Combined FDG PET / rTMS Study of Speech Processing Neurocircuitry in Schizophrenic Patients with Auditory Hallucinations

Ralph Hoffman, M.D.

Yale University, New Haven, CT

Grant Program:

David Mahoney Neuroimaging Program

Funded in:

September 2000, for 4 years

Funding Amount:


Investigator Biographies

Ralph Hoffman, M.D.

Associate Professor of Psychiatry, Yale University School of Medicine



Constant auditory hallucinations arise from excessive activation of multiple brain regions underlying speech processing that can be normalized by repetitive transcranial magnetic stimulation (rTMS).

1. To delineate components of speech processing neurocircuitry that are excessively activated in schizophrenic patients with continuous auditory hallucinations using [18F] fluorodeoxyglucose positron emission tomography (FDG PET).
2. To demonstrate that these hallucinations can be reduced or eliminated by low frequency rTMS positioned according to patient-specific cortical maps of excess metabolic activation.
3. To demonstrate that positive treatment response to rTMS is accomplished by normalization of metabolic brain maps of speech processing neurocircuitry.

The researchers plan to study fifteen schizophrenic patients with constant auditory hallucinations resistant to medication using FDG PET. Regional rates of brain metabolism for each of these patients will be compared to FDG PET data from a group of normal control subjects via a region of interest analysis and statistical parametric mapping. The two cortical regions most activated according to FDG PET data will be identified for each patient. A randomized, double blind, crossover research design will be used to compare active vs. sham rTMS delivered simultaneously to these two sites.

Follow-on Funding: NIMH 1R01 MH073673, rTMS clinical trial for auditory hallucinations. Hoffman, R. (PI), 2/06-1/11.

The Dana Foundation has supported critically important studies of patients with schizophrenia who experience hallucinations of spoken speech, or “voices.” This is a very disruptive and often disabling symptom that often does not respond to treatment and can lead to impulsive and dangerous behavior. Due to technical difficulties, we switched our neuroimaging approach from positron emission tomography to functional magnetic resonance imaging (fMRI). These methods were combined with repetitive transcranial magnetic stimulation (rTMS), which allows us to selectively stimulate small parts of the cerebral cortex using highly focused magnetic waves administered using powerful electromagnet positioned on the scalp.

Our fMRI study has identified a unique pattern of brain activation during hallucination events. Of special interest is that we have identified a discrete site in the right middle temporal cortex that reliably activates across patients just prior to onset of their hallucinations, with subsequent activation more widely distributed in superior temporal areas coinciding with the experience of hallucinations themselves. One interpretation of these data is that activation within this middle temporal site is a critical conduit or trigger leading to disseminated activation that then is responsible for the conscious experience of hallucinated speech.

A second study has used these fMRI maps to identify 3-6 cortical sites per patient that were probed with 1-hertz rTMS and sham stimulation. 1-hertz rTMS is known to reduce excitability or reactivity of brain areas accessed. Delivering 1-hertz rTMS to a region in the superior temporal lobe known as Wernicke’s area was accompanied by a greater rate of improvement in voices compared to sham (placebo) stimulation and rTMS delivered to anterior temporal sites. Some patients also responded to rTMS delivered to the right hemisphere equivalent of Wernicke’s area. These findings suggest that these specific regions are critically involved in producing hallucination experiences in schizophrenia, and that rTMS delivered to these areas could represent a possible treatment strategy.

We consequently have embarked on a more extended clinical trial funded by a grant from NIMH to study the usefulness of rTMS delivered to Wernicke’s area and the right hemisphere equivalent area in reducing hallucinated voices. 90 patients will be studied and will receive either active stimulation or sham (placebo) stimulation for a four week trial. An MRI anatomic scan is used to position the TMS coil on the scalp so that it precisely overlies these two brain areas. We currently are completing the first year of this grant and have obtained promising results showing much greater improvement in hallucinated voices for active rTMS compared to sham stimulation.

Selected Publications

Hoffman R.E., Hampson M., Wu K., Anderson A. Gore J., Buchanan R.J., Constable T., Hawkins K., Sahay N., and Krystal J.H.  Probing the pathophysiology of auditory hallucinations by combining functional magnetic resonance imaging and transcranial magnetic stimulation.  Cereb Cortex. 2007 Nov;17(11):2733-43 .

Hoffman R.E., Boutros N.N., Hu S., Berman R.M., Krystal J.H., and Charney D.S.  Transcranial magnetic stimulation and auditory hallucinations in schizophrenia.   Lancet. 2000 Mar 25;355(9209):1073-5 .

Hoffman R.E., Boutros N.N., Berman R.M., Roessler E., Krystal J.H., and Charney, D.S.  Transcranial magnetic stimulation of left temporoparietal cortex in three patients reporting hallucinated “voices.”  Biol Psychiatry. 1999 Jul 1;46(1):130-2 .