by Bor, Daniel
Over the next seven years, there were numerous legal petitions, appeals, motions, and so on, with Michael, on the one hand, trying to let Terri die, and Terri’s parents, on the other, determined to stop him. The case violently pitted civil liberties against religious values. Terri was a devout Catholic, and questions were raised as to whether she would have wanted her life to be ended in a potentially sacrilegious way. The pope himself ruled on the case at one point, stating that Terri should be kept alive. At times the story looked like a sorry soap opera, with claims being bandied about that money was a driving motivation on both sides of the family divide—Terri’s estate was worth a considerable amount by this stage, because of successful litigation against the fertility doctor who had failed to spot her bulimia. At the same time, since 1993 Michael had been dating another woman, with whom he’d had two children, one in 2002 and the second in 2004, but he refused to get a divorce from Terri. His stated reason was a continued emotional attachment to Terri, but her parents claimed this was simply a strategy to lay claim to her money.
As the years rolled by, the feeding tube was removed numerous times, only to be replaced on Terri’s parents’ appeal. The public became increasingly fascinated by the case, and the politicians began to step in, realizing that cheap votes from the far right could be won by taking a firm pro-life stance. Jeb Bush, the governor of Florida, passed legislation giving him the right to intervene in the case and prevent Terri’s death, although this was later found to be unconstitutional. Eventually even the president, Jeb’s brother, George W. Bush, weighed in on the pro-life side, canceling a vacation and flying back to Washington, finally signing emergency legislation late in the night, in his pajamas. The main purpose of this bill was to keep one person alive: Terri Schiavo.
In the end, the political and legal wrangling fell irreversibly on Michael’s side. Terri Schiavo died on March 31, 2005, thirteen days after her feeding tube was removed for the last time.
At the heart of this long, painfully public battle were two heart-rending questions. First, was Terri Schiavo still there, even in some shrunken form, as the personality by which she had been defined prior to her collapse? And second, if the brain damage had indeed stripped her of any chance of normal awareness, might there have been a treatment on the horizon that could have restored it? Terri Schiavo is by no means a unique case—there are in all likelihood well over 100,000 similar examples worldwide, with the families of each of these patients obsessively asking themselves these two exact questions.
A THIN VEIL BETWEEN LIFE AND DEATH
The first stage in assessing how little consciousness is left in patients like Terri is by arriving at a formal diagnosis. If a patient is entirely unconscious with eyes persistently closed, then he is in a coma. If he has some form of sleep-wake cycle, and sometimes opens his eyes, but shows no signs of awareness, then he is in a vegetative state. Many vegetative state patients can make at least a partial recovery and even regain normal consciousness, but about half do not, and the longer the patient is in a vegetative state, the less likely he is to recover. In line with research on how normal consciousness is supported in the brain, the vegetative state is closely linked with damage to the thalamus, along with its connective pathways to the prefrontal cortex, especially when there are no signs of long-term recovery. But this syndrome is also linked with many other damaged brain areas, along with a range of initial causes, including head injuries, infection, drug overdose, and so on. If the vegetative state continues for a month, then it is classed as a persistent vegetative state. If it continues for a year (depending on the injuries and the country of diagnosis), then the patient is classed as being in a permanent vegetative state, or PVS. At this stage, chances of recovery are increasingly slim.
But if the patient’s condition does improve and he shows some slight signs of awareness, such as tracking an object with his eyes, or responding to commands to move his arm, then he is upgraded to being in a minimally conscious state. It’s relatively straightforward to ascertain whether a patient has moved from coma to vegetative state—you just have to see whether his eyes open and close—but it’s notoriously difficult to tell the difference between a vegetative state and a minimally conscious state. Although the vegetative state might appear to be an improvement over a coma, the truth is that consciousness may still be entirely absent. Therefore, to provide reassurance to the patient’s family that he is consciously present and might recover, an improved diagnosis of minimally conscious state is crucial.
For the patient’s family, it’s easy to mistake certain outward signs as hopeful evidence of consciousness—the patient might weep, smile, or grind his teeth—but in fact these are effectively reflexes and can occur with no consciousness whatsoever. Doctors, too, although mindful of overinterpreting these surprisingly primitive behaviors, can often struggle to disentangle the subtle, inconsistent signs of real awareness from random sounds and movements. Clinicians distinguish between PVS and minimally conscious state by looking for any behavioral hints that the patient is conscious. Can he speak at all, for example? Does he seem to follow your eyes? Does he make purposeful movements? Their search for such evidence is within the usual challenging context of a patient who produces copious twitches or spastic limb movements and who might be drifting in and out of sleep on a regular basis. This somewhat haphazard approach means that it is extremely difficult to diagnose such patients accurately—with some estimates suggesting that 43 percent are misdiagnosed by the standard clinical bedside observations, usually because patients are assumed to be in a vegetative state when they should be upgraded to minimally conscious. Better behavioral assessments are slowly filtering through in the form of carefully designed rating scales, which standardize and quantify the diagnostic markers that place a patient in one group or another. But even this improved measure can be completely misleading in some cases.
VIEWING CONSCIOUSNESS FROM WITHIN
When we watch a friend open her mouth wide, raise her eyebrows as high as they can go, and point an urgent finger at a naked man wearing only a top hat who is going down the middle of a busy street on a unicycle, we have a good idea as to the contents of our friend’s mind, and have no trouble knowing that she is awake and normally conscious. But it’s easy to forget that this is only ever an indirect measure: Her hand movements or the expression on her face are merely the outputs of her brain’s activity, and not her brain activity or consciousness itself. If the connection between brain and body becomes severed as a result of brain damage, blocking a person’s ability to move, then it’s no longer valid to infer anything about what his body language shows about his conscious thoughts. It certainly is invalid to fully assume that a person has no conscious life because he is completely paralyzed. Of course, we rarely have to worry about this problem, but with patients in vegetative or related states, it is a real concern, and one that no behavioral rating scales, however standardized or carefully constructed, will ever solve.
Adrian Owen, my primary supervisor during my PhD work and then for many years my boss as I became a research scientist, has spent much of his career searching for better, more direct ways to gauge the level of inner life in such patients, with my occasional assistance. A charismatic ginger-haired man, full of humor and uncannily resembling Vincent Van Gogh, Owen has used brain-scanning as a powerful tool to probe the level of consciousness in vegetative patients. These emerging techniques have proven, for some patients, to be far more sensitive than any based on behavioral observation.
One recent large-scale study he organized involved 41 patients who were all in either a vegetative or a minimally conscious state. Owen and his team placed patients in the fMRI scanner and presented them with various kinds of sounds via earphones. The sounds were distinguished by their complexity. First there was random noise for the patients’ brains to process. Then came normal speech, via standard sentences—for instance, “Her secrets were written in her diary.” Finally, and most tricky of all to process, were sentences co
ntaining ambiguous words—for instance, “The shell was fired towards the tank.”
By seeing which brain regions activated for each condition, and comparing this with the way that the brains of normal, healthy subjects light up on the same tasks, he would thus have four possible levels of cognitive processing distinguishable by brain-activity patterns: inability to process any sounds, ability to process random sounds, ability to process linguistic sounds, and ability to process the meaning of words. Surprisingly, two patients who had been firmly classified on behavioral scores as being vegetative were able to pass the highest level and process the meaning of the words they heard, as shown by their brain activity. So they were probably minimally conscious instead—or better. More exciting still was Owen’s discovery that the extent to which these patients passed the sound tests was a strong predictor of how much they would recover six months down the line.
This research is groundbreaking and may in future years become an important tool in the clinician’s standard arsenal of assessing the level of consciousness in such patients. It should also give hope in certain situations, backed by good evidence, that a future improvement in symptoms is likely. Recall from Chapter 3 that understanding the meaning of words—a skill unavailable to those under general anesthesia—is one of the simplest functions requiring consciousness. Therefore, showing brain activity for meaning in this experiment is a solid clue that consciousness is present. It is also a sensible reason for why patients who pass this part of the test go on to recover further degrees of awareness.
But such data do not definitively show that these patients are conscious because it is too indirect and circumstantial. It is technically possible that consciousness is not present in these patients, even if meaning is appropriately processed. To answer such a vital question definitively, more convincing evidence would be needed.
This is exactly what Owen and colleagues provided in a landmark study using a somewhat different approach. Owen reasoned that if these patients could demonstrate volition, if they could choose to follow a complex command that he gave them, then that would provide unbreakable evidence that they were indeed conscious, regardless of what the clinicians’ diagnosis was. He was able to test this idea in 2006 with a twenty-three-year-old woman who entered a vegetative state following a road traffic accident. She had already passed the previous sounds test with flying colors, showing robust activation to ambiguous words. But then Owen asked her while she was again in the scanner to carry out one of two tasks. Sometimes he would ask her to imagine playing tennis, and at other times he asked her to imagine navigating around all the rooms of her house (recall Chapter 1, when my friend Martin Monti carried out a similar experiment on me). When Owen looked at her brain activity for the two commands, her brain responses were indistinguishable from those of normal, healthy controls performing the same imaginary tasks. The only possible conclusion he could draw from this finding was that the doctors had misdiagnosed her, that she wasn’t in a vegetative state, after all, but was in fact clearly conscious. Following this scan, she continued steadily to recover, and within a year she was able to answer yes-or-no questions by pressing a button with her foot.
Such a discrepancy between the standard clinical diagnosis and the brain-scanning results is dramatic, if relatively rare. In a later study, Owen, Martin Monti, and colleagues showed that only about 17 percent of apparent PVS patients could successfully show appropriate brain activity in response to the imagination commands. Still, this experimental approach shows that brain scanning can at times be a far more sensitive way of gauging levels of consciousness in these severely brain damaged patients than any bedside assessment by the doctor. Owen also believes that this result shows that some seeming PVS patients may already have leapfrogged over the intermediate minimally conscious state and have an almost normal level of consciousness—but they are simply locked inside their heads with no way of demonstrating awareness.
Because an fMRI scanner is a very expensive, nonportable machine, Owen’s group has recently been testing the use of EEG instead. EEG is a more practical clinical tool, since it is far cheaper than MRI and can be taken to the bedside—an important factor for such patients, who are difficult to move. Although picking up a reliable signal with this technique poses a challenge, Owen and colleagues have nevertheless shown by a similar command-following protocol that it can be used to demonstrate that some apparently vegetative patients are indeed conscious.
COMMUNICATION BY BRAIN ACTIVITY
Such methods open up the possibility that, even though some of these patients give all outward appearances of being robbed of consciousness, not only are they clearly aware, but we might even be able to communicate with them via the brain scanner. When Martin Monti piloted his fMRI experiment on me, I answered yes-or-no questions by imagining moving around my house or playing tennis. He would then read off my answer in the brain activity the scanner picked up. Exactly this technique was later used on one patient, a young Belgian man, who had, it was assumed from outward appearances, been in a vegetative state for five years following a traffic accident. When he was asked about the name of his father and other personal questions, he answered robustly and correctly via his selective brain activity to the first five questions. This was a striking result, completely at odds with his PVS diagnosis.
It’s important to emphasize in all these studies that, as usual, a negative result on scans is hard to interpret—for instance, if a patient fails to generate the appropriate brain activity, she might just have moved too much in the scanner, which in itself spoils results. However, passing this test clearly reflects consciousness, and as methods are refined, it should become easier both for patients to prove they are aware and for them to communicate.
Although none of these methods are close to being a conventional clinical tool, there is every reason to assume that the situation will change in the next five years or so. Having a standard, brain-based method by which patients could express their thoughts and wishes, despite being completely unable to move a muscle, would give them a much-needed voice and provide hope and reassurance to their anxious families.
CHECKING THE INTEGRITY OF CONSCIOUS NEURAL HIGHWAYS
The methods described above essentially seek a lucky glimpse of conscious communication in the peaks and valleys of flittering brain activity. But what if the patient is having a bad day when scanned—for instance, spending much of this time asleep, when in fact he is usually very conscious? Or what if the patient really is only partially conscious, but all the neural architecture is present for a solid recovery? There is certainly mileage in applying less direct neural investigations that bypass questions of online consciousness or communication. Indirect methods with the most potential for these patients include those that examine whether the brain is sufficiently physically wired to support consciousness, or whether regions are still communicating with each other, in a way that reflects awareness. These methods, although less sexy, could nevertheless prove more robust than other techniques that rely on the patient obeying instructions, and they might also offer more reliable predictions of recovery.
And, given the level of maturity of the science of consciousness today, it is certainly possible to exploit existing knowledge about consciousness and the brain in order to make accurate diagnostic tests of PVS.
First, it is well known that the thalamus, that central relay station in the brain, has a critical role in consciousness. Although it might not be as important as the prefrontal parietal network in supporting the richness of our various experiences, its normal functioning, as a means of transmitting and combining information freely throughout the cortex, is necessary for awareness to occur. Significant damage to the thalamus, as seen in anatomical scans, is an obvious indicator that consciousness will be lacking and recovery unlikely. But even if the thalamus seems intact, that doesn’t mean that its connections with the rest of the brain are still functional. So Davinia Fernandez-Espejo, Adrian Owen, and colleagues recently used a relatively n
ovel MRI scanning technique, diffusion tensor imaging, in order to examine how well the thalamus was still connected with other regions in such patients. It turned out that the integrity of these fiber pathways was an excellent diagnostic measure of whether patients were classed as PVS or minimally conscious, securing a 95 percent accuracy. This is a considerably higher success rate than is typically reached by means of standard clinical behavioral assessments, which, you may remember, have a lamentable 43 percent misdiagnosis rate.
Looking directly at the pathways of the prefrontal parietal network is another clear research goal, emerging out of the large bank of evidence arguing that this is the most central set of brain regions for consciousness. Using a complex neuroimaging technique that assesses how one brain region causes activity in another, Melanie Boly and colleagues were able to show that, while in PVS patients the prefrontal cortex was no longer influencing regions toward the back of the brain, in minimally conscious patients and normal subjects there was a more robust, active link between the regions, with information flowing both away from and toward the prefrontal cortex.
Another principled approach, this time springing from information integration theory, uses the TMS-EEG technique mentioned in Chapter 6, where an initial burst of cortical stimulation, initiated by the TMS machine, is then monitored by EEG for its spread and duration. A prolonged bouncing of activity throughout the cortex is found in awake, aware individuals, while only a local, short-lived response is present when we’re in a dreamless sleep. Although this research is in the early stages, the approach is also being used to measure the level of residual awareness in vegetative patients.
