In the 19th century, many people in the western world were terrified of the possibility that they might be buried alive, after physicians and family wrongly thought them dead. From Edgar Allen Poe’s story, Premature Burial, to the creation of the Society for the Prevention of Premature Burial, to the invention of various “safety coffins,” society played out this fear until improved medical knowledge (including, among other things, the invention of the stethoscope) led to the fear abating.

A modern version of this fear was perhaps seen in the Terri Schiavo saga, where the patients’ parents were insistent that their daughter was conscious in ways ignored by the medical staff and her husband – “buried alive” in a nursing home for those in a persistent vegetative state. The 19th century fear was fed by the occasional finding, when bodies were exhumed or graveyards moved, of a disturbed corpse, that had moved in the coffin or that had apparently damaged its hands in trying to escape or even of corpses altogether outside their coffins. Today’s fears may be fed by new functional magnetic resonance imaging (fMRI) studies.

Last week the New England Journal of Medicine published, on line, a fascinating, and disturbing, research report on the use of fMRI to detect – possibly – consciousness in people with profound brain damage.

This article builds on several years of research by two laboratories, those of Adrian Owen at Cambridge and Steven Laureys at Liege. Owen got attention a few years ago for a report in Science, showing that one woman, diagnosed as being in a vegetative state, had brain activation patterns in fMRI scans similar to those of healthy controls when asked to imagine herself walking through her home or playing tennis. Laureys has done a long series of studies of minimal consciousness and vegetative states. His work last garnered attention late last year when, as part of studying indicating a frighteningly high rate of patients in a minimally conscious state being diagnosed as being in the lower vegetative state, he showed that a man long thought to be in a vegetative state actually had moments of consciousness. (This story got some unwanted attention when others claimed that the man could actually communicate through “facilitated communication,” a controversial method where a third party helps “guide” the patient’s hand.)

Owen and Laureys worked together in this study and showed two different things of great interest.

First, they did fMRI studies of 54 patients (31 at Cambridge, 23 at Liege) who had been diagnosed as either in a vegetative state (23) or in a minimally conscious state (31). While these apparently unconscious and unresponsive were inside the scanner, they asked them both to think about hitting a tennis ball back and forth with an instructor and to think about walking through the rooms of their home or through the streets of a city they knew. In healthy people they would expect to see activation of the supplementary motor area in the first task and in the parahippocampal gyrus in the second. For five of the patients, they saw supplementary motor area activation during the tennis task; for four of those five (and no one else) they saw activation of the parahippocampal gyrus during the spatial task.

One of the five patients whose brains “responded” had been diagnosed as being in a minimally conscious state, but four of the five patients had been diagnosed as being in a vegetative state. On closer examination, the researchers decided that two of those four had been misdiagnosed and were actually in a minimally conscious state. The five responders had suffered their brain injuries 1.3 months, 2 months, 6 months, 30.2 months, and 60.8 months earlier. All five had had among the 32 subjects, out of 54, who had had traumatic brain injury. No one responded whose damage was caused by anoxic (lack of oxygen) brain injury (14), meningitis (3), cerebrovascular accident (2) brain stem stroke (1), or anoxic brain injury plus traumatic brain injury (1), or anoxic brain injury plus encephalitis (1).

Second, they took one of the five responders, a 22 year old man who had been diagnosed in a vegetative state (and who they agreed was appropriately so diagnosed) as a result of traumatic brain injury over 5 years earlier, and tried to communicate with him. Along with 16 healthy controls, he was asked yes and no questions, such as “Do you have any brothers?” All 17 subjects were told either to indicate yes by thinking about the tennis task or by thinking about the spatial task. (Which one was “yes” and which was “no” switched back and forth during each individual’s scan.) The healthy control subjects were able to “say” yes or no this way perfectly, based on looking for activations in either their supplemental motor areas (tennis) or their parahippocampal gyri (spatial). So was the man in a vegetative state. He answered the first five questions they asked him perfectly, although they got no response from him on the last question.

What does this mean?

Well, for the patient who was able to communicate “yes” or “no” it is very hard to see how this does indicate at least some consciousness. He not only processed the question but responded to it accurately. More trials will surely be useful to increase confidence in this result, but he surely seems “locked in” – conscious, at least occasionally, but with no power at the bedside to make an visible response.

We do not know whether the other four patients could communicate. At least as of the time this article was written, it hadn’t been tried. Their brains “responded” appropriately to the tasks they were given, while apparently unconscious, but it is not clear that they were actually conscious. We are confident, for example, that the brains of sleeping people (another form of unconsciousness) respond differently to some noises than to others, filtering out familiar and unimportant sounds while often waking us to urgent sounds – a child’s cries, a telephone, the alarm clock. We do know that they seem different from the other 49 (roughly 90 percent) of the subjects, but, then again, if Laureys and Owen had tested all 54 subjects again, on a different day, we do not know whether the results from either the responders or the non-responders would have stayed the same. For people diagnosed as in a minimally conscious state, responsiveness comes and goes quickly and sometimes appears only a few times a week.

We do not know whether these signs of responsiveness are a good prognostic indicator. The subject Owen wrote about in 2006 had some significant recovery after the test. Will these patients also improve? Two of the five had been injured two months earlier or less, but one had been injured 2 ½ years ago and the communicating one had been injured over five years earlier.

We do not know whether the people whose brains responded, or the person who communicated, are feeling any sensations. Their brains responded to speech – do they respond to visual stimulation? Do they feel pain? Do they have emotions?

We do not know whether the people whose brains responded do well with rehabilitation efforts and, if so, what of rehabilitation.

We do not know whether the person who communicated was, in any sense, “competent.” He could say, accurately, whether or not he had brothers. If he said he wanted to have life support withdrawn – or that he didn’t want to have life support withdrawn – how could we know whether he understood the question, his answer, or the consequences? Should we act on that kind of communication? We don’t know.

And, finally, we do not know how this paper will – or should ¬– affect the care of patients in vegetative or minimally conscious states. Will family members (or courts) demand fMRI scans? Should they? And what will, or should, we do differently if those scans provide signs of responsiveness, or of an ability to communicate?

One thing we do know – the issues swirling around these patients have just gotten a lot more interesting, and difficult, as neuroscience has, in yet another context, helped us “read minds.”