The article by de Jong1 addresses the fascinating yet challenging issue of the limits of consciousness in patients with profound neurologic disability and highlights the rapidly changing scientific landscape in this area. When I first began caring for patients with severe brain injury, the persistent vegetative state was a largely unquestioned clinical entity, with respect to both its lack of conscious awareness and its persistence. It was of little interest to the medical community, the subject of very little research, and the subject of ethical discussion primarily within the framework of struggles over the "right to die." In the ensuing decades, nearly everything about this clinical entity has changed. Dramatic research advances led by some of the most sophisticated neuroscientists have altered our ability to visualize the inner workings of the brain, clinical assessment tools and guidelines have focused greater attention on the prevalence and social costs of misdiagnosis, and part of the ethical discussion now concerns the failure to adequately address the needs of this population. Indeed, even the "persistence" of PVS has come under question as a result of periodic cases of late recovery of consciousness and the demonstration of at least temporary reversal of unconsciousness with the drug zolpidem. Yet, one thing has not changed: the fact that an individual's subjective state of conscious awareness can never be directly observed or measured but inferred only from observations of an organized pattern of responses. We simply have devised increasingly sophisticated methods for observing these responses. Moreover, the inference of consciousness, like the testing of a scientific hypothesis, is a fundamentally asymmetric proposition. We can reject the null hypothesis, or fail to reject it, but we can never prove it; we can reject the conclusion of unconsciousness or fail to reject it, but we can never prove unconsciousness.

Although I certainly agree with the authors that the terminology and means of diagnosing the vegetative and minimally conscious states are primitive, given our evolving understanding of the phenomena, I personally favor an interim acceptance of this complexity rather than a rapid redefinition of terms based on our still rather rudimentary understanding. Conscious responding requires the integrity of a complete input-computation-output loop, modulated by arousal and attention. That is, an individual must have adequate alertness and attention, the means of perceiving and decoding some sensory event in the environment (eg, a verbal command, a moving visual stimulus), must decode its meaning or significance, and then must organize some form of response signal to convey to the observer that the prior stages have occurred. A break anywhere in this chain may lead to a failure to demonstrate meaningful responses that could provide evidence of conscious awareness. In the case of typical locked-in syndrome (LIS), we benefit from knowledge of the location of the focal lesion and the relevant neural circuitry, and we place the break at the output side. Locked-in syndrome from a pontine stroke is no different in principle from a high cervical cord lesion; it is just "higher." And the very limited motor responses that most such patients can provide can still demonstrate a full repertoire of cognitive activities, as long as we transform our examination into questions that allow yes/no responses.

In patients with limited responding due to traumatic injuries or other diffuse pathologies, we are faced with a different problem: we do not know precisely where the processing breakdown occurs, or whether it occurs in a single step or multiple interacting steps in this processing chain. That is, in principle, a patient could have a severe limitation in arousal and attention, in multisensory input, in specific cognitive computations, in motor output, or in a combination of all of these. And the more limited the response repertoire, the more difficult it is to sort out these possibilities.

We know from examining diffusely injured patients meeting the current definition of MCS that their behavioral responses are often quite inconsistent. That is, in addition to having only limited motor responses available, they often fail to demonstrate these responses reliably and may fail to respond as our questions become more complex, may vary by time of day, etc. Similarly, when we identify a diffusely injured patient whose only responses are demonstrated through functional imaging or electrophysiologic methods, there is no reason to assume that the only deficit is at the motor output stage. Indeed, there are examples of patients who demonstrate these voluntary patterns of brain modulation but only inconsistently. Thus, like the patients with overt motor responses, we face the complex job of identifying the (potentially) multiple impairments in this input-computation-output loop. In that context, "minimal cognitive responsiveness" fails to specify the nature of the problem. The authors are quite correct that some patients in the minimally conscious state might have preserved consciousness with minimal response capacity. Indeed, there was considerable debate among the members of the Aspen Working Group between the older term "minimally responsive state" and the final recommendation of the term "minimally conscious state." One reason that the latter was selected was to highlight the fact that such patients usually have dysfunction in more than responding.

As research evolves, we may make progress by combining information about the integrity of specific neural networks with careful and repeated cognitive "examination" using these brain modulations as responses. But for the moment, "complete motor LIS" seems both redundant and potentially inaccurate-redundant because LIS has always referred to the absence of motor responses with preservation of sensory and cognitive processing and therefore "complete LIS" is sufficient to define the complete loss of motor responding and potentially inaccurate because many such patients, once we are able to fully characterize them, may turn out to be impaired in multiple ways, only one of which is in organizing their motor response.

Regardless of what term we currently apply, the ethical and resource utilization implications of this phenomenon are great. Given that we can never prove unconsciousness, how hard are we obligated to look for evidence of consciousness? The evidence for consciousness depends on how widely we sample sensory and cognitive domains and how variable the patient's performance-it is sample size dependent. Extensive and repeated examinations with fMRI or similar technologies for all such patients would be enormously costly. And if we can identify patients with some evidence of consciousness, but lack the ability to connect that consciousness to meaningful interaction with the environment, who has been served? I raise these questions not to imply answers but to point out that because of recent advances in research, the clear clinical and ethical distinctions between consciousness and unconsciousness have become shifting sands beneath our feet.

REFERENCE