The Global Neuronal Workspace Model and Probing
Consciousness in Nonresponsive Patients
Door: Katinka Pennings
Bachelorthesis
University of Amsterdam
Supervisor: Simon van Gaal
Student number: 10047603
Number of words: 6600
Index
Abstract ……….. 3 Introduction ……… 3 Chapter 1: The neural correlate of consciousness in healthy wakeful people ……... 5 Chapter 2: Nonconscious and conscious processing in altered states of consciousness … 10 Chapter 3: Detecting awareness in nonresponsive patients ………... 14 Conclusion and discussion ………. 20 Bibliography ………. 23
Abstract
This paper will examine the global neuronal workspace model of consciousness and will integrate the model in the effort to probe awareness in patients suffering disorders of
consciousness. First, several studies on healthy wakeful people have shown unconscious processing in higher order brain areas. Besides, awareness is due to great interconnectivity in the brain whereby information becomes globally available. In other studies researchers have demonstrated that the model also applies for a conscious state. At last, further studies have shown the possibility of detecting awareness in nonresponsive patients and offer a reliable technique of bedside testing according to the global neuronal workspace model. What can be concluded from these studies is that the global neuronal workspace model provides a good theory about the concept of consciousness and is useful when probing awareness in nonresponsive patients.
Introduction
Consciousness is a concept in psychological science that has been discussed for centuries. The philosopher Descartes already mentioned the interest of consciousness many years ago. His “cogito ergo sum” (I think, therefore I am) can still be seen as the essence, yet at the same time the problem of consciousness. It implies that only we know ourselves whether we are conscious or not, which ensures our existence. Defining the consciousness of someone else should be impossible in this view. But in addition, Descartes also proposed a way to solve uncertain issues, gradually decomposing knowledge into true statements; a way that we now know as science. Defining the elusive concept of consciousness in a scientific way is something psychologist Bernard Baars (1989) tried. “Conscious experience is notoriously the great, confusing, and contentious nub of psychological science. We are all conscious beings, but
consciousness is not something we can observe direcly, other than in ourselves, and then only in retrospect” (UCLA, 1989). Nowadays, observing consciousness directly seems to be more
plausible. Although Baars had no access to modern brain imaging, his theory of consciousness has been the basis of recent theories of consciousness. Baars proposed consciousness as the result of a Global Workspace in the brain that distributes information to a huge number of parallel unconscious processors in the rest of the brain (McDermott, J., Harvard Undergraduate Society for Neuroscience, 1995). Stanislas Dehaene and colleagues proposed a modernized version of Baars’ theory called ‘the global neuronal workspace model’ (Dehaene et al., 1998). The key idea behind this model is that because of its great interconnectivity, workspace neurons, particularly in frontal and parietal regions, can project their coherent activity to other brain processors, by which the processed information becomes globally available (Gaillard et al., 2009). What we experience as consciousness, Dehaene says, is this global availability.
Consciousness therefore means brain-wide information sharing. Modern brain imaging such as functional magnetic resonance imaging (fMRI) is available to detect this global availability of information processing in the brain, therefore making it possible to distinguish between
whether a person is conscious or not. If that is possible, it might also distinguish between people who are not able to report their consciousness, such as people who are in a state of vegetation (Kalat et al., 2014).
Consciousness is divided into two components: content and state. Content relates to awareness, whereas state relates to vigilance or wakefulness. Fully aware means the ability to have a subjective experience, whereas fully awake equals arousal; a state of heightened physiological activity in which the body is reactive to stimuli and ready to respond. Besides wakefulness, the state of consciousness can differ. The vegetative state (VS) is a clinical condi-tion that is often described as ‘wakefulness without awareness’. The patient has a normal sleep-wake cycle and can breathe on his own, but shows no other cognitive or physical response to stimuli. When a patient emerges from the vegetative state showing some cognitive function it is labeled as being in a minimally conscious state (MCS) (Fernandez-Espejo et al., 2013). Every
year, around 2000 people get into a state of vegetation (Hersenstichting, n.d.). Approximately 40% of these patients are minimally conscious but are misdiagnosed as being in the vegetative state, because of showing signs of awareness (Monti et al., 2010). This demands a more accurate method of bedside testing of consciousness, because it is relevant for predicting recovery of their consciousness and cognitive abilities. It will also provide a way to communicate with these patients and question the withdrawal of nutrition (Fernandez-Espejo et al., 2013).
This paper will examine consciousness in several ways, leading to answer the main question; does the global neuronal workspace model offer a reliable theory as regards to the neural correlate of consciousness and does it provide a basis for developing ways of bedside testing based on this theory? In the first chapter I will take a look at the neural correlate of consciousness in healthy wakeful people and argue the global neuronal workspace model. To further align the model and provide a way of answering the main question, in the second chapter I will highlight some findings on the neural correlate of consciousness during several altered states of consciousness, like anesthesia. Last, taking Dehaene’s model and previous findings into account, in the third chapter I will take a look at awareness in patients suffering disorders of consciousness and will come up with a reliable method of bedside testing.
1. The Neural Correlate of Consciousness in Healthy Wakeful People
As noted by Baars (1989), the study of the underlying mechanisms of consciousness requires an experimental design in which certain information is processed consciously and in a similar situation the exact same information is processed nonconsciously. Two techniques to realize this design are pattern masking and dichoptic masking. Pattern masking has been used as a technique to create a subliminal stimulus, one that is undetectable due to a close spatial and temporal contingence of other stimuli (Dehaene et al., 2011). During pattern masking, a stimulus is presented quickly followed by a mask, resulting in the invisibility of the stimulus. This
stimulus could also be an additional prime, called masked priming. Dichoptic masking is induced when different images are presented in rapid succession to the two eyes, leading to the
perception of just one of the images (van Boxtel et al., 2007). As already noted, determining whether the global neuronal workspace model is a plausible theory of consciousness, first, information is needed about healthy wakeful people and the actual detection of awareness. The first study uses pattern masking to visualize nonconscious processing in the brain.
Dehaene et al. (2001) used fMRI and event-related potentials (ERPs) to visualize the cerebral processing of unseen masked words. They tried to image the brain areas that were activated by masked words. This experiment included three conditions in which the participants were exposed to a visual stream of random shapes and blanks, either not interrupted or
interrupted by unmasked- or masked words. Brain activity was measured with fMRI and EEG. ERPs were obtained by analyzing the results to state clear which brain activity was caused by masked words alone. The masked words showed activation in left extrastriate, fusiform and precentral regions, areas normally related to word processing. The results also revealed three ERP components corresponding to the activity in extrastriate, fusiform and precentral areas. In addition, the masked words also caused reduced activity in prefrontal and parietal brain areas. On the other hand, the visible words were related to activation in the fusiform gyrus, left parietal cortex, inferior frontal, anterior cingulate, precentral cortex and supplementary motor area. The visible words also led to an increase in functional connectivity. These results suggest the
presence of unconscious processing in areas related to conscious reading and the absence of unconscious processing in prefrontal and parietal regions, in contrast to conscious processing.
Masking provides a method for distinguishing automatic processes that occur without consciousness. However, these results do not implicate lexical or semantic processes. It is plausible that the prior study only suggests activation at the letter level and therefore it is interesting to further investigate whether the automatic processes that are found by Dehaene et al. (2001) can be generalized to actual semantic processes. Comparing unconscious processing due to nonwords and words will discriminate semantic processes from non-semantic processes.
In the following study the researchers compared fMRI activation related to masked words and masked nonwords.
Diaz and McCarthy (2007) examined whether masked words are processed to the semantic level. The participants first performed a lexical decision task (LDT). The researchers used pattern masking by quickly priming the participants with several (non)words followed by a masked word/nonword, semantically related or unrelated to the prime. In the LDT, the
participants needed to decide quickly whether a (non)word existed or not. Then, masked words/nonwords were presented to the participants, while lying in the scanner. The results showed that the participants performed faster on the LTD when the masked words were
semantically related. On top of that, fMRI showed specific activation in the frontal gyrus, angular gyrus and inferior temporal gyrus, on the left side of the brain, due to masked words. These effects show that unconscious processing in word related areas in the left side of the brain is not just a result of a series of vowels but is also semantically related. Unconscious word processing thus occurs at a high level that is related to specific areas on the left side of the brain.
The research of Diaz and McCarthy (2007) has made considerable progress in the notion that unconscious word processing activates specific left parts of the brain, in the same way conscious word processing does. Now it is known that although we are not conscious of reading words, that does not mean we do not take them into account.
From prior research it has been known that the fusiform face area (FFA) is activated during detection or identification of faces (Summerfield et al., 2006, in Kouider et al., 2009). However, whether activity in face-responsive regions always correlates with perceptual
awareness remains unclear (Kouider et al. (2009). Kouider et al. (2009) examined the activity in face-responsive brain regions by masked and non-masked faces. To do so, the researchers used masked priming, making use of subliminal and visible primes, while the participants performed a fame-judgement task on famous and nonfamous faces. The researchers examined brain activity by using fMRI while the participants tried to perform the task as quickly as possible. The results showed that reaction times were faster on familiar faces, rather than nonfamiliar faces. The
participants also reacted more quickly when the familiar faces were primed due to the effect of the already processed prime. A lower reaction time on primed faces showed related decreased activity in classic face-processing areas, including the FFA, occipital face area, and superior temporal sulcus. Besides, when the participants perceived the visible faces consciously, which caused additional activity in frontal and parietal regions. These results provide evidence that the processing of faces in classic face-processing areas can occur without the presence of perceptual awareness and that frontal and parietal regions play a role in conscious processing of faces.
Previous research has shown that information processing can occur without awareness. The question is why information processing in word- or face related areas not necessary
implicate awareness. Conscious perception seems to be related to increased activation in fronto-parietal regions, but which underlying neuronal process turns complicated neuronal responses into conscious perception? According to Dehaene this is where the global neuronal workspace emerges. The next study will take a look on his model by focusing on the relationship between category tuning (low level of processing) and figure-ground segregation (high level of
processing). Category tuning does not provide conscious perception of object recognition. Prior research has shown that object recognition influences figure-ground assignment and might even precede it (Fang & He, 2005, in Fahrenfort et al., 2009). Resolving the link between those two processes might give us more information about the nature of conscious perception.
Fahrenfort et al. (2012) explored this question and examined the neural correlate of conscious face perception. The researchers used dichoptic masking, using visible and invisible faces, houses and other objects, while imaging the brains of healthy wakeful subjects with fMRI and EEG. To resolve the neuronal process of category tuning, the researchers contrasted faces with other categories among each other. They compared these processes with those of
homogenous textures, leading to the neuronal processes underlying figure-ground segregation. The results showed that invisible and visible faces showed category-specific activity in the ventral visual cortex, whereas only visible faces caused changes in neural oscillatory
synchronization (EEG), as well as increased functional connectivity between higher and lower
visual areas (fMRI). These findings represent recurrent processes in the visual cortex during figure-ground segregation and imply that the difference in neuronal activity between category tuning and figure-ground segregation displays the difference in nonconscious and conscious information processing. These results also confirm that conscious perception is a result of a more complex neuronal integration system.
Fahrenfort et al. (2012) suggest that the difference between face-category tuning and figure-ground segregation is actually the difference between unconscious and conscious
perception. It now seems conceivable that the cause of conscious perception is the integration of neuronal structures and this provides a basis for the global neuronal workspace theory. The next study also gives an argument for the same theory by exploring Dehaene’s theoretical proposal, displaying the same neuronal integration with EEG instead of fMRI.
Gaillard et al. (2009) examined the neural difference between conscious and nonconscious information processing. They used visual pattern masking while recording intracranial EEG in epilepsy patients. Subjects were exposed to either masked words, masked blanks, visible words or visible blanks. Nonconscious processing of masked words was observed in several cortical areas, accompanied by feedforward activity. A long-lasting effect was the result of conscious processing of unmasked words, accompanied by voltage changes in the prefrontal cortex, recurrent processing and increased neural synchronization. What can be concluded is that awareness is due to long lasting synchronized activity between the prefrontal cortex and other brain areas and recurrent processing rather than feedforward activity. It implies a neural correlate of consciousness that is in alignment with the theory of a global neuronal workspace.
Profound research offers evidence for a definition of consciousness that can be seen in the brain. Prior research provides a basis for ascerting that the brain is able to process
information that we are not aware of and that conscious perception is due to a different form of information processing. Conscious perception is the result of an integrated neuronal network that has a long lasting effect and ignites in parietal and frontal areas. Given the content of
consciousness, the question arises, whether a global neuronal workspace also applies when taking a look at altered states of consciousness. To start answering these questions it is relevant to take a look at subjects in different altered states of consciousness and if evidence for a global neuronal workspace will be found when we put their brains to the test.
2. Nonconscious and Conscious processing in Altered States of Consciousness
Awareness seems to be the result of integrated brain wide information sharing in fronto-parietal regions of the brain. In a state of wakefulness the content of consciousness can differ. This is true for the state of consciousness as well. The results found in the first chapter account for the global neuronal workspace model, but only the content of consciousness mattered. It is imaginable that the underlying processes causing conscious perception only occur while being vigilant. Before detecting awareness in nonresponsive patients, it is therefore relevant to get a clear picture of the neuronal processes in different states of consciousness, when awareness is absent. In this chapter, the previously described neural correlate of consciousness, will help to answer the question what the difference is between nonconscious and conscious processing, looking at altered states of consciousness.
Prior research (Plourde et al., 2006; Heinke et al., 2004, in Davis et al., 2007) has already shown that neural activation of speech processing is impaired in patients under conditions of anesthesia. Whether this means that speech comprehension or perception is also absent
remained unclear. Although no specific speech activity has been found in a state of anesthesia, it is not that they do not show activity at all (Coleman et al., 2007, in Davis et al., 2007). Given the altered states of consciousness, does neural activity imply intact comprehension and perception or not?
Davis et al. (2007) examined whether speech processing is affected by reduced levels of consciousness. To do so, there were three levels of sedation; nonsedated, lightly sedated and
deeply sedated, induced by the anesthetic agent propofol. The participants listened to either high ambiguity sentences, low ambiguity sentences or signal-correlated noise (SCN), while fMRI was used to measure neural activity. While listening, increased auditory activity was present in all subjects, at all levels of sedation (compared to silent rest). Comparing low ambiguity
sentences with SCN showed a neural contrast in anterior and posterior regions in the superior and middle temporal gyri. Semantic comprehension was measured by contrasting high- and low ambiguity sentences. Listening to high ambiguity sentences showed increased neural activity in the inferior frontal gyrus (IFG), but only in fully awake participants. These results show that semantic processes can be impaired at relatively low levels of sedation, whereas general speech perception can still take place at all levels of deep sedation.
As to states of consciousness previous study states that we do not process information in the same way we do when we are fully awake. This is consistent with the first chapter, implying that there must be a neural process defining a difference in the content of consciousness, considering the fact that reduced and changed activity has been observed. In a state of wakefulness, awareness can be defined as a global neuronal workspace. The next experiment will take a further look into that statement.
Schrouff et al. (2011) examined whether the loss of consciousness caused by propofol is due to a reduced ability of the brain to integrate information. The participants were fully awake or deeply sedated by the anesthetic agent propofol. fMRI data were acquired and by using statistical methods, brain integration was determined. The results showed less neuronal
integration in deeply sedated subjects. The researchers also identified six brain networks during a state of wakefulness (for example motor and default mode networks) and looked at the
between network integration. When deeply sedated, the between network integration decreased significantly. There was also a decrease in regional functional connectivity between these
networks, especially between parietal and frontal regions. These findings show less neuronal integration during deep sedation in contrast to wakefulness. Decreased neuronal integration
takes place in the brain as a whole as well as between brain networks, especially in frontal and parietal brain areas.
The previous study assumes brain integration as the neural correlate of consciousness. Whether this is specifically true has to be proved by further studies, but it does account for the global neuronal workspace model in a way that the integration between parietal and frontal areas is very important for the perception of consciousness. However, it still remains unclear, whether this decreased connectivity means direct cortical action of anesthetics or actual disruption of corticothalamic-cortical connectivity. Many studies have reported a reduction of thalamic metabolism and blood flow during anesthesia (Fiset et al., 1999; Alkire and Miller, 2005; Alkire et al., 2008, in Boly et al., 2012). However, during the first few minutes of
anesthesia, thalamic activity seems to be preserved, while cortical activity is suppressed (Velly et al., 2007, in Boly et al., 2012).
Boly et al. (2012) examined whether brain connectivity changes during propofol-induced loss of consciousness. To do so, they examined the excitability of the thalamus and the amount of corticocortical backward connectivity. EEG recordings were acquired in three conditions:
wakefulness, mild sedation and clinical loss of consciousness. Sedation was induced by the anesthetic agent propofol. The results showed that the excitability of the thalamus increased during mild sedation (compared to wakefulness) and remained the same during loss of consciousness, while corticocortical backward connectivity from frontal to parietal cortices diminished during loss of consciousness. It can be concluded that propofol has a direct effect on the communication between cortical areas. Thalamic activity is preserved during altered levels of consciousness, while recurrent processes do not take place at levels of sedation. Since bilateral thalamic lesions can cause a vegetative state (Adams et al., 2000, in Boly et al., 2012), thalamic activity might be necessary, but not sufficient, for consciousness. Boly and Moran (2012) have highlighted the importance of recurrent cortical connections in the preservation of consciousness and the plausibility of a global neuronal workspace.
Another state worth looking at is sleep. Most people do not report any signs of awareness during sleep. Although we think we are not conscious, our brains are still active (Sciencedaily, n.d.). Why then does consciousness fade away? The next two studies will describe brain activity during non-REM sleep. Non-REM sleep consists of three stages. In stage 1, we produce slow eye movement and are often in the belief that we have been fully awake. In stage 2, no eye movement occurs and we are quite easily awakened. Deep sleep occurs in stage 3 and the sleeper is unaware of any stimuli (Evans et al., 2013; Massimini et al., 2005). If it is true that consciousness is related to a degradation of brain integration, the fading of consciousness during NREM sleep should be associated with a deterioration in the cortical effective connectivity of the brain, especially in stage 3.
Massimini et al. (2005) examined cortical effective connectivity during NREM sleep. They used transcranial magnetic stimulation (TMS) and high-density
electroencephalography (HD-EEG) to measure cortical effective connectivity, while
participants progressed from a state of wakefulness to a state of deep sleep. Direct cortical stimulation does not activate the reticular formation and bypasses the thalamic gate. Thus, it directly probes the ability of cortical areas to interact (Massimini et al., 2005). The
researchers looked at the transmission of information from the premotor cortex to the rest of the brain. The results showed that in fully awake subjects, the TMS response was made of recurrent waves of activity and occurred in a high frequency. With the onset of deep sleep, brain responses remarkably changed. First, EEG waves became double as large and occurred in a lower frequency. After 150 ms all TMS-evoked activity diminished. Second, during wakefulness, information transmission occurred, which involved the premotor cortex and prefrontal areas. This in contrast to deep sleep, which was related to a decrease in
intracortical information transmission. It can be concluded that during deep sleep, corticocortical effective connectivity diminishes compared to a state of wakefulness.
Whether deep sleep is not only related to a breakdown in effective connectivity, but also a breakdown in functional connectivity, has been examined in another study
(Spoormaker et al., 2010). During the experiment subjects went from a resting state to a state of light sleep, followed by a state of deep sleep, while lying in a fMRI scanner. The researchers found an increase in corticocortical connectivity during light sleep (compared to a resting state). Corticocortical connectivity, heavily decreased during deep sleep. During light sleep, functional activity occurred in a random way, which moved towards a great local clustering of neurons during deep sleep. These results imply a breakdown of functional brain
connectivity during NREM sleep.
As Massimini et al. (2005) already revealed, loss of consciousness during deep sleep is the result of a breakdown of corticocortical connectivity. Neural activity during sleep remains active, but local. This has been confirmed by Spoormaker et al. (2010) who showed that loss of consciousness is also related to a breakdown in functional connectivity.
In the previous chapter it has become clear that during anesthesia the global and local brain integration decreases, mainly in frontal and parietal brain areas. Besides, the degradation of consciousness during sleep is also due to a gradual decrease of effective and functional connectivity in the brain. On top of that, it seems that the neural pattern that is related to conscious content in fully awake people is similar to the neural pattern related to conscious state. The previous described neural correlate of consciousness does also apply for the state of consciousness, which makes the global neuronal workspace model even more likely. In the next chapter, the consensus on the neural correlate of consciousness will constitute the search for awareness in patients suffering disorders of consciousness.
3. Detecting Awareness in Nonresponsive Patients
Although it is possible to measure awareness in a state of wakefulness and in subjects who are in an altered state of consciousness, it is not as quite as difficult as in patients suffering disorders of consciousness. Most experiments that have tried to probe awareness in these patients rely on the patient’s ability to follow commands, verbally or behaviorally, as a proxy
measure for awareness (Fernández-Espejo & Owen, 2013). This requires preserved language comprehension, hearing or sustained attention, which makes the detection of awareness difficult when a patient has motor or cognitive deficits.
Therefore, one needs to know whether these patients can be aware, before discussing the global neuronal workspace theory. As described, vegetative state patients are in a state often called as ‘wakefulness without awareness’. The ‘without awareness’ part remains doubtful, because these patients would never be able to reject that statement. De Jong et al. (1997) and Menon et al. (1998) already did research on nonresponsive patients and measured ‘normal’ brain activity within their brains. In addition, other functional neuroimaging studies have suggested that ‘islands’ of preserved brain function may exist in a small percentage of patients who have been diagnosed as vegetative (Schiff et al., 2002, in Owen et al., 2006). In the following study I will describe one of the first groundbreaking results on probing awareness in a
nonresponsive patient.
Owen et al. (2006) tried to detect awareness in the vegetative state. They measured brain activity, related to spoken commands, of a 23-year old woman in a state of vegetation. First, researchers used fMRI to determine preserved speech processing by playing several sounds. They detected speech-specific activity bilaterally in the middle and superior temporal gyri (indistinguishable from healthy subjects). Now speech processing was still intact, they performed a second experiment. The patient had to imagine playing tennis or walking through the rooms of her house. The imagination of playing tennis evoked neural activation in the supplementary motor area, whereas the imagination of walking through rooms evoked neural activation in the parahippocampal gyrus, posterior parietal cortex and the lateral premotor cortex (indistinguishable from healthy subjects). These regions are specifically involved in the imagination of these behaviors; the real behavior causes the same neural activation as the imagination. These results confirmed a retained ability to understand spoken commands and respond to them. The results imply conscious awareness in this patient and therefore the possibility to detect awareness in nonresponsive patients.
Previous results provide a basis for further research. Although it seemed that the patient was still aware, this is not a certainty; beside the fact that she was the only subject, it is not sure whether she was actually aware or that the words ‘tennis’ or ‘house’ activated automatic responses. It has been known, however, that the processing of words can occur without the presence of conscious awareness (Laureys et al., 2000; Koelsch et al., 2006; Davis et al., 2007). To solve this problem, in the following study the researchers used fMRI in nonresponsive patients to generate willful answers.
Monti et al. (2010) examined the generation of willful responses in nonresponsive patients. The experiment was divided into two phases. In the localize phase, subjects were either asked to imagine playing tennis or to imagine navigating through a familiar city or home. In the communication phase, subjects performed a communication task in which they had to respond to simple ‘yes or no’ questions, by using either motor or spatial imagery respectively. In both phases fMRI was used to image brain activity. Motor and spatial imagery showed activation in the supplementary motor area and parahippocampal gyrus respectively, in 5 out of 54 patients (indistinguishable from healthy subjects). Out of all patients, one was able to use the imagery technique to answer simple ‘yes or no’ questions. This means that it is possible to generate willful answers in nonresponsive patients and the ability to detect awareness in these patients.
These results show the possibility for functional brain imaging to close the gap between behavioral shortcomings and the actual level of conscious awareness. Although previous research had a great impact on detecting awareness in the absence of behavioral responses, investigating every patient in this way is expensive and time-consuming. The fact that only one patient was able to generate willful answers perhaps means that the other patients were not aware or the method of detecting awareness is not as accurate as it can be. Compared to fMRI, EEG can detect neural changes within a millisecond timeframe and it has the ability to exam brain activity at the bedside (Faugeras et al., 2011). In the following studies I will take a further look at other methods of detection and will try to approach awareness more closely.
The mismatch negativity (MMN) is an ERP component and seems to be a good indicator of awareness. For example, the MMN evoked by non-self-related stimuli has shown to be able to predict awakening from acute coma (Fischer et al., 1999, in Qin et al., 2008). The MMN reflects pre-attentive sensory memory processes and the difference between ERP responses to standard stimuli and responses to deviant stimuli (Qin et al., 2008).
Qin et al. (2008) examined the mismatch negativity in MCS patients, VS patients and control subjects. To do so, the patients were told their own name (SON; the deviant stimuli) and tones of 800Hz (the standard stimuli), while the researchers measured their brain activity with EEG. After three months, the patients underwent the exact same procedure. Two ERP
components were identified: a N100 for both stimuli and a MMN for the subject’s own name. The N100 was present in all controls and in 9 out of 12 patients. A SON-MMN was present in all controls and in 7 out of 12 patients. After re-diagnosing the patients, 6 out of 7 (who showed a SON-MMN) changed from coma or VS to MCS. Patients who did not show a SON-MMN, did also not show clinical improvement. There was a correlation between the SON-MMN and later
clinical recovery from coma or VS to a MCS. This correlation was absent for the N100. Identifying a mismatch negativity is a way to detect awareness in patients suffering disorders of
consciousness and an indication of clinical improvement.
Although the mismatch negativity detects novel auditory stimuli, prior research also demonstrated a later neural response for novel stimuli, following the MMN, named P3b. It is related to working memory (Donchin et al., 1988; Wetter et al., 2004, in Bekinschtein et al., 2009), because of maintaining previous stimuli in working memory, regardless of long stimulus intervals. The MMN, in contrast, disappears during a short stimulus interval (Mantysalo et al., 1987, in Bekinschtein et al., 2009). Besides, the MMN is also resistant to attention in the sense that the MMN can be measured in anesthesia, coma or in response to subliminal stimuli (Fischer et al., 1999; Bekinschtein et al., 2008; Faugeras et al., 2012). The P3b, however, was only present in conscious controls when subjects were attentive and conscious of the global irregularities. These results imply that the P3b reflects a marker of conscious awareness. The following studies
use a local-global paradigm. This contains two levels of auditory regularity within trials (local level) and across trials (global level). Violation of auditory regularities at a local level should lead to the detection of the MMN, in both a conscious and nonconscious state, while violation of auditory regularities at a global level should lead to the detection of the P3b, a late and spatially distributed response (Bekinschtein et al., 2009). Besides, this is where the global neuronal workspace emerges. The local-global test is based on this theory of consciousness, which states that consciousness reflects the use of the working memory and globally integrate information to become aware of occurring deviances (Bekinschtein, Dehaene, et al., 2009; Dehaene & Naccache, 2001). Besides, the global effect should activated a brain-scale network including prefrontal, cingulate, parietal and temporal regions. In the following studies I will take a further look at this so called global effect in relation to conscious awareness.
Bekinschtein et al. (2009) examined the signature of consciousness using a local-global paradigm. To do so, they used different combinations of sounds. They created local auditory irregularities within trials (local level) and global auditory irregularities across trials (global level). To determine the full presence of awareness, investigators executed two other
experiments. In experiment 2, controls were instructed to mind wander during the paradigm. In experiment 3, controls performed a visual detection task while ignoring the sounds. Finally, a regularity awareness score (RAS) was calculated from all three experiments. Examining control subjects, the results showed a MMN at the local level and a P3b at the global level. Mind
wandering showed a decrease in the P3b effect, and in subjects performing the visual detection task there was no P3b effect at all. On top of that, fMRI showed that local violation locally activated the bilateral superior temporal gyri, including primary auditory cortices, whereas global violation activated a distributed network, including bilateral dorso-lateral prefrontal, anterior cingulate, parietal, temporal and occipital areas. Examining the patients suffering disorders of consciousness, both the VS patients and the MCS patients showed the MMN. Only the MCS patients showed a P3b effect and fully regained their consciousness in the weeks that followed. It can be concluded that less attention to (and awareness of) auditory violations
results in the decrease of a P3b. Full attention to the paradigm can lead to a P3b, which reflects a global effect and a late and spatially distributed response. Only subjects who are attentive and aware of the auditory rule and violations show a P3b effect, which implies conscious processing.
To further align these results, the following study used the same local-global paradigm, but related this to early recovery in VS patients as well. Faugeras et al. (2011) tried to probe consciousness in VS patients with ERPs. ERPs were measured during the already described local-global paradigm. The results showed a P3b effect for all controls. Only 2 out of 22 patients showed this P3b effect as well. Both of these patients recovered from a vegetative state to a minimally conscious state within a few days. These results imply a consistency in the presence of a P3b and early recovery from the vegetative state.
From the previous chapter, it has become clear that it is possible to detect awareness in patients suffering disorders of consciousness. Several methods have been used to probe
awareness; the use of spoken commands, the detection of the MMN and the detection of the P3b. The last one reflects a late and spatially distributed brain response, which requires a patient to be fully attentive and aware. Besides, the global effect showed activation of a globally distributed network including prefrontal, parietal, temporal, cingulate and occipital regions and is therefore consistent with the global neuronal workspace model. It seems that the P3b effect is a reliable measure to probe awareness in patients who are suffering from disorders of consciousness. Although detection of the P3b effect is a measure of a late and spatially distributed response, it only occurs if the subject is fully attentive and consciously aware.
Conclusion and Discussion
When it comes to detecting awareness, it seems that Bernard Baars (1989) himself was unaware of the great discoveries that would take place: “conscious experience is notoriously the great, confusing, and contentious nub of psychological science”. Prior research has been helpful to get a less confusing understanding of this so called nub. Consciousness is something we
actually can observe directly and not only in retrospect. The first chapter showed the reliability of the global neuronal workspace model of Dehaene, a theory that changed consciousness into a measurable concept, which can be observed in the brains. According to him, consciousness is due to an integrated global network of information processing, caused by a great
interconnectivity of the brain (Dehaene et al., 1998). As described, processing of words
(Dehaene et al., 2001) and faces (Kouider et al., 2009). can occur without perceptual awareness. What makes this processing conscious relies on the importance of feedback connections
resulting in late increased neuronal integration, which has its ignition in frontal and parietal brain areas (Fahrenfort et al., 2012; Gaillard et al., 2009). The first chapter has made it plausible that consciousness reflects a global neuronal workspace, which has been found in fully awake subjects. In the first chapter, the neural mechanisms underlying consciousness were related to the content of consciousness. Since the conscious state of nonresponsive patients differs from healthy awake subjects, it was important to find out, whether awareness is specifically related to conscious content or that these neural mechanisms also apply for the state of consciousness. To be able to say something about awareness in nonresponsive patients, it has been interesting to look at other states of consciousness. The second chapter has provided evidence for the global neuronal workspace model as well. As described, sedation has no effect on automatic speech processing (Davis et al., 2007). However, semantic speech processing is impaired at relatively low levels of sedation. Corticocortical backward connectivity decreases during sedation, in frontal and parietal brain areas (Schrouff et al., 2011; Boly et al., 2012). These results were confirmed by studies that looked at brain activity during sleep, which have shown that unconsciousness is a result of a breakdown in corticocortical functional (Spoormaker et al., 2010) and effective (Massimini et al., 2005) connectivity as well. This knowledge has become the link to probe consciousness in nonresponsive patients. Consciousness as a result of a highly integrated neuronal workspace does not only apply for the content of consciousness, but also for the state of consciousness. In the third chapter, studies have shown the possibility to detect awareness in these patients. Asking the patients to imaging specific behavior showed brain
activity in brain areas related to that specific behavior (Owen et al., 2006) and by using the same method, the researchers were able to generate willfull answers as well (Monti et al., 2010). By using EEG, one has been able to detect the mismatch negativity in nonresponsive patients, which reflects local auditory irregularities (Qin et al., 2008) and even distinguish patients that are attentive and aware from patients who are not (Bekinschtein et al., 2009; Faugeras et al., 2011). Attentive an aware refers to the P3b effect, which represents global neuronal integration, the use of working memory and the presence of recurrent processes in the brain, among others
involving frontal and parietal areas. On top of that, the P3b effect is an indicator of the early gain of a conscious state.
However, prior research deserves some criticism. First, research in patients suffering disorders of consciousness is still very challenging. Although it is possible to detect awareness in these patients and it has been shown that it correlates with early recovery, the amount of
patients in whom this happened is still extremely low. Before taking a certain statement, further research needs to be done on a large amount of patients to confirm the results already
described.
The research in these patients is also challenging because of physical or cognitive
limitations. The methods used for the detection of awareness could be insufficient due to certain affected abilities. Detecting ERP’s that reflect auditory irregularities for example requires preserved hearing, sustained attention and working memory.
Besides the notion that a subjective experience of consciousness can differ, one has to take the individual differences into account. Not every nonresponsive patient has the exact same brain-injury. Early recovery is thus something that not only depends on the amount of
awareness, but also on the severity and location of the injury.
The global neuronal workspace model together with EEG offers methods of bedside testing that are theoretically justified. On this basis, one is able to develop other methods of bedside testing to detect awareness in these patients. Previous results offer more precision about the neural concept of consciousness, provide a way to predict recovery of patients
suffering disorders of consciousness and will support families and doctors in their decision to withdrawal nutrition.
Bibliography
Bekinschtein, T. A., Dehaene, S., Rohaut, B., Tadel, F., Cohen, L., & Naccache, L. (2009). Neural signature of the conscious processing of auditory regularities. Proceedings of the National
Academy of Sciences, 106(5), 1672-1677.
Bekinschtein, T. A., Shalom, D. E., Forcato, C., Herrera, M., Coleman, M. R., Manes, F. F., & Sigman, M. (2009). Classical conditioning in the vegetative and minimally conscious state. Nature
neuroscience, 12(10), 1343-1349.
Boly, M., Garrido, M. I., Gosseries, O., Bruno, M. A., Boveroux, P., Schnakers, C., ... & Friston, K. (2011). Preserved feedforward but impaired top-down processes in the vegetative state. Science, 332(6031), 858-862.
Boly, M., Moran, R., Murphy, M., Boveroux, P., Bruno, M. A., Noirhomme, Q., ... & Friston, K. (2012). Connectivity changes underlying spectral EEG changes during propofol-induced loss of consciousness. The Journal of Neuroscience,32(20), 7082-7090.
Boly, M., Perlbarg, V., Marrelec, G., Schabus, M., Laureys, S., Doyon, J., ... & Benali, H. (2012). Hierarchical clustering of brain activity during human nonrapid eye movement sleep. Proceedings of the National Academy of Sciences,109(15), 5856-5861.
Coleman, M. R., Davis, M. H., Rodd, J. M., Robson, T., Ali, A., Owen, A. M., & Pickard, J. D. (2009). Towards the routine use of brain imaging to aid the clinical diagnosis of disorders of consciousness. Brain, 132(9), 2541-2552.
Davis, M. H., Coleman, M. R., Absalom, A. R., Rodd, J. M., Johnsrude, I. S., Matta, B. F., ... & Menon, D. K. (2007). Dissociating speech perception and comprehension at reduced levels of
awareness. Proceedings of the National Academy of Sciences, 104(41), 16032-16037. Dehaene, S., Kerszberg, M., & Changeux, J. P. (1998). A neuronal model of a global workspace in
effortful cognitive tasks. Proceedings of the National Academy of Sciences, 95(24), 14529- Dehaene, S., Naccache, L., Cohen, L., Le Bihan, D., Mangin, J. F., Poline, J. B., & Rivière, D. (2001).
Cerebral mechanisms of word masking and unconscious repetition priming. Nature
neuroscience, 4(7), 752-758.
Dehaene, S., & Changeux, J. P. (2011). Experimental and theoretical approaches to conscious processing. Neuron, 70(2), 200-227.
Dehaene, S., & Naccache, L. (2001). Towards a cognitive neuroscience of consciousness: basic evidence and a workspace framework. Cognition, 79(1), 1-37.
Diaz, M. T., & McCarthy, G. (2007). Unconscious word processing engages a distributed network of brain regions. Journal of Cognitive Neuroscience, 19(11), 1768-1775.
Faugeras, F., Rohaut, B., Weiss, N., Bekinschtein, T. A., Galanaud, D., Puybasset, L., ... & Naccache, L. (2011). Probing consciousness with event-related potentials in the vegetative
state. Neurology, 77(3), 264-268.
Faugeras, F., Rohaut, B., Weiss, N., Bekinschtein, T., Galanaud, D., Puybasset, L., ... & Naccache, L. (2012). Event related potentials elicited by violations of auditory regularities in patients with impaired consciousness. Neuropsychologia, 50(3), 403-418.
Fahrenfort, J., Snijders, T., Heinen, K., van Gaal, S., Scholte, S., & Lamme, V. (2013). Neuronal integration in visual cortex elevates face category tuning to conscious face
perception. Journal of Vision, 13(9), 1033-1033.
Fernández-Espejo, D., Junque, C., Vendrell, P., Bernabeu, M., Roig, T., Bargallo, N., & Mercader, J. M. (2008). Cerebral response to speech in vegetative and minimally conscious states after traumatic brain injury. Brain Injury, 22(11), 882-890.
Fernández-Espejo, D., & Owen, A. M. (2013). Detecting awareness after severe brain injury. Nature Reviews Neuroscience, 14(11), 801-809.
Fischer C, et al.(1999) Mismatch negativity and late auditory evoked potentials in comatose patients. Clin Neurophysiol 110:1601–1610.
Gaillard, R., Dehaene, S., Adam, C., Clémenceau, S., Hasboun, D., Baulac, M., ... & Naccache, L. (2009). Converging intracranial markers of conscious access. PLoS biology, 7(3), e1000061.
Kalat, J. W. (2014). Consciousness and the Brain: Deciphering How the Brain Codes our Thoughts. Journal of Undergraduate Neuroscience Education, 12(2), R5.
King, J. R., Sitt, J. D., Faugeras, F., Rohaut, B., El Karoui, I., Cohen, L., ... & Dehaene, S. (2013). Information Sharing in the Brain Indexes Consciousness in Noncommunicative Patients. Current Biology, 23(19), 1914-1919.
Kouider, S., Eger, E., Dolan, R., & Henson, R. N. (2009). Activity in face-responsive brain regions is modulated by invisible, attended faces: Evidence from masked priming. Cerebral
Cortex, 19(1), 13-23.
Lau, H., & Rosenthal, D. (2011). Empirical support for higher-order theories of conscious awareness. Trends in cognitive sciences, 15(8), 365-373.
Massimini, M., Ferrarelli, F., Huber, R., Esser, S. K., Singh, H., & Tononi, G. (2005). Breakdown of cortical effective connectivity during sleep. Science,309(5744), 2228-2232.
Monti, M. M., Vanhaudenhuyse, A., Coleman, M. R., Boly, M., Pickard, J. D., Tshibanda, L., ... & Laureys, S. (2010). Willful modulation of brain activity in disorders of consciousness. New
England Journal of Medicine, 362(7), 579-589.
Owen, A. M. (2013). Detecting consciousness: a unique role for neuroimaging. Annual review of
psychology, 64, 109-133.
Owen, A. M., Coleman, M. R., Boly, M., Davis, M. H., Laureys, S., & Pickard, J. D. (2006). Detecting awareness in the vegetative state. Science, 313(5792), 1402-1402.
Qin, P., Di, H., Liu, Y., Yu, S., Gong, Q., Duncan, N., ... & Northoff, G. (2010). Anterior cingulate activity and the self in disorders of consciousness. Human brain mapping, 31(12), 1993-2002.
Qin, P., Di, H., Yan, X., Yu, S., Yu, D., Laureys, S., & Weng, X. (2008). Mismatch negativity to the patient’s own name in chronic disorders of consciousness. Neuroscience letters, 448(1), 24-28.
Schrouff, J., Perlbarg, V., Boly, M., Marrelec, G., Boveroux, P., Vanhaudenhuyse, A., ... & Benali, H. (2011). Brain functional integration decreases during propofol-induced loss of
consciousness. Neuroimage, 57(1), 198-205.
Spoormaker, V. I., Schröter, M. S., Gleiser, P. M., Andrade, K. C., Dresler, M., Wehrle, R., ... & Czisch, M. (2010). Development of a large-scale functional brain network during human non-rapid eye movement sleep. The Journal of neuroscience, 30(34), 11379-11387.
