Our understanding of the pathophysiology of chronic disorders of consciousness continues to be illuminated by creative functional neuroimaging studies, yet the diagnosis and classification of these disorders remain based on clinical examination findings.1Naci andOwen2 supplement thegrowingevidence that functionalmagnetic resonance imaging (fMRI) can uncover cognitive functioning that cannot be elicited by neurological examination. They show that selective auditory attention and the capacity to follow commands and communicate canbedetected ina fewpatients in aminimally conscious state (MCS) or a vegetative state (VS) who are otherwise utterly unresponsive. Their convincingdata raise themore general question that I consider here of how these findings impact the nosology and diagnostic criteria of chronic disorders of consciousness. In their classic text The Diagnosis of Stupor and Coma3 in 1966, Plum and Posner explained that consciousness can be assessed in its 2 clinical dimensions: wakefulness and awareness. Healthy people have both and comatose patients have neither. They also defined locked-in syndrome (LIS) as a state of profound paralysis (de-efferentation) with intact cognition that may be mistaken for unconsciousness by an unwary examiner who is misled by the patient’s markedlydiminishedvoluntarymotor repertoire. In 1972, Jennett and Plum4 defined the persistent VS (PVS) as the ironic and tragic dissociation of the 2 components of consciousness: wakefulness remained preserved despite abolished awareness. In 1994, theMulti-Society Task Force on PVS provided clinical criteria for PVS that have been accepted for the past 2 decades,which at aminimumrequire the complete absence of any evidence of awareness of self or environment by a careful neurological examination.5,6 The task force offered the caveats that any determination of the quality of a person’s awareness was entirely inferential based on a reasoned interpretationof theperson’s responses to stimuli and that the absence of clinical evidence of awareness did not necessarily constitute proof of unawareness, a point made more forcefully by critics.7 The task force’s clinical criteria for the VS were delineated almost entirely as negatives—stipulating what behaviors patients inVS lacked the capacity todo—all ofwhichmust be absent.5,6 The task force’s choice of negative criteria for the syndromewas accurate clinically but invited false-positivediagnosticerrorsbecause the fact thataparticularbehaviorcould notbeobservedoreliciteddidnotnecessarilyprove that itwas fully absent. The failure to elicit a behavior could result from factors other than the direct effects of the primary brain injury, such as from inadequacies of the neurological examination, samplingerror, the failureof thepatient to respond to the voice of a physician to the same extent as to a beloved family member, concomitantmetabolic or toxic encephalopathies, or adverse effects of medication. The situation became better defined in 2002 when a new diagnostic entity, the MCS, was delineated to classify profoundly unresponsive patients who demonstrated behaviors that requiredat leastperiodicor fragmentaryawareness.8 Like theVS, theMCS is a heterogeneous clinical syndromeencompassing diverse pathologies but sharing certain clinical findings. When the diagnostic criteria were scrupulously followed, it became clear thatmanypatients initially believed to be inVSwereactually inMCS,which is farmorecommon.1Naci and Owen cited studies showing that false-positive diagnostic error for VS remains disturbingly high. It also seems likely that the task force’s high estimate of VS prevalence resulted frommisclassification of some patients inMCS as being in VS because MCS was not a recognized diagnosis in the brain injury databases studied by the task force. In2006, theassumptionthatVScanbediagnosedwithperfect accuracybyaneurological examinationwas shakenby the brief reportofOwenetal9of ayoungwoman inVS for 5months following a traumatic brain injury whose fMRI measurements duringmental imagery tasks shewas asked to perform disclosedthatshewasaware.Theneurologyandcognitiveneuroscience communities initially were skeptical and questioned whether the elicited findings were artifact or, if they were valid, what quality of conscious cognitive life they indicated. Since then, Owen et al and other investigators have reported similar findings of “willful modulation of brain activity”10measured in fMRIparadigms in a fewpatients inVS and MCS who were clinically unresponsive, as in their current report. Notably, the number of these cases remains small andmost patients had traumatic brain injury, particularly diffuse axonal injury, and not hypoxic-ischemic neuronal injury. Many of the patients in VS showing this capacity subsequently improved to clinically demonstrable awareness, suggesting that the finding of willful modulation of brain activity also may be a predictive marker for clinical improvement in those patients for whom VS or severe MCS represented a transient state of unresponsiveness on a path to spontaneous improvement after traumatic brain injury. The findingof fMRIevidenceofawarenessdespitenoclinical evidence of awareness suggests that, at times, even anoptimal neurological examinationmay not be sensitive enough todetect awareness.Most studies haveused theComaRecovery Scale–Revisedor a similar specializedneurological examiEditorial page 1229