Terri Schiavo's Brain:
A Neuroethicist Clarifies Her Condition
Recent discussions of the Terri Schiavo case have focused, understandably enough, on its legal and political dimensions. But what propelled this case through umpteen courts and into the US Congress are the neuroethical issues it raises, and we’ve heard almost nothing about them lately. I do not think the lack of attention to the neuroethical issues is simply because they are settled and need not be discussed further. (Indeed, if someone has solved the mind-body problem and I missed it, email me quick!) The issues at stake here concern the relations between conscious awareness, brain and behavior, and these are hard problems. What gives mental life to a brain and how can we tell whether a given damaged brain supports some type of awareness?
Until recently, a patient’s behavior was the one source of evidence we could use to infer their mental state. Now functional neuroimaging has been brought into the mix. Both sources of evidence are trickier to interpret than they appear at first.
Some observers of Terri Schiavo find her behavior indicative of conscious awareness and intentionality. One observer, writing on the website terrisfight.org wrote: “I was pleasantly surprised to observe Terri’s purposeful and varied behaviors... I never imagined Terri would be so active, curious, and purposeful. She watched people intently, obviously was attempting to communicate with each one in various ways and with various facial expressions and sounds.” For me, watching Terri Schiavo in the website videos, it was difficult not feel I was seeing a person interacting with others and aware of her surroundings.
However, clinical and experimental neuroscience have taught us some surprising things about the range of behaviors that can emerge from a decorticate brain. Such behaviors include orienting with eye and head movements toward sights and sounds, generating facial expressions, and producing nonverbal vocalizations that have meaning for us, if not the person producing them, such as cries and laughter. In light of this, we must interpret the behavior seen in the videos cautiously and with a measure of skepticism.
The most natural interpretation for the behaviors we see on the video is not the only interpretation. For example, when a dozing Terri is loudly ordered to “open your eyes!” and does so, does that mean she understood what was said? Or would she have done the same thing if roused with an equally loud order to “open your mouth!” or “stand on your head!”
Humans are hardwired to interpret the behavior of others in terms of mental states. In the psychology literature this tendency is part of a suite of abilities termed “Theory of Mind” (ToM) and in most situations we apply our ToM automatically, without weighing alternative reasons for the behavior. For a particularly striking demonstration of this fact about ourselves, consider the typical response to the robot Kismet. Kismet is part of a research effort at the MIT Artificial Intelligence Lab to design machines that interact socially with humans. Kismet has been programmed to gaze at humans who approach it, orient to salient objects moving within its field of view, pull back avoidantly if an object is thrust forward at it, and so on. People attribute all manner of cognitive and emotional states to this robot on the basis of a fairly small set of simple behaviors, and have been known to become quite attached to it. And this is a contraption made of metal and plastic, not a human being! My point is emphatically not to liken Terry Schiavo to Kismet, but rather to suggest a similarity in our reactions to the woman and the robot.
Recent functional brain imaging results have added a new twist to the debate over Terri Schiavo and the treatment of persistent vegetative state (PVS) more generally. Brain imaging is potentially helpful in understanding the mental life of neurological patients. Behavior is an imperfect measure of cognitive state in anyone, but especially in neurological patients whose verbal and motor systems may be damaged or disconnected from cognitive systems. Brain imaging offers seemingly more direct access to the workings of cognitive areas of the brain. However, at this point in the development of functional brain imaging, the meaning of different patterns of brain activity is not well understood.
A much publicized article in the journal Neurology used fMRI to document cortical responses to speech in patients who were in a minimally conscious state (MCS). This result, reported on the front page of the New York Times, was seized upon by some as evidence that Terri Schiavo may retain more awareness than her behavior suggests. There are two problems with this conclusion.
The first problem is that the patients in this study were in a MCS, not a PVS. What is known about brain activity in PVS? A few studies have shown that the primary sensory cortices of PVS patients respond to touch and sound but higher-level cortices associated with cognition are not reliably activated. However, one PVS patient imaged with PET showed more cortical activation in response to a story told by his mother than to nonsense words.
The second problem is that brain activation, even activation that discriminates between meaningful stimulation and nonsense, does not imply awareness. The cognitive neuroscience literature contains a number of studies in which people’s awareness of stimuli was manipulated while brain activity was measured. In studies with healthy normal participants, awareness of visual stimuli was eliminated by the use of brief, subliminal presentations. In studies with neurological patients, a phenomenon called “extinction” was used to control awareness of the stimuli; visual stimuli on one side of space are extinguished (not consciously perceived) when another stimulus is presented simultaneously on the other side of space. These studies have shown that stimuli can activate relevant regions of cortex even when people are unaware of the stimulus. This is consistent with the idea that the brain activity underlying perception is graded, ranging from nonexistent to the levels of activity observed in normal conscious perception, and that conscious awareness of external events accompanies only the higher end of that range.
The bottom line is that we have two windows through which to look for an answer to the question of conscious awareness in brain-damaged patients, and while neither is crystal clear, both are useful. The first is extended behavioral observation (as opposed to snippets of video), undertaken with an awareness of our susceptibility to the “Kismet” phenomenon. The other is functional neuroimaging, interpreted cautiously and with an awareness of how much remains to be learned about activation-cognition correlations in damaged brains.
