Neural Correlates of Consciousness: A Comprehensive Review
In this review, we summarize the current understanding of consciousness including its neuroanatomic basis. We discuss major theories of consciousness, physical exam-based and electroencephalographic metrics used to stratify levels of consciousness, and tools used to shed light on the neural correlates of the conscious experience. Lastly, we review an expanded category of ‘disorders of consciousness,’ which includes disorders that impact either the level or experience of consciousness. Recent studies have revealed many of the requisite EEG, ERP, and fMRI signals to predict aspects of the conscious experience. Neurological disorders that disrupt the reticular activating system can affect the level of consciousness, whereas cortical disorders from seizures and migraines to strokes and dementia may disrupt phenomenal consciousness.
Introduction
The term “consciousness” often carries different definitions.
- Phenomenal consciousness: our personal, subjective experience of the present moment.
- Level of consciousness: the state of wakefulness assessed by various scales described in terms such as comatose, stuporous, obtunded, delirious, confused, and alert.
A better understanding of consciousness and its neurobiology may improve the clinical care of patients with disorders of consciousness, including not only anoxia and severe traumatic brain injury but also delirium, epilepsy, migraine aura, cortical strokes, and cortical dementias.
Theories of Consciousness
There are at least twenty-two supported neurobiological explanations for the basis of consciousness. Each of these theories works towards understanding the neural correlates of consciousness-the neurobiological mechanisms necessary and sufficient to generate a conscious experience.
Here are some prominent theories:
- Higher Order Theory (HOT): claims that consciousness depends on meta-representations, representations that occur higher in a processing hierarchy.
- Global Workspace Theory (GWT): postulates a centralized resource shared by multiple independent threads of processing.
- Local Recurrency Theory (LRT): argues that recurrent processing in sensory cortices is sufficient to generate conscious experience while parietal and frontal regions are required only for reports and judgements pertaining to a stimulus.
- The Memory Theory of Consciousness (MToC): proposes that the phenomenonal consciousness evolved from, and functions as a part of, episodic memory and other explicit memory systems (sensory, working, and semantic memory). The entire cerebral cortex and hippocampus function together to serially integrate and time stamp parallel unconscious brain processes to form a linear, coherent stream of conscious experiences. This theory proposes that the experience of consciousness is actually a memory, preceded temporally by unconscious processes of sensation, decisions, and/or actions. The theory also proposes that different cortical areas produce different aspects of consciousness, each with its own neural correlate. This theory can explain long-standing puzzles in consciousness such as postdictive effects-the finding that conscious perception of a stimulus can be affected not only by conditions preceding a stimulus, but also by the conditions following a stimulus.
Global Workspace Theory
Clinical Metrics of Levels of Consciousness
Level of consciousness, that is, the state of wakefulness, is commonly defined clinically through descriptors of psychomotor activity in response to stimulus.
Commonly used scales include:
- The Glasgow coma scale (GCS): is perhaps the most widely used clinical scale of consciousness, favored for its three metric simplicity (eye opening, verbal response, motor response). However, in recent years, it has drawn criticism for lack of content validity, standardization, and inter-examiner reliability.
- The Richmond Agitation-Sedation Scale: A simpler measure favored in the intensive care unit (ICU) setting is the Richmond Agitation-Sedation Scale, which plots wakefulness on a single dimension from combative behavior to complete unresponsiveness.
- The Coma Recovery Scale Revised (CRS-S): Another scale with high inter-examiner reliability, content validity, standardization, and prognostic capability is the Coma Recovery Scale Revised (CRS-S).
Patients may, however, be clinically unresponsive and maintain electroencephalographic (EEG) or functional MRI (fMRI) activity that suggests underlying consciousness. In a large, ICU-based study of brain activity, Claasen et al. used EEG patterns to detect occult brain activity in response to spoken motor commands in patients with acute brain injury, observing that about 15% of clinically unresponsive patients have a “conscious” EEG pattern in response to spoken commands. Follow-up investigations found that 41% of patients with covert consciousness make a complete recovery versus only 10% of patients without covert consciousness.
EEG Activity as a Correlate of Consciousness
EEG is the gold standard in determining the sleep-wake phase and is a useful tool in assessing the level of consciousness. The Bispectral Index (BIS) is used by anesthesiologists to assist in the titration of anesthetics. Raw EEG data is collected from four electrodes on the forehead and analyzed by a proprietary, algorithm which outputs a unidimensional score, the BIS, which quantifies the level of wakefulness from 0 (comatose-flat line EEG) to 100 (awake-low amplitude, fast frequency).
EEG frequency ranges and measures of connectivity have been considered as potential correlates of phenomenal consciousness (or at least aspects of it). The alpha frequency (8-12 hz), canonically considered to be an idling rhythm, has recently been thought to have a functional role in the inhibition of task-irrelevant information. Pre-stimulus alpha activity predicts poor conscious perception; occipital alpha power might filter out which stimuli will be perceived consciously, and which stimuli will be deemed “task irrelevant” and ignored. For example, Krasich et al. showed that about halfway between the maximum positive and negative peaks of occipital alpha activity, a stimulus is mostly likely to be consciously perceived. In the post-stimulus period, a drop in parietooccipital alpha power has been suggested as a marker of conscious perception. Hutchinson et al. studied a paradigm of inattentional blindness and confirmed that alpha power 250-500 ms post stimulus is negatively associated with conscious perception.
ERP Activity as a Correlate of Consciousness
Event-related potentials (ERPs) are EEG changes time locked to sensory, motor, or cognitive events that provide a means to study electrophysiologic correlates of mental processes. ERPs provide excellent temporal resolution and are a practical (non-invasive, relatively inexpensive) method of studying cognitive functions.
Each of these factors can be experimentally controlled and modified in ERP study paradigms to determine the brain signals that are together necessary and sufficient for conscious perception-the neural correlates of consciousness. The conscious perception of a stimulus depends on a variety of factors including internal features of the stimulus itself (such as its salience). External factors that affect stimulus perception include observer state (such as attentional) and context (such as masking) which may affect perception preceding, simultaneously, or following the stimulus.
Early ERP components, peaking within 100 ms after stimulus, are known as “sensory” or “exogenous,” as they depend mostly on the physical parameters of the stimulus itself. The middle and late components, generated 100-300 ms and 300-600 ms after stimulus, respectively, reflect stimulus evaluation and are termed “cognitive” or “endogenous” potentials.
Early ERP Components (P50, N100, and Others)
In their historic work, Libet et al. reported in 1967 that early ERP sensory evoked potentials over the primary somatosensory cortex are not always sufficient to cause perceptual experience. This finding, replicated and expanded upon by others, started the search for other early potentials to determine the first signals sufficient to predict sensory experience, also known as perception.
Middle ERP components (P100, N140, ERN/Pe)
- P100: The P100, a positive occipital ERP component occurring 100 ms after stimulus presentation, has been implicated as the earliest ERP involved in conscious perception and associated tasks. Other work, however, found the P100 in both perceived and unperceived conditions.
- N140: The N140 is a negative ERP peaking around 140 ms. Multiple studies highlight the N140 as important for perception but there is debate as to whether the signal should be interpreted as awareness, perception, detection, or attention. Given the prolongation of the N140 in patients with clinical deficits in contralateral spatial attention, we favor this specifically as a marker of somatosensory hemi-attention rather than conscious perception.
Late ERP Components (late Pe, P300, VAN)
Our review of early and middle ERP components suggests that they are mainly correlated with attention or unconscious sensory processes. Later, ERP components may be more likely to represent true neural correlates of consciousness.
- Positivity Performance Error (Pe): The late positivity performance error (Pe) component (400-600 ms) has been associated with conscious awareness of error detection.
- P300: The P300 is a posterior ERP observed 290-450 ms after a stimulus. Historically, the P300, which includes components P3a (central maximum) and P3b (parietal maximum), has been thought to represent a gold-standard marker of conscious perception. Recently, however, the P300 has been favored to specifically represent post perceptual processing, rather than phenomenal conscious awareness.
- VAN: The visual awareness negativity (VAN) is a negative difference wave, observed 200 ms after stimulus presentation, most prominently over occipitotemporal sites, which differentiates aware from unaware states. The VAN may be the earliest reliable electrophysiologic marker that correlates with aware perception after the presentation of a visual stimulus. There is an analogous auditory awareness negativity leading to recent proposals to consider these potentials as parts of a broader termed “perceptual awareness negativity”.
Connectivity as a Correlate of Consciousness
In the conscious patient, a TMS pulse sets off a complex chain of events as information reverberates through cortical brain regions, but in the unconscious patient, a similar pulse may induce only a short-lived local effect. This finding supports and provides a testable metric for IIT that applies a connectivity approach to evaluate levels of consciousness.
Studying EEG recordings in patients with diminished levels of consciousness as determined by GCS, Toker et al. proposed that state of consciousness correlates with the proximity of slow cortical oscillatory potentials to a critical pattern described as the edge-of-chaos. In a serial awakening paradigm, exploring EEG phase connectivity as a correlate of consciousness, patient states were differentiated across sleep stages by measures of EEG phase connectivity in posterior brain regions. Similarly, using fMRI, Demetrezi et al. found that compared to healthy controls, those with decreased level of consciousness showed impaired baseline connectivity in bilateral executive control, default mode, and auditory networks. Luppi et al. employed resting-state fMRI with measures of integration and entropy to identify consciousness-specific patterns of brain function including temporal states of high integration and functional diversity.
Aspects of Consciousness
If we assume that consciousness is unitary, prompting us to look for a single neural correlate of consciousness, then it is difficult to make sense of these divergent results from various studies using different measures of brain physiology. If, however, we take the approach suggested by the MToC-that there are various aspects of consciousness such as visual, auditory, emotional, phonological, motor, and more-then each study may reveal the neural correlate of consciousness for one aspect of consciousness but not others. This modular view of consciousness is entirely consistent with the disorders of consciousness encountered daily by the neurologist.
Disorders of Consciousness
Classically, this group included only disorders that directly impact the level of consciousness, such as TBI and cardiac arrest. Following the MToC, however, we broaden this definition to include disorders that affect phenomenal consciousness.
Levels of Consciousness
Seizure
Epileptic seizures offer a window into the requisite anatomy and function for maintaining consciousness as different seizure subtypes affect consciousness differently. A generalized onset seizure includes bilateral motor involvement, loss of consciousness, and involvement of bilaterally distributed networks. The fact that complete loss of consciousness in seizure requires generalized involvement of both hemispheres argues strongly that level of consciousness is a diffuse cortical phenomenon, consistent with MToC but not HOT, GWT, IIT, or LRT. Focal seizures are subclassified as impaired awareness or unimpaired awareness. Focal seizures with altered awareness also impair phenomenal consciousness, as manifested by, for example, amnesia, automatisms, and behavioral arrest.
Migraines
Although migraine does not generally affect the level of consciousness, Budson et al. argued that migraine aura is, in fact, a focal disruption of phenomenal consciousness. Rarely, in the case of the basilar sub-type, migraine can affect the level of consciousness, causing confusion. More commonly, migraine aura results in the impairment of various aspects of phenomenal consciousness, such as visual associated with posterior onset of cortical spreading depression. Positive followed by negative visual phenomena (scintillating scotoma) are common changes associated with focally impaired conscious awareness.
Strokes and Other Discrete Brain Lesions
Stroke, discrete in border and stable over time, is the prototype brain lesion and offers a means to evaluate the neuroanatomic basis of consciousness. We will first discuss how strokes affect the level of consciousness, and then how strokes can affect various aspects of phenomenal consciousness. At onset, level of consciousness is impaired in about 36% of patients with large hemispheric infarcts and 14% of stroke patients as a whole. Clinical attention to level of consciousness is critical as early decreased consciousness may portend a 2.2-fold risk of in-hospital mortality.