There are many reasons I have been hesitant to talk about consciousness, the biggest being that I have no idea what it is. Many colleagues also have difficulty in clarifying it, and the subject of this month's column does nothing to allay our hesitancies. The content involves the extraordinary recent news about a man who had been in a minimally conscious state for 6 years. Researchers were able to arouse him and then send him back into minimal consciousness with the flip of a switch. Exactly what was being stimulated and what was missing before that stimulation is an open question.
To discuss the finding and its intriguing (some might say disturbing) possibilities, I will examine arousal and attention, review some brain anatomy, and then progress directly to the research. Feel free to skip to "Data" if intralaminar thalamic nuclei and calbindin-positive cells are working parts of your vocabulary.Problems defining consciousness
Defining consciousness means discussing arousal and attention, which are very difficult subjects to investigate at the neurological level. Millions of sensory neurons fire simultaneously throughout waking life, all carrying messages to various brain regions that are clamoring for attention. Only very few inputs will succeed in breaking through to awareness--the rest will be partially or fully ignored. It is easy to alter this balance and effortlessly grant airplay to one of the many messages you were previously ignoring. (While reading this sentence, can you feel where your elbows are?) This fact led some researchers to conclude that attention involves a simultaneous 2-step filtering process: the ability to focus awareness on certain stimuli and the ability to filter others out. Both steps involve the concept of consciousness, which is sometimes defined as the part of the mind where awareness resides.
But exactly what is awareness? Beginning in 1994, the Tucson Conscious- ness Conference has convened every year. It is a gathering of august scientific professionals whose goal it is to specifically define and characterize consciousness at a neurological level.
At this meeting every year, the failure of researchers to find the region or regions in the brain that are responsible for consciousness is notable; in some years they cannot even agree on what they are discussing. Until that benchmark is achieved, it will not be possible to identify the overarching brain region where consciousness resides. (It is the admittedly grumpy view of some--a cadre that currently includes yours truly--that no such region exists.)
Even if an agreement can be made on exactly what consciousness is, there is still the uncomfortable observation that there may be many subtypes. Sometimes called "levels of awareness," this multiplicity was brought home in an extraordinary fashion almost 40 years ago by studying patient memory formation during surgery. Physicians were taught that a patient under anesthesia was fully unconscious. Naturally, surgeons felt free to discuss anything they chose, including some uncomfortably personal things about the patient on whom they were operating.
In the 1950s, a series of clever experiments were performed while patients underwent surgery, including a cruel one in which the surgeons jokingly made false statements that the surgery was failing and that the patient was going to die. These experiments demonstrated that patients could be surprisingly "aware" even when "unconscious." It was shown that patients could hear, remember, and later repeat conversations that occurred while they were under anesthesia during surgery. The memories of these patients were so powerful that they could affect the length and quality of their postoperative recoveries (patients exposed to the doctors' depressing false statements did not recover as well or as fast). How do we reconcile these data with the notion of conscious awareness? We do not. Clearly, there are levels of awareness, but we have a long way to go before we fully understand what that awareness really means.
It is much easier for both researchers and clinicians to describe consciousness as the absence of something rather than the presence of something. There are a number of ways to characterize catastrophic cognitive losses, 3 of which are described below. The diagnostic borders defining these states are fuzzy, and there have been calls for more rigorous research to better classify what is sometimes characterized as a continuum of consciousness.
Coma signifies catastrophic loss. No spontaneously organized behavior occurs in this state and there is no sensory awareness or response to aversive stimuli. There is a distinct lack of physical interaction with external cues beyond brainstem-mediated reflexes, and the patient appears to be permanently and irretrievably "sleeping."
Persistent vegetative state (PVS) is identical to coma with one crucial exception: the patient does not appear to be sleeping. PVS is usually described in terms of the eyes' movements; the eyes may open and close, with evidence of roving nondirected motion. Brainstem activity still occurs primarily, although there are suggestions that discrete but disconnected cortical regions remain active.
In a minimally conscious state, patients show some signs of arousal and occasionally organized behavioral responses to external stimuli. Imaging studies have shown that parts of the cortex are functional, and there can be intermittent self-awareness and environmental awareness in the patient.Neurological substrates and deep brain stimulation
The lack of a consistent definition of consciousness has not deterred some courageous scientists from trying to describe its neuro- biological substrates, and there have been some quite spectacular successes. A great deal of research has pointed to a pivotal role for the thalamus. The thalamus is a multitalented egg-shaped clearinghouse involved in sensory perception and is organized as clusters of nuclei. The best characterized of these nuclei serve as relay stations for signals coming from the skin, ears, and eyes, and rout them to appropriate areas in the sensory cortex.
In non-human primates, a group of neurons exist that have a different job description. They appear to be involved in a more generalized activation of cortical networks in the brain. Their axons terminate more broadly in cortical areas and among the layers of tissue that the relay-oriented cells do not target. They produce a unique protein called calbindin and are found throughout the thalamus. These cells form the bulk of cells in the intralaminar nuclei of the thalamus--an area long suspected of involvement in arousal and consciousness.
I discuss this anatomy because of the manipulation used in the experiments described next. The manipulation, which has received a great deal of interest by the press, is called deep brain stimulation (DBS). The technique involves the surgical implantation of electrodes that can stimulate specific tissues in the brain. The technology is quite old; the first recorded application was performed in 1870 with dogs and showed that the direct electrical stimulation of the motor cortex could elicit limb movement. Today, implantation of DBS pacemakers in specific areas of the brain has proved extremely useful in a variety of therapies, such as alleviat- ing the symptoms of chronic pain and the tremors associated with Parkinson disease.
It was the fact that specific regions in the thalamus were involved in awareness that led researchers to examine a potential role for DBS techniques in minimally conscious patients. The question was simple: If electrical stimulation were to be applied directly to the intralaminar nuclei of the thalamus, would patients wake up? Or would they at least show changes in behavior normally associated with a more sustained state of awareness? A group of researchers at Cornell in New York and at the Cleveland Clinic in Ohio decided to research this, and the results of their efforts are where I turn next.Data
Researchers treated a 38-year-old man who had suffered what they thought was irreversible damage to large regions of his cerebral cortex and had been in a minimally conscious state for almost 6 years. Although he was severely impaired, functional MRI showed the preservation of some of his cerebral language network. There was evidence of intermittent interactive behavior; thus, he was an ideal candidate for this study.
Electrodes were inserted bilaterally into the anterior intralaminar thalamic nuclei and into the adjacent paralaminar regions of the thalamic association nuclei. They were situated to cover most of the calbindin-positive regions (Figure). The experiment commenced with the first application of current to those electrodes.
Two types of behavioral tests were deployed during the investigations and fell into both primary and secondary outcome assessments. Primary outcomes involved the measurement of cognitive-mediated behaviors (levels of awareness, response to voices, and so on); secondary tests involved an assessment of motor behaviors (such as limb control and oral feeding). Scores in all these inventories were obtained over 3 weeks to establish a baseline before surgery. The researchers then applied the electricity.
Initial results were, to put it mildly, spectacular. Within 48 hours of the first stimulations the patient began to dem- onstrate sustained eye opening, an overall increase in attentional arousal, and even a bilateral ability to turn his head in response to human voices. When tested formally, both primary and secondary postsurgical scores on the various inventories were quite different from the presurgical scores, and in the positive direction. It appeared as if the electrical stimulation of specific thalamic nuclei had created a behavioral effect associated with increased awareness and reactivity.
To unequivocally establish that the observed behavioral results originated from the electrical stimulation and not from some artifact of the surgery, sever- al tests were run over about a year. These manipulations could be divided into 3 periods.
Confirmation period. After the initial success, the patient was allowed to "rest" for 50 days before the electricity was reapplied to the electrodes. During this rest period, a decrease in the various arousal and motor scores was observed, and the patient returned to his minimally conscious state. This was followed by a subsequent and dramatic increase in those outcome scores when current was reapplied.
Titration period. An 18-week assessment followed this confirmation period. In these experiments, various parameters of the stimulation were manipulated in an attempt to optimize the observed behavioral changes. A range of stimulation frequencies and intensities were examined. During this titration phase, remarkable behavioral changes continued to be observed. The patient became verbally competent (he had not spoken for more than 6 years), and showed a consistent ability to name objects given to him. He also began manipulating physical objects with his hands. The patient even regained the ability to swallow food and could receive food by mouth. The gastric feeding tube was disconnected.
Double-blind crossover period. A final 6-month, double-blind, alternating crossover study was performed. In these manipulations, the researchers turned the electricity off and on in a systematic fashion, assessing the manipulation with the blinded rigor of a clinical trial. Correlative relationships were easy to establish; the highest state of arousal and the ability to perform the motor tests were always associated with optimal periods of stimulation. At this stage of the experiment, many functions remained in an improved state, even when the electricity was withdrawn. It became apparent to the researchers that a carry-forward effect was being observed, as if the patient was responding to some type of treatment. Behavioral improvement continued to be observed without stimulation.Conclusion
As you might suspect, this finding has caused quite a stir in the research community. And it raises hope for patients with brain injuries around the world.
It also raises a few questions, including the necessary calls for more research, and even a few caveats. As rigorous as the study was, the researchers had an "N" of exactly one person. That means it is a clinical case study only and not a statistically salient (and therefore portable) result. Second, the patient was selected partially on the basis that some of his neurological substrate was preserved. As mentioned, the patient had displayed spotty but quite clear evidence of interactive behavior before surgery. These data, exciting as they are, do not currently apply to anyone in an unconscious state. They do not even apply to all minimally conscious patients.
These concerns hardly cast a shadow on the findings, however. From a research perspective, the data raise important questions--as tantalizing as they are frustrating--about the neurobiological basis of conscious behavior. There are regions in the thalamus that are clearly involved in arousal. Yet, we know from countless other patients that the brain harbors many other networks that also contribute to conscious awareness. How does stimulation of these calbindin-positive intralaminar cells interact with these other networks? How much of these other networks must go offline before stimulating the thalamus becomes useless or becomes the treatment of choice? And what, if anything, do these experiments contribute to establishing the overarching definition of human consciousness?
At this point, they do not contribute much to this elusive goal. Stimulating one part of the engine of a complex car does not completely inform us about a vehicle's overall shape, which is why I have chosen to ignore the subject all these years. It is thrilling to watch the mind of a seemingly irreversibly injured patient suddenly spring back to life. At the same time, it is frustrating to observe how much further we have to go. Maybe there will be more to say in another decade or so, and if I am still around, I will be happy to write about it.