What does it mean, to be conscious? For Wilder Penfield, consciousness represented “a thinking, a knowing, a focusing of attention, a planning of action, an interpretation of present experience … (an) integrated perception of the present.”1 Notwithstanding philosophical, medicolegal, and sociocultural definitions, by Penfield’s time, in the mid-20th century, it had become clear that consciousness was dependent on a specific region in the brainstem known as the reticular activating system (RAS), which together with pathways that traveled between the brainstem, thalamus, and cerebral cortex provided the neuroanatomic basis for consciousness.

Our current understanding of the brainstem and its fundamental importance to consciousness is remarkable considering that, until the late 18th century, the brainstem was still widely regarded as a simple extension of the spinal cord.2 It was not until the 19th century that anatomists began to progressively localize cranial nerves, motor pathways, and respiratory centers to different levels within the pons and medulla oblongata, lending growing support to the notion of the brainstem as a mediator of life-sustaining functions.3 By the 20th century, the core structures involved in consciousness had been mapped to regions of the brainstem and their rostral projections. In modern clinical practice, the central importance of the brainstem is manifested by its significance in neuroprognostication after brain injuries and its central place in medical and legal definitions of brain death/death by neurologic criteria (BD/DNC).

In this narrative review, we describe the evolution of our modern anatomic understanding of the RAS and explore how that knowledge has informed our understanding of disorders of consciousness and current definitions of BD/DNC. We begin by reviewing the foundational 20th-century experiments that identified the neuronal networks which mediate arousal and awareness. We then outline the connections between the RAS and link our anatomic knowledge of these structures to their central importance in consciousness and in the neurologic determination of death. We conclude by exploring important differences across jurisdictions in BD/DNC definitions and the question of inferring complete and permanent loss of consciousness—one of the core requirements to declare brain death—in patients with isolated infratentorial injuries.

Progressive anatomic mapping of the brainstem, RAS, and thalamocortical circuits

By the 19th century, neuroanatomists had shown that different parts of the brainstem mediated vital physiologic functions without which life could not be sustained. In the early part of the century, the French anatomists Cesar Julien, Jean Legallois, and Jean Pierre Flourens separately performed transection experiments in rabbits and pigeons, in which they localized respiratory activity to the brainstem and more specifically to a 1-mm region at the base of the brain, which Flourens called the noeud vital (“vital knot”).4,5 In 1822, Bell identified a similar respiratory center in donkeys, which he was able to localize to the lateral medulla oblongata.6 These transection experiments grossly localized the respiratory centers to the brainstem, but another one hundred years would pass until more sophisticated experiments allowed precise identification of the neural networks that control respiration.

The advent of new methods of neuromonitoring, specifically electroencephalography (EEG), allowed for more detailed insight into the topographical importance of the brainstem to consciousness. In 1937, the Belgian physiologist Frederic Bremer found that cutting the base of the feline brainstem at the level of the medulla (the so-called encephale isolé preparation) did not disrupt the animal’s EEG sleep-wake cycle, whereas cutting within the midbrain between the superior and inferior colliculi (cerveau isolé) resulted in a persistent EEG pattern of spindles and slow waves.7 Experiments by Giuseppe Moruzzi and Horace Magoun showed that stimulating a region between the pontine and midbrain tegmentum in Bremer’s encephale isolé animals rapidly converted a slow-wave, sleep-like EEG pattern into one resembling wakefulness. Stimulation would not modify the slow-wave EEG trace with more rostral lesions at the cerebral peduncles or tectum.8 These findings led Moruzzi and Magoun to delineate with remarkable precision the regions of the brainstem that were necessary for regulating arousal and normal sleep (of note, the specific neurons necessary for sustaining consciousness remain unknown).9 Figure 1 indicates the major findings and years over which these experiments occurred.

Fig. 1
figure 1

Historical advances in understanding the anatomy of consciousness

Full size image

Around the time that Berger, Moruzzi, and Magoun were performing their animal experiments, 20th-century clinicians were beginning to link the brainstem to clinical disorders of consciousness. In 1917, the Austrian neurologist Constantin von Economo described his findings in patients afflicted by encephalitis lethargica, an atypical form of encephalitis that targeted neurons in the brainstem and caused a profound and progressive somnolence.10 Later, Canadian neurosurgeon Wilder Penfield observed that damage to structures within the brainstem produced a clinical state that was incompatible with wakefulness, whereas patients with extensive resections of their frontal lobes could still maintain arousal.11 Exploring more superior and caudal structures, Jack French and Horace Magoun further observed that destroying specific regions of the median cephalic portion of the brainstem including the midbrain tegmentum, subthalamus, hypothalamus, and medial thalamus in animal models produced “stupor waves” on the EEG similar to those of Bremner’s cerveau isolé preparation and extinguished environmental awareness or purposeful movement.12

These experiments have led to our modern understanding of the reticular formation as a collection of neurons within the brainstem, extending from the level of the rostral midbrain to the rostral medulla. Named for its loose, web-like (“reticular”) appearance, this meshwork supports distinct groups of nuclei that cluster together and mediate various physiologic functions, including breathing, circulation, pain modulation, and generation of repetitive motor movements (e.g., chewing and swallowing). The ascending RAS, which is relevant to arousal and consciousness, is localized in the central and dorsal brainstem between the caudal midbrain and midpons. Projections from the RAS extend to the basal forebrain, hypothalamus, and intralaminar nuclei of the thalamus. The thalamus is a primary relay point for ascending fibers from the RAS, and thalamocortical circuits play a key role in generating oscillatory activity associated with conscious states, slow-wave sleep, and EEG spindles; disruption of these circuits has been implicated in various disorders of consciousness.13 A small number of fibers bypass the thalamus and project directly to the cortex, as depicted in Fig. 2.

Fig. 2
figure 2

Anatomy of the reticular activating system

Full size image

The RAS and thalamocortical system in disorders of consciousness

The clinical impact of lesions of the RAS, thalamus, and cortex on consciousness vary depending on the location and completeness of the lesion. Lesions of the pontine tegmentum are capable of producing a state of coma, in which the sleep-wake cycle is lost, the eyes remain closed, and there is no speech or purposeful movement.14 Cortical metabolism and activity may be virtually undisturbed by such lesions and yet the clinical syndrome is one of profound unresponsiveness.15 Diffuse cortical or selective bilateral thalamic injuries may produce a persistent vegetative state or unresponsive wakefulness syndrome,16,17 in which patients maintain sleep-wake cycles but show no ability to communicate or engage purposefully with their environment. Autopsies in such patients (including Karen Ann Quinlan, whose case gained worldwide prominence)17 have confirmed severe damage to the thalamus and/or thalamocortical white matter tracts.18,19 Less profound cortical or thalamic injuries can lead to a condition known as the minimally conscious state. Patients in this condition show some level of awareness of their environment: they can often reach for objects and track movements with their eyes, although such reactions are not always consistent.20

Modern EEG and imaging studies have reinforced the association between the degree of damage to the global RAS and the severity of the resulting disorders of consciousness. Using quantitative EEG in patients with subarachnoid hemorrhage, Claassen et al. found that a greater degree of disconnection between the cortex, thalamus, and basal forebrain was associated with a lower Glasgow Coma Score.21 In studies using functional magnetic resonance imaging (fMRI), researchers observed a higher level of wakefulness in patients when both corticocortical and corticothalamic connections were preserved, whereas more profound impairments were seen when corticocortical connections were preserved but corticothalamic ones were not.22,23 Reports using MRI-based diffusion tensor imaging sequences have been used to show white matter disruption within the RAS in patients with profound disorders of consciousness (e.g., due to traumatic brain injury or cardiac arrest).24,25,26,27 Our understanding of the neuroanatomical basis for consciousness continues to evolve with new insights into the topography and connectivity among these regions. Nevertheless, the core notion that some viability of the RAS is necessary for consciousness remains unchanged.

The central importance of the brainstem and RAS in BD/DNC

Brain death/death by neurologic criteria, defined conceptually by the permanent and complete loss of the capacity for consciousness and all brainstem function, including the capacity to breathe, is distinct from disorders of consciousness.28 These foundational criteria for BD/DNC reflect the concept that vital functions for life and arousal are governed by pathways that originate from, or travel through, the brainstem. Consequently, the assessment of brainstem function plays a central role in the determination of BD/DNC, based on the premise that permanent and complete damage to the RAS (as shown by unresponsive coma and a lack of demonstrable brainstem function) precludes consciousness. This premise is further supported by evidence showing that BD/DNC is associated with damage to the brainstem on postmortem examination.29

The central importance of the brainstem and RAS for consciousness has made the permanent loss of brainstem function an enduring criteria of BD/DNC from its first description by Mollaret and Goulon in the 1950s30 to modern definitions and guidelines.31,32,33 Since the publication of the seminal Harvard Criteria, which provided the first set of formal criteria for establishing “irreversible coma,”31 the construct of BD/DNC has been adopted across jurisdictions with subsequent updates incorporating more specific instructions for the clinical exam and ancillary testing.28,32,33,34 Although there are important historic, national, and even institutional variabilities in the procedures for determining BD/DNC, all definitions consistently include the loss of brainstem function as a central feature.28

“Brainstem death,” “whole brain death,” and unanswered questions

Despite recent initiatives to reach consensus in the definition of BD/DNC, which center around the primacy of the brainstem and RAS,28 there continues to be heterogeneity in definitions across jurisdictions, including differing formulations of “whole brain death” and “brainstem death.” The UK and several other jurisdictions adopted the stance that brainstem death, as defined by the permanent and complete loss of all brainstem function and the capacity for consciousness, was sufficient for death of the whole person. This position was justified on the premise that the core processes of life and consciousness require some preserved brainstem function.35,36 By contrast, the USA adopted legislation stating that “an individual who has sustained an irreversible cessation of all functions of the entire brain, including the brainstem, is dead.”37 Despite conceptual differences and theoretical arguments that the brainstem definition entails a lower burden of proof than the whole brain definition,38 both formulations share common prerequisites and are supported by the same clinical assessment.39 This concordance stems from the fact that BD/DNC is most commonly the result of a severe supratentorial or diffuse global injury that leads to brainstem destruction by mechanical compression or complete infarction from loss of intracranial blood flow.40

Nevertheless, where the two formulations may potentially diverge is in the uncommon scenario of an isolated infratentorial injury causing BD/DNC.41,42 In jurisdictions that follow the whole brain death formulation, ancillary testing (e.g., with angiography or radionuclide imaging) is suggested when BD/DNC is suspected from an isolated brainstem lesion. Loss of intracranial blood flow with flow-based ancillary tests supports the clinical determination of BD/DNC in this scenario.28 Ancillary testing is not mandated in the brainstem formulation. Notwithstanding the limitations of ancillary testing and the fallacy of inferring intact function from intact flow, a concordant ancillary test that supports the clinical exam in patients with isolated brainstem injuries may improve clinicians’ confidence in diagnosing BD/DNC.43 One Canadian study found that clinicians ordered ancillary testing in 10/12 (85%) patients undergoing assessment for BD/DNC after infratentorial ischemic or hemorrhagic strokes compared with only 25/47 (53%) patients after supratentorial strokes.42 Interestingly, the authors documented presence of some supratentorial blood flow or perfusion in five out of six patients with infratentorial strokes, although in all cases infratentorial blood flow was absent or abnormal.42 Ancillary testing may also be perceived as providing additional certainty of BD/DNC in patients with isolated brainstem injury. This perception was supported by a Canadian survey of intensive care physicians, which found that over half believed ancillary testing should be performed in this context.44

Despite the apparent inconsistency between the two formulations of BD/DNC, there are reassuring data that the natural history of patients who appear brain dead after isolated infratentorial injuries is to progress to infarction of the supratentorial brain, thus eventually satisfying criteria of the “whole brain death” construct.28 Mechanistically, this progression may occur through progressive hydrocephalus and increased intracranial pressure leading to whole brain infarction. One study of patients with isolated brainstem injuries and clinical exams consistent with BD/DNC observed that all patients who had initial evidence of supratentorial blood flow on computed tomographic angiography (CTA) lost that flow with repeat imaging.41 A neurophysiologic study found that EEG activity persisted in 3.5% of patients who were clinically determined to be brain dead, but this activity disappeared with repeated recordings.45 Although not always temporally synchronized, there are data showing eventual convergence of ancillary testing (e.g., CTA or EEG) and a clinical diagnosis of BD/DNC.46 Nevertheless, specific situations should raise caution, as in the case of patients with posterior fossa decompressions for a primary infratentorial lesion. A case was recently reported of a patient with a cerebellar hemorrhage who underwent posterior fossa decompression and who had return of spontaneous respiration 12 hr after completing a clinical examination for BD/DNC.47 Return of respiratory activity was thought to be caused by relief of posterior brainstem compression by delayed herniation of the cerebellum through the skull defect. Ancillary testing with documentation of intracranial circulatory arrest has therefore been advised to support the determination of BD/DNC in this scenario.47

Conclusion

Clinical and neurophysiologic experiments have made clear that the RAS is a fundamental requirement for consciousness. This is further emphasized in modern medicolegal requirements of BD/DNC, in which destruction of the RAS and loss of capacity for consciousness are inferred by the complete absence of brainstem function. Nevertheless, a particular challenge for clinicians may arise in patients with isolated posterior fossa injuries, where the possibility of residual viability of the RAS or its projections makes it difficult to infer complete and permanent loss of consciousness in all cases with complete certainty. Notwithstanding that these concerns are mostly theoretical, they do provide an occasion to reflect on current clinical practices as they relate to this rare scenario. In whole brain death jurisdictions, ancillary testing has been suggested in patients with isolated infratentorial injuries before finalizing the determination of BD/DNC.28 A supportive ancillary test may also provide reassurance to clinicians when the capacity for consciousness cannot be definitively disproven—whether because of the nature or pattern of the brainstem injury or because of residual confounders. Furthermore, since clinical criteria and ancillary tests seem to achieve eventual if not immediate concordance, there is little concern that waiting will change the initial presumption of BD/DNC. It may, however, mitigate the perception of having made a premature diagnosis. These factors may collectively favor obtaining ancillary tests to support the clinical determination of death in patients with isolated infratentorial injuries.