Файл: PLUM AND POSNER S DIAGNOSIS OF STUPOR AND COM-1.pdf

a very low metabolic rate, incompletely preserved MEG patterns of spontaneous and evoked gamma-band responses were seen. Taken together, these imaging data suggest the modular sparing of cortical networks associated with language functions.117 Nevertheless, despite this patient, any verbal output suggests function better than vegetative until proved otherwise.

Another patient, a 26-year-old man, remained in a behaviorally unremarkable VS for 6 years following a motor vehicle accident (Figure 9–8A). MRI T1 images revealed bilateral paramedian thalamic injury, severe bilateral infarction of the tegmental mesencephalon, and diffuse white matter injury. However, nearly normal cerebral cortical metabolism was measured by quantitative FDG-PET in the brain. EEG showed diffuse low-voltage, lowfrequency activity that did not change with arousal patterns correlating with the marked loss of metabolic signal in the paramedian mesencephalon and thalami. Isolated damage to the paramedianthalamusandmesencephalonalone

may cause PVS,112,113 so that in this patient, the preserved cortical metabolism may reflect multiple preserved but isolated networks that fail to integrate because of the overwhelming injury to the paramedian mesencephalon and thalamus.105

Neuroimaging of Isolated Cortical

Responses in Persistent Vegetative

State Patients

In a widely discussed Lancet paper, Menon and colleagues114 described selective cortical activation patterns using a 15O-PET subtraction paradigm in a 26-year-old woman described as being in PVS 4 months following an attack of ADEM. MRI studies of the patient’s brain showed evidence of both diffuse cortical and subcortical (brainstem and thalamic) lesions. Although the patient inconsistently demonstrated visual tracking (leading to some debate as to whether her condition at the

time of study reflected PVS or MCS), no other features of her examination were inconsistent with the diagnosis of VS. Improvements in responsiveness unequivocally consistent with the MCS level were noted by 6 months, with emergence from MCS occurring some time after 8 months. As noted above, it is now generally recognized that prognosis in ADEM includes later recoveries at time periods 6 months or longer after the injury. Thus, patients with ADEM may harbor residual integrative capacities despite a long convalescence. By contrast, similar clinical examination findings in a patient 6 months following cardiac arrest would not portend such a cerebral reserve. The patient eventually made a full cognitive recovery.115 Imaging studies in this patient at 4 months, when described as being in PVS, demonstrated selective activations of right occipitaltemporal regions (in a subtraction paradigm

comparing familiar faces and scrambled images). The investigators interpreted activation of the right fusiform gyrus and extrastriate visual association areas as indicating a recovery of minimal awareness without behavioral manifestation. The findings in this patient, however, point out a significant limitation of

brain imaging techniques in this clinical context and have been extensively debated.111,116

The selective identification of relatively complex information processing associated with visual processing of faces as shown here may not alone provide an index of recovery of cognitive function or even potential for recovery. Specific cortical responses to faces are obtainable in anesthetized animals118 and, if found in isolation of any other imaging evidence or bedside demonstration of awareness, do not guarantee that these patterns of activation represent cognitive function per se. Without

further clinical evidence, the present state of imaging technologies cannot provide alternative markers of awareness. While neuroimaging studies hold the promise of elucidating underlying differences between VS/PVS and MCS patients, at present no techniques are able to identify awareness in such patients unambiguously.

Owen and colleagues119 have subsequently developed a new imaging framework to evaluate volitional responses in VS and MCS patients that address the ambiguities of the methods used in the Menon study.114 Applying these new methods,120 they identified unequivocal neuroimaging evidence of a patient remaining in VS at 5 months following a severe traumatic brain injury being able to follow commands to imagine various visual scenes. The

commands were associated with activation of appropriate areas of the cerebral cortex, despite lack of an external motor response. At the time of examination the patient showed evidence of brief visual fixation, a possible transitional sign for evolution into MCS.76 Another examination 11 months later revealed visual tracking to a mirror, another transitional sign, but no evidence of object manipulation or behavioral manifestations of command following. Figure 9–9 shows the main result of the study. The imaging findings demonstrated preservation of cognitive function for this particular patient that the clinical bedside examination failed to reveal, and indicated a cognitive level at least consistent with MCS. It is important to recognize that command following is a cardinal feature of MCS that does imply

communication. MCS patients may show consistent evidence of command following with visible motor responses that cannot be used to establish communication. Neuroimaging studies in such MCS patients also show preservation of large-scale cerebral networks (see below). As a result, it is unclear from the methods used in the Owen study whether or not the patient’s level of consciousness was consistent with MCS or a higher level of recovery.

FUNCTIONAL IMAGING OF MINIMALLY CONSCIOUS STATES

Only a few studies have examined brain activity in MCS.121 In five MCS patients,15 O-PET

identified activation of auditory association regions in the superior temporal gyrus not seen in PVS patients.122 In addition, stronger correlation of the auditory cortical responses with frontal cortical regions was observed in both MCS patients and control subjects than in PVS patients. Median nerve electrical stimulation activated the entire pain network, similar to the response in normal subjects123 (see Figure 9– 6). These findings stress the need for analgesic medications when MCS patients undergo painful procedures.

Multimodal imaging studies using functional MRI activation paradigms and FDG-PET in two MCS patients near emergence from MCS uncovered unexpected evidence of widely

preserved large-scale cerebral network responses121 (Figure 9–10). Both patients suffered sudden brain injuries (blunt trauma, spontaneous intracerebral hemorrhage) leaving them in MCS for longer than 18 months. The acute phase of injury of each patient included herniation to a midbrain level. This historical feature is commonly associated with MCS and other poor outcomes following traumatic brain injuries as a result of focal infarction or indirect (axonal shearing, ischemic) damage.124 Neither patient demonstrated consistent command following or functional communication (either gestural or verbal) on repeated examinations. Both patients, however, did demonstrate best responses that included command following or occasional verbal output (single words).

Figure 9–10. Functional magnetic resonance imaging (MRI) maps obtained during listening to spoken narratives, in a minimally conscious state patient (left) and control subject (right) measured by functional MRI. Yellow color indicates response to spoken narratives, blue color indicates response to time-reversed narratives, and red color indicates regions of overlapping response to both conditions. See text for details of paradigm. (Adapted from Schiff et al.,121 with permission.)

Significant fluctuations in their responsiveness occurred across examinations. Figure 9–10 shows cortical activity for one patient and one normal control associated with receptive language comprehension during presentation of 40-second narratives prerecorded by a familiar relative, presented as normal speech, and also played as time reversed (backward). Brain activations in response to normal speech are shown in yellow. Selective responses to backward presentations are shown in blue. Normal speech generated robust activity in languagerelated areas of the superior and middle temporal gyri for both the control subject and the MCS patient. In addition, the normal speech stimuli produced brain activations in the MCS patient’s brain in the inferior and middle frontal

gyri, primary and secondary visual areas including the calcarine sulcus, and precuneus. The pattern of brain activations for normal speech in this patient overlapped with that in the normal controls. However, different from the normals, neither patient activated in response to reversed speech. The findings also indicate that the functional MRI technique alone is insufficient to characterize the presumably wide differences in brain function that separate the patients and the control subjects. In addition, unlike PVS patients who fail to produce activation of polymodal association cortices in response to natural stimuli, the two MCS patients retain potentially recruitable cerebral networks that underlie language comprehension and expression despite their inability to

execute motor commands or communicate reliably. The preservation of large-scale forebrain networks associated with higher cognitive functions such as language provides a clinical foundation for wide fluctuations sometimes observed in MCS patients. Other inves-

tigators have obtained similar neuroimaging findings from single MCS patients.99,125

The same limitations of imaging techniques for determining awareness in VS/PVS patients also limit assessment in MCS patients. One cannot determine whether or not the functional MRI activations indicate awareness without communication, and by definition these patients cannot communicate. In addition, when they do awaken, they typically are amnestic for this period of time. Neuroimaging studies of visual

awareness in patients and normal subjects implicate certain patterns of coactivation across cortical networks as the principal correlates of awareness, including coactivation of prefrontal and parietal cortices along with the occipitaltemporal cortex.126 Although the activation patterns identified in the MCS patient shown in Figure 9–10 include several of these specific regions, the patient is unable to communicate reliably to indicate whether visual or selfreflective awareness is present. The coactivation of prefrontal, parietal, and occipital regions suggests awareness but is potentially consistentwithotherinterpretations.Similarconcerns arise in the interpretation of the Owen120 findings shown in Figure 9–9.

In the future, functional brain imaging techniques in combination with electrodiagnostics may identify patients with rehabilitative potential, and conversely, those in whom further recovery is not expected. The introduction of the MCS nosologic category is aimed at directing efforts to identify patients who may have some substrate for further recovery despite very limited behavioral evidence of awareness. On the other hand, fragmentary cortical networks may remain in VS patients without heralding further recovery or signifying awareness. The ‘‘gray zone’’ between VS and the lower functional boundary of MCS in Figure 9–1 reflects a probable overlap region where patients may acquire a reliable sensory-motor loop response of very limited cerebral systems that, despite contingency with environmental stimuli, may not reflect awareness or potential for further recovery. It is critical, then, to identify residual capacity as opposed to isolated functional activity in the cortex. This will require prospective studies of large numbers of patients with early VS, to determine if there are indices on functional imaging that can predict eventual improvement.

POTENTIAL MECHANISMS UNDERLYING RESIDUAL FUNCTIONAL CAPACITY IN SEVERELY DISABLED PATIENTS

The neuroimaging studies reviewed above raise the question of what mechanisms might limit further recovery in MCS patients who harbor widely connected and responsive cere-

bral networks. Fluctuations of cognitive function in MCS patients91,127(and occasional late

spontaneous emergence from MCS [see below]) demonstrate an underlying residual cognitive capacity in some severely injured brains. At present, no studies have addressed this question by systematically correlating brain structural integrity, cerebral metabolism, and electrophysiology across a large sample of patients with severe disability. Nonetheless, several careful observations of variations in structural injury patterns, patterns of normal resting metabolic activity, and abnormal brain dynamics provide potentially important clues and directions for future research.

Variations of Structural Substrates

Underlying Severe Disability

Clinical observations and quantitative measurements of neuronal loss following complex brain injuries do not support an invariably straightforward relationship of recovery of cognitive function that is simply graded by the degree of vascular, diffuse axonal, and direct ischemic brain damage. Although indirectly measured volumetric indices do offer some prediction of long-term outcome in PVS following overwhelming traumatic71 or anoxic brain injury,38 pathologic studies comparing patients remaining severely disabled following brain injuries to those remaining in VS demonstrate that severely disabled and some MCS patients may have only focal brain damage, whereas PVS patients suffer diffuse axonal injury.128 Severely disabled patients with diffuse axonal injury appeared to have lesser quantitative damage than PVS patients. These findings suggest that significant variations in underlying mechanisms of cognitive disabilities and residual brain function accompany MCS and other severe but less disabling brain injuries.

It is also well known that enduring global disorders of consciousness can arise in the setting of only focal injuries.129 These injuries are typically concentrated in the rostral teg-

mental mesencephalon and paramedian thalamus.112,130 Patients with moderate, diffuse

axonal injury combined with limited focal damage to these paramedian structures have not been systematically studied, but this pathology

probably plays an important role in causing severe disability.128,131 The paramedian thalamic and upper brainstem structures are specifically vulnerable to injury during periods of acute cerebral edema produced by traumatic brain injuries, infarctions, hemorrhages, infections, and brain tumors, as reviewed in Chapters 3 and 4.

Recent studies suggest that slowly developing structural remodeling may be a potential source of late recovery following severe brain injury. Voss and associates132 longitudinally characterized brain structural connectivity and resting metabolism in a 40-year-old man who recovered expressive and receptive language after remaining in MCS for 19 years after a traumatic brain injury. The patient continued to improve over the next 2 years. MRI revealed extensive cerebral and subcortical atrophy particularly affecting the brainstem and frontal lobes; there was marked volume loss throughout the brain with ventricular dilation (Figure 9–11A). Diffusion tensor imaging (DTI) data revealed severe diffuse axonal injury, as indicated by volume loss in the medial corpus callosum (Figure 9–11B, C). In contrast to the overall severe reduction of brain connectivity demonstrated by DTI fractional anisotropy maps, measurements also revealed large regions of increased connectivity in posterior brain white matter not seen in 20 normal subjects (Figure 9–11B). These large, bilateral regions of posterior white matter anisotropy were reduced in directionality when measured in a second DTI study 18 months later (Figure 9–11C). At this time, repeat imaging identified significant increases in anisotropy within the midline cerebellar white matter that correlated with significant clinical improvements in motor control over the intervening time period.132 Figure 9–11D shows the quantitative changes in an index of fractional anisotropy and left-right fiber directions; the open circle shows measurements at the time of the first scan (Figure 9–11B), and the open square shows the marked increase in fractional anisotropy corresponding to the increased intensity of the red region within the midline cerebellum (Figure 9–11C). These findings suggest the possibility of structural changes within the patients’ white matter playing a role in their functional recovery. Recent experimental studies provide some support for such a mechanism of late remodeling of

white matter connections after structural injuries133,134 and in normal human adults.135

Although suggestive and fascinating, individual case studies of this sort must be interpreted cautiously. Nevertheless, such findings indicate the need for larger prospective studies examining whether slow structural changes do arise in the setting of severe traumatic brain injuries and, if present, whether they influence functional outcomes.

The Potential Role of the Metabolic ‘‘Baseline’’ in Recovery of Cognitive Function

As discussed on page 370 and illustrated in the example shown in Figure 9–10, the abnormal response to speech reversal in some MCS patients provides a potentially important clue to the mechanisms underlying their profound cognitive impairment. Control subjects were instructed to listen passively to the sounds; however, the time-reversed narratives elicited an involuntary attempt to decode the speech. The failure of the time-reversed narratives to activate the large-scale languageresponsive networks identified by the forward presentations in MCS patients suggests a significant difference in the resting state of the brain in MCS patients and control subjects. The recruitment of a normal network activation pattern suggests that MCS patients may require very salient stimuli to activate these language systems (e.g., clear human speech, familiar voice, emotional content, etc.).

Raichle and colleagues have proposed that the normal human brain has a ‘‘baseline’’ state of metabolic activation (as reflected by oxygen uptake) reflecting ‘‘default self- monitoring.’’136–138 Specific areas of brain that are active at rest (e.g., posterior cingulate cortex and ventral anterior cingulate cortex139) form a network that deactivates during tasks that activate other areas of the brain. Data obtained from these investigators provide some evidence supporting a functional role of a resting state of monitoring environmental factors and an internal state that might be

sensitive to salient events such as emotionally meaningful human speech.138–141 Loss of sig-

nal in these regions is a common finding in VS and partial recovery of this metabolic signal is seen in MCS.141,99