May 22, 2010 (HP31).  In Superconductor Week (SW) appeared recently (Issue 2406 of April 24th, 2010) a news item on the use of magnetoencephalography (MEG) at the National Institute of Mental Health (NIMH, USA) to study cortex activity of schizophrenics.  This interesting news contained quotations of statements by Dr. Dietmar Plenz of NIMH.  One could also read that “This study might pave the way for MEG technology to drastically ease the diagnosis of schizophrenia”.  While this statement was not directly attributed to Dr. Plenz, it could lead the reader to believe that a breakthrough towards clinical use of MEG is imminent.  Indeed, such a breakthrough would have been welcome news to companies still developing and supplying MEG systems, and could be also of significance to many current MEG users.

Noninvasive functional imaging diagnostics utilizing arrays of SQUID sensors, such as MEG and MCG (the magnetocardiography), has been for many years holding promise of broader use in clinical medicine.  One clinical application of MEG, the pre-surgical localization of epileptic foci (centers of activity) has, in fact, received an insurance reimbursement code in the US.  Nevertheless, the hopes of a broader use of MEG in medical diagnostics, and thus of a wider market for MEG systems, did not materialize thus far.  As of today, MEG remains a par excellence research tool that is available only in slightly more than 100 best-equipped laboratories around the world.  This situation recently led to the demise of two small companies involved in the development and manufacturing of MEG systems: in 2007 the CTF of Vancouver, Canada, and in 2009 the 4D-Neuroimaging of San Diego, USA.  The latter was the final reincarnation of BTi (Biomagnetic Technologies, Inc.), first to pioneer the development of multichannel SQUID systems for MEG in the 1980s.  These two companies supplied a large share of MEG systems currently used around the world.

The MEG situation briefly characterized above prompted us to follow the lead indicated by SW and to contact NIMH, specifically Dr. Richard Coppola, who is the Director of the NIMH MEG Core Facility at Bethesda, MD.  In turns out that the study referred to in SWwas presented at the 2008 meeting of the “Society for Neuroscience”.  The results have not been published in a paper, but Dr. Coppola, who is one of co-authors of the study, made the abstract of the presentation available here, so those interested can form their own opinion.  He emphasized that this study was aimed at investigating group differences and not at the diagnostics per se.  This research is still in progress, and more data is needed before full publication.  Furthermore, NIMH doesn’t run clinical trials, which would be a necessary prerequisite of any diagnostic claims.  Rather, NIMH is interested in group and perhaps individual differences that could potentially lead to treatment targets.  Certainly, Dr. Coppola is optimistic that MEG coupled with dynamic measurements will be very useful in extending the understanding of the neurophysiological substrate of schizophrenia.  Nevertheless, diagnosis of schizophrenia is not really the focus of most neuroimaging research.

Dr. Coppola commented also that there are several promising MEG studies of neuropsychiatric disorders, not just of schizophrenia, which eventually might lead to successful therapies.  However, the whole field is still far away from any clinical application.  He remains optimistic, but cautions against thinking that any “breakthrough” might be imminent.  As a historical note, when computerized EEG methods were introduced several decades ago, there were great expectations for massive clinical applications, which did not materialize.  The electrophysiological research has continued to be fruitful, but there was no singular breakthrough.

The final note of Dr. Coppola, no doubt of interest to readers involved in MEG, was that MEG as a research modality is most in need of improved software that will allow for simplified and more uniform analyses across laboratories.  This appears more pressing than any expensive hardware improvements resulting in multimodal operation, such as the MEG combined with low-field MRI.  In the best case, the latter could become an aid in co-registration, but not competitive with high-resolution structural MRI.  In neurological research, high-field, high resolution MRI data must be and are routinely collected from all investigated subjects.