  The activation observed with these neuroimaging techniques is an indirect measure of brain activity reflected by hemodynamic changes.  The observed activation may reflect either inhibitory or excitatory activity that is unique to the condition under evaluation.  However, one cannot determine which it is from these neuroimaging approaches.  

As you note, the PET and fMRI literature on pain has been inconsistent in observing activation in S1 (primary somatosensory, usually BA 3, 1, 2) and S2 (secondary somatosensory, primarily BA 5) regions.  This may be a reflection of study differences in methodology (type of stimulus, level of pain perceived [some studies give low pain -- see Becker et al's discussion of ERP pain research], statistical analysis cutoff points), choice of subjects, etc.  A recent PET study by Iadarola and his colleagues at NIH (ref below) provides an excellent overview of the literature and interprets their data into a systems approach.  In our fMRI studies of pain (refs below) we do see activation in S1 in both attend and hypnotic analgesia, but the amount of activation observed varies across subjects.    

Somatosensory event-related amplitude data cannot be directly compared to the above approaches.  SERP approaches reflect electrical activity recorded at the scalp or intracerebrally, whereas the fMRI, PET and rCBF approaches reflect various aspects of the hemodynamic exchange.  

Chen and others have demonstrated quite well a correlation between level of perceived pain and amplitude of the P250 (or P200, P300)SERP component.  Quite consistently across the literature, we find attentional deployment(including the disattentional techniques of hypnotic analgesia) leads to a reduction of the amplitudes of these positive SERP components.  In our research (e.g., IJCEH 1998) we have demonstrated an increase in the negativity of N140 (anterior frontal) and N250 (fronto-central) SERP components suggestive of increased inhibitory processing.  Kropotov et al. (1997) found increased negativity in intracerebral recordings from the anterior temporal region.  The N140 component is thought to reflect the "complex reciprocal interactions between posterior and prefrontal [anterior frontal] cortex and subcortical structures" that play "a key role in governing sequential attention processes" (Desmedt & Tomberg, 1989, p. 343).    

Future research that combines imaging techniques (e.g., fMRI and ERP) should assist us in illuminating the relationships and differences between dramatically different approaches to recording brain activity.    
Many Thanks!

Helen J. Crawford
James E. Horton

references:

Crawford, H.J., Horton, J.E., Harrington, G.S., Vendemia, J.M.C., Plantec, M.B., Yung, S., Shamro, C., & Downs III, J.H. (1998).  Hypnotic analgesia (disattending pain) impacts neuronal network activation:  An fMRI study of noxious somatosensory TENS stimuli. NeuroImage, 7, 436.

Crawford, H. J., Gur, R. C., Skolnick, B., Gur, R. E., & Benson, D. (1993).  Effects of hypnosis on regional cerebral blood flow during ischemic pain with and without suggested hypnotic analgesia.  International Journal of Psychophysiology, 15, 181-195.

Crawford, H. J. Knebel, T., Kaplan, L., Vendemia, J., Xie, M, Jameson, S., & Pribram, K. (1998).   Hypnotic Analgesia:  I. Somatosensory Event-Related Potential Changes to Noxious Stimuli and II.  Transfer Learning to Reduce Chronic Low Back Pain.  International Journal of Clinical and Experimental Hypnosis, 46, 92-132.

Desmedt, J. E., & Tomberg, C. (1989).  Mapping early somatosensory evoked potentials in selective attention: critical evaluation of control conditions used for titrating by difference the cognitive P30, P40, P100 and N140. Electroencephalography and Clinical Neurophysiology, 74, 321-346.

Downs III, J.H., Crawford, H.J., Plantec, M.B., Horton, J.E., Vendemia, J.M.C., Harrington, G.S., Yung, S., & Shamro, C.  (1998).  Attention to painful somatosensory TENS stimuli: An fMRI study.  NeuroImage, 7, 432.

Iadarola, M.J., Berman, K. F., Zeffiro, T. A., Byas-Smith, M. G., Gracely, R. H, Max, M. B., & Bennett, G. J. (1998).  Neural activation during cue capsaicin-evoked pain and allodynia assessed with PET.  Brain, 121, 931-947.

Kropotov, J. D., Crawford, H. J., & Polyakov, Y. I.  (1997)  Somatosensory event-related potential changes to painful stimuli during hypnotic analgesia:  Anterior cingulate cortex and anterior temporal cortex intracranial recordings.  International Journal of Psychophysiology, 27, 1-8.

______________________________

Previous Memo from Rainville to Crawford:

On Fri Dec 11, Pierre Rainville wrote
-------------------------------------
>Dear Dr Crawford,

>I think that your suggestion that hypnotic analgesia involves frontal control of thalamocortical is very interesting and promising. The frontal activation that you first reported in your 133-xe study in 1993 is replicated in our data showing massive frontal increase in rCBF associated with suggestions for pain modulation.

>However, there is one important result in your brain imaging study that seems difficult to reconcile with the rest of the litterature on SERP and with your and our more recent brain imaging results. In your 133-xe study, you found an INCREASE in somatosensory cortex activity during hypnotic analgesia using an ischemic model. If I remember correctly, this increase was modulated by the hypnotic susceptibilty of subjects, right? In contrast, pain-evoked activity in other studies is found to decrease with hypnotic analgesia, and activity in pain-related regions (S1, S2, Insula and Anterior cingulate) has been found to correlate positively with pain perception. In your early study, activity in S1 appears to be inversely related to pain perception. Did I get this right? If so, what could be the difference between that study and the more recent ones that might explain this discrepancy?

>There is ongoing controversy concerning the involvement of S1 in pain processing and early brain imaging studies of pain-evoked changes showed both increases and decreases in S1. Part of the answer might rely in the dynamic of activation in S1 with possibly an early activation at stimulus onset and a late de-activation (e.g. see an early study by Apkarian). Presumably, your study have been more sensitive to the late part of the stimulation. However, I am not sure how this dynamic effect could be integrated in your model. Do you have any insight on how these results might reconcile with your model of thalamocortical inhibitory processes?

>Pierre R.
>

[ This message was edited on Wed Dec 16 by the author ]