Jan 20 2011


I recently received this question in my Topics Suggestions page:

Dear Dr. Novella,

An interesting and rather weird neurological diagnostic technique…


I would be interested in hearing your thoughts.

dave, Melbourne Australia

Thanks for bringing this to my attention, Dave. Electrovestibulography, or EVestG, is a technique that measures and digitally analyzes the electrical signals produced by the vestibular system during certain manipulations, such as changes in head position. The vestibular system is that part of the brain and inner ear that senses the orientation of the head with respect to gravity, and also acceleration. The primary sensory organ is three semicircular canals, each oriented along a different three-dimensional axis. These canals are filled with fluid (endolymph) and are lined with neurons that project tiny hairs into the canals. When the fluid moves in response to head rotation they flow past the hairs, bending them and triggering neuronal activation. Separate structures, the otolithic organs, respond to gravity or linear acceleration. This information is processed by the brain to give us the subjective sense of our own position with respect to gravity and acceleration.

There already are standard tests for investigating vestibular function and localizing any dysfunction. Vestibular dysfunction usually results in vertigo – a subjective sense that either one is moving or the world is moving or spinning. (The colloquial term “dizzy” is often used to describe this sensation, but “dizzy” is a vague word with multiple meanings, so “vertigo” is the preferred technical term.”)

What these inventors appear to have done is devise a system that measures the electrical output of the vestibular system during tilting or other maneuvers and then analyzes that output in order to search for patterns of activity. I have not found the technical information online to know if this basic part of the claim is true or not, but it is plausible. Let’s assume that their machine does what they claim it does – measure and analyze the electrical activity  of the vestibular system.   This remains to be determined, but there is nothing surprising about this claim.

The question is – what are the clinical applications of this new diagnostic technology. The obvious application will be for diagnosing vestibular problems. This application is reasonably plausible, but cannot be taken for granted. Looking for patterns of electrical activity that act as a reliable marker for specific vestibular dysfunction is not easy, and requires a great deal of clinical study. We would need to know what the sensitivity and specificity of the test is in a variety of vestibular conditions, and then see if the test has validity – meaning that is predicts something. Do the results of the test correlate in any way with prognosis or response to specific treatments? Do they correlate with any gold standard tests already in use, or with any pathology that can be demonstrated in other ways?

I did a PubMed search on “Electrovestibulography” and came up with 6 papers. All six appear to be by the same team of researchers – the developers of the technology, which is a bit of a red flag. It could just mean that the technology is new and has not yet garnered attention from other researchers, and therefore there has not yet been any replication. This is fine – but means that we need to be very cautious about the claims until such replication can be done.

Of the six papers, three deal with vestibular disease, two in animals and one claiming that their technique can distinguish between Meniere’s disease and benign positional vertigo – two different vestibular diseases. Sorry – I only have access to the abstract at the moment, which states:

“Separation of Meniere’s disease and benign paroxysmal positional vertigo (BPPV) has been accomplished using a “DC” biomarker.”

The abstract does not give any details as to the type of study that was done, but the absence of any details is concerning – no methods, no results, just a claim. It seems like an introductory paper just outlining the technology, and not the presentation of a detailed study.

If this were the end of the story I would say that this is a new and interesting technology that requires some investigation and replication to determine if it works as claimed, and to see if it really does detect biomarkers of specific vestibular conditions. So far I am not impressed with the data being presented, and I would be concerned that the developers are going the commercial rather than the academic route. But I would keep an open mind until actual detailed studies are done.

But that’s not the end of the story. The developers of this technology then go on, from the most preliminary of data, to expand the potential clinical uses of their technology to the diagnosis of schizophrenia, depression, and bipolar disorder with suggestions that it may also work for Alzheimer’s disease and Parkinson’s disease. This would mean that there are discernible vestibular biomarkers not only for vestibular diseases, but a host of distinct and very different neurological disorders. Plausibility has now plummeted and the red flags are flying.

In one study of schizophrenia they look at 4 affected subjects and 10 controls, without any apparent blinding. This sample size is too small to be of any real use, and is questionable if it should even be published – especially given that their outcome is not something concrete (like death or survival) but is the very complex action of pattern recognition – distinguishing different patterns of activity between schizophrenics and normal controls. This type of analysis would require many more subjects (50 or more sounds about right) and blinded analysis before I would take it seriously.

I am concerned that before the developers of this technology have done extensive research to validate its use for the most obvious application – vestibular dysfunction – they have done pilot studies for depression, schizophrenia, and Parkinson’s disease. So now we have a small group of almost worthless pilot studies, rather than a series of increasingly tight and powerful studies for the most likely application. This is an odd way to proceed – unless one is trying to bolster the commercial appeal of a proprietary device by generating lots of low grade evidence for a wide range of applications.

Further – the evidence they have provided so far is akin to retrodicting – taking known quantities then looking for patterns. Astrologers can do this quite well – use astrology to “explain” why something happened in the past. But the real test of any system is its predictive power, and we have no data on that so far. The developers may be falling victim to simple pattern recognition and confirmation bias. They appear to be so impressed with the technical aspects of their device that they are neglecting the medical aspects – the plausibility of their claims and the type of data that would be needed to really substantiate them.

I predict that electrovestibulography will become just another fringe crank device with overblown and dubious claims. It seems to be heading rapidly in that direction. It may also have some genuine utility, but that is yet to be determined by proper studies. I would recommend to the developers that they focus on vestibular disease and perform (in collaboration with outside experts) large and well-designed studies to determine if it has true utility for such applications. Demonstrate first that the thing works and is useful. Then you can explore the more unlikely claims.

12 responses so far