A functional approach to consciousness : why access explains consciousness

(1)

A Functional

Approach to

Consciousness

Why access explains consciousness

Bachelor Thesis

Psychology Brain and Cognition University of Amsterdam Author Oosterholt, P. J. Student number 10192263 Supervisor Olympia Colizoli

(2)

subject-object dichotomy. Philosophers took it for granted that a subject, the knower, is aware of another thing, the object. The knower is the (non-materialistic) mind or soul; the known object might be a material object, or in some cases another mind or our own (Russell, 1946). William James radically departed from this view and denied the existence of consciousness as an entity: “That entity is fictitious, while

thoughts in the concrete are fully real. But thoughts in the concrete are made of the same stuff as things are” (James, 1904). Consciousness is not made of magical

stuff; it is a biological phenomenon. Therefore we can study consciousness scientifically, just like molecules and atoms. Nevertheless it took almost a century before the scientific research of consciousness emerged. Behaviorism, the paradigm that dominated the first half of the century, mistrusted consciousness on the grounds that it was a subjective phenomenon. Subsequently cognitive psychology put the emphasis on other processes such as memory and perception—yet left consciousness itself untouched. Only since the last decade of the 20th_{century consciousness was back on the map. Philosophical and}

technological advances made the scientific study of consciousness possible.

Before diving in the research we need to roughly define consciousness. Consciousness can mean a numerous of things. First of all it can refer to a state.

(4)

While reading you are in a conscious state and when asleep you are not. The second meaning is being conscious of something, such as a color. This is the type of consciousness is called access consciousness. We have access to the content of our experience and we can use this content for other purposes, such as communicating. Some argue that consciousness has separate phenomenal qualities (Block, 1995). According to Block there is a difference between being able to report the color of an apple and what it is like to experience the color of an apple. Theories proposing a dissociation between access and phenomenal consciousness are called dissociative theories.

The goal of this thesis is to identify what the function of consciousness might be. In order to do so we shall analyze empirical data in the light of the access and dissociative theories. After deciding which theory is the most parsimonious with the empirical data we can explore what the function of consciousness entails. Before doing so we briefly discuss the consciousness research paradigms and discuss the idea that consciousness has no function at all. The Scientific Research of Consciousness

In the late 1980’s consciousness was back on the agenda of science. The development of new functional brain imaging techniques, such as the fMRI, was crucial for the scientific research of consciousness. Suddenly it was possible to watch what happens in the brain right at the moment we become conscious of

(5)

something. To research the contents of consciousness we must ask the subject what he is currently conscious of, we thus require introspection (Baars, 1989). Introspection as a tool for researching one’s own mental processes is clearly flawed; we generally have no access to the processes. Nevertheless the content of consciousness itself should not be doubted. Even in the case of a fata morgana we can say that the experience of phenomenon is real—however the experienced object is not. We can ask the subject what the nature of the experienced content is (i.e. did you see the word, hear a sound et cetera). Introspection therefore returns to the field of science, but now it is in the form of raw data (Dehaene, 2014).

An important step is to develop a paradigm in which we can investigate with scrutiny. Baars (1989) had just the right idea, in order to research consciousness, experimental situations should have minimal contrast with each other. Ideally the only difference between the two is that in only one the stimulus is consciously perceived. Differences in the brain then reflect conscious processing. Binocularly rivalry is a great example of minimal contrast. Normally our eyes receive roughly the same input. However when one eye is presented a different image our brain has to resolve this ambiguity. Instead of imposing a fusion between the two images, we see only one. The perception switches randomly between the two images. Here lies a unique way to research consciousness. The stimuli are exactly the same; the only difference is the conscious percept. Another useful paradigm is masking (Dehaene, 2014). When

(6)

flash a word for a short period of time, e.g. 30ms, the word is still perceived. However when we quickly substitute the first image with another, the mask, then the first image remains unconscious. Now the changes this image brings upon the brain can be researched. This enables us to research what kind of functions can operate without consciousness and which cannot. With these techniques we are able to delineate the borders of consciousness and explore its functions. Before going down that road we should entertain the thought that consciousness has no function—can we make a convincing argument for this stance?

Consciousness: An Unnecessary Condition?

Can it possibly be that every single fleeting moment of our conscious existence does not actually play a role in our behavior? Two sorts of arguments are made along these lines: conscious inessentialism and epiphenomenalism (Seth, 2008).

“For any intelligent activity i, performed in cognitive domain d, even if we do i with conscious accompaniments, i can in principle be done without these conscious accompaniments” (Flanagan, 1992). Inessentialism does not necessarily

imply that consciousness has no function; it merely states that in principle the intelligent activity could be done without being conscious. Conscious inessentialism is most likely true for some activities. Many activities can be done without consciousness, such as breathing and walking. Furthermore we can imagine that a computer can execute intelligent activities, such as solving a

(7)

mathematic equation—even though that we are incapable to do so without the help of consciousness. Most problems have multiple solutions and some of our cognitive activities are no exception to this rule. However, this is no a priori evidence that consciousness has no function.

Epiphenomenalism is the view that consciousness has no casual role in behavior. The most famous proponent of this view is Chalmers. Chalmers (1996) argues that it is possible to conceive a world wherein all people are zombies, lacking conscious experience entirely, yet still acting exactly the same. These zombies are biologically identical to humans. Since Chalmers deems these zombies conceivable (in a philosophical sense, not a practical), he concludes that the phenomenological aspect of consciousness must be an epiphenomenon. Chalmers still believes that phenomenal consciousness exists, however not as a part of the physical world. Chalmers therefore takes a (neo)dualistic stance. Most people do not agree with this radical conclusion. Nevertheless this thought experiment illustrates a problem. Chalmers himself differentiates between easy

and hard problems of consciousness. The easy problems are explaining

phenomena such as categorization of visual stimuli. The hard problem is explaining how a bunch of neurons can create qualia; the subjective experience of what it is like to experience something.

There is considerable amount of evidence that at least some aspects of consciousness are indeed epiphenomena. Libet (1983) found that unconscious

(8)

brain activity preceded the conscious intention to flick their wrist. This activity takes place about half a second before their conscious intention to move. The unconscious brain activity was found to be predictive of the actual behavior (Soon, 2008). This begs the question if consciousness plays a role in initiating motor actions and argues for epiphenomenalism stance on consciousness.

Another relevant phenomenon is blindsight, the ability to “see” without consciously seeing. Individuals with blindsight have a lesion in the primary visual cortex and have no conscious perception. When asked to discriminate between two objects in a forced-choice paradigm the patients perform above chance, which indicates that at least some form of perception is possible without consciousness (Weiskrantz, 1986). Again this is an argument for epiphenomenalism. However we should realize that the abilities of people with blindsight are seriously affected (even though monkeys with blindsight seem to be less affected). Only confronted with a forced-choice paradigm the patients show above chance performance. The lack of consciousness clearly deteriorates our ability to process the environment. People with blindsight are far from the zombies Chalmers envisioned.

Although inessentialism and epiphenomenalism illustrate the difficult nature of consciousness we should not accept their stance a priori. Nevertheless, it is a good idea to remain open for more than one possibility: a conscious processes can be inessential, an epiphenomenon or functional. The only way to settle the

(9)

debate is to delineate the borders between the unconscious and consciousness: which processes can function without consciousness and which cannot.

Unconscious Processing

The function of consciousness is intricately tied with unconscious processing. Which operations can the unconscious mind perform on its own—and which only with the help of consciousness? Understanding the strength and limitations of unconscious processing helps us to determine the function of consciousness. A large part of the processes in the brain remain unconscious. Every physiologic process is unconsciously regulated; all our learned motor skills are flawlessly executed without the help of consciousness, and our habits take place without conscious effort. It would indeed be terribly strenuous if we have to consciously regulate all these processes. Unconscious processing even extends to operations that we usually associate with consciousness. For example it has been found that we are able to detect errors unconsciously (Bechara et al., 1997). Furthermore we can unconsciously read words and numbers (Naccache & Dehaene, 2001); understand the meaning of words (van Gaal et al., 2014), and solve complex problems (Dijksterhuis et al. 2006). We are even able to overrule habitual responses without the help of consciousness (van Gaal et al., 2010).

Unconscious processes can be viewed as a massive set of specialized processors. The problem however is that these specialized processors are in a way

(10)

blind for other processors. The specialized processors only receive input from processors in a lower hierarchical level, and usually do not communicate reciprocally. This should be no surprise. Clearly it would highly inefficient if for example the visual cortex would have direct connection to every area that makes use of visual information. As William James remarked: The study a posteriori of the

distribution of consciousness shows it to be exactly such as we might expect in an organ added for the sake of steering a nervous system grown too complex to regulate itself (1890, p.89). In order to steer, communicate and control we need to

architecture that supports these operations (Baars, 2005). The Global Workspace Theory might just be the solution.

Global Workspace Theory

Baars (1988) developed the Global Workspace Theory (GWT), which states that, the function of consciousness is making information globally available to the brain. The GWT can be likened to a theater. In this metaphor consciousness represents the bright spot, and attention the spotlight itself (Baars, 2005). In the dark sits the unconscious carefully listening to the broadcasted content. The global workspace allows unconscious processors to interact with each other. Furthermore the unconscious processors compete (or cooperate) with each other for access to the global workspace. Once the information gained access other unconscious processors can use it. The primary role of consciousness is thus integrating,

(11)

provide accessing, and coordinating of the function of unconscious processors that otherwise would operate autonomous (Baars, 2005). Following the GWT we can say that whatever is conscious can be reported. To clarify, any means of reporting would suffice.Monti et al. (2010) for example were able to communicate with patient in a coma with the use of a fMRI.

Dehaene et al. (1998) proposed a brain interpretation of the GWT, called the global neuronal workspace theory. They have accomplished this by modeling the theory in formal neuronal network models, and by using brain imaging techniques to acquire evidence that shows that global broadcasting only occurs when one is conscious (Baars, 2011). Dehaene (2014) showed that there are four signatures present when a stimulus is consciously perceived and absent when not. He emphasizes that these are not mere correlates. The signatures are casually related to consciousness—correlates are obviously not per definition casually related to each other. The signatures are (1) the sudden ignition of activity in the parietal and prefrontal circuits; (2) a late slow wave called the P3; (3) a late burse of high-frequency oscillations; and (4) synchronization of information exchanges across distant brain regions. The signatures shall now be discussed in detail.

Dehaene et al. (2001) performed a fMRI and ERP study on the perception of masked words and compared conscious to unconscious trials. Whenever the flashed word was followed by a mask and thus not perceived, activity in the primary visual cortex was detectable, including in the left extrastriate, fusiform and

(12)

precentral areas. However the activation lost its strength and did not reach parietal and frontal areas. Conscious trials showed a different image. The activity did not extinguish but was amplified and caused activation in a number of areas, including the parietal, prefrontal and cingulate cortex. The ignition of activity in the parietal and prefrontal cortex is the first signature of consciousness and can be likened to the gaining access part of the GWT. The activity in the visual cortex was strong enough to gain access and subsequently made its contents available to distant parts of the brain.

This ignition is by no means a gradual process. Using the attentional blink paradigm Sergent, Baillet, and Dehaene (2005) manipulated the parameters until the subjects consciously perceived the stimulus on half of the trials. Initially both conscious and unconscious words showed the same activity. The first waves in the posterior part of the visual cortex, the positive wave P1 at 100ms, and negative wave N1 at 170ms, were identical for both unconscious and conscious words. This is considered the feedforward sweep (Lamme, 2003); the stimulus is processed through a hierarchy of visual areas. Between 200 and 300ms after the onset of the stimulus something interesting happened: the brain activity in unconscious trials died out, however in conscious trials it rapidly grew stronger. The ignition of activity seen in conscious trials usually starts from 270ms and peaks between 350-500ms (depending on the area). This is called the P3 wave and this is the second

(13)

the P3 wave represents the point where the content gains access to the fronto-parietal network. Fascinatedly this implies that our consciousness is very slow. It takes at least 1/3 of a second to become conscious of something. Which would explain why conscious control of motor actions seems to be impossible as seen in the Libet studies. Evolutionary it would be disastrous to depend on such a slow mechanism. The unconscious mind is however quick and has emergency plan ready, either learned or innate.

The third signature of consciousness is a late burse of oscillations. Fish et al.

(2009) studied the signatures with intracranial recordings and backward masking and found similar results as previously mentioned. Furthermore Fish et al. found a massive increase in gamma-band fluctuations after 300ms in the 30-70 Hz range. This synchronization is hypothesized to facilitate information transmission and is therefore crucial for the global workspace.

Finally the fourth signature of consciousness is long distance synchrony. Gaillard et al. (2009) used the same paradigm as Fish et al. and found another crucial difference 300 to 500ms after the stimulus onset. The unmasked stimuli were associated with an increase in phase synchrony between distant pairs of electrodes, i.e. areas, in the beta range (13-30 Hz), which was absent when the stimuli were masked and thus not consciously perceived. The casual relations between the distant pairs of electrodes were found to be significant and

(14)

bidirectional. This is hypothesized to enable the long distant interaction between distant regions.

The studies conducted by Dehaene and others convincingly show that consciousness is at the very least intractably tied to the signatures. The power of the signatures is that the processing before 300ms is exactly the same. Only afterwards the unconscious and conscious trials start to divert. Be that as it may, these signatures are rather crude indications of consciousness. Considerable refinement is thus necessary. These studies investigate the entering of one percept in the consciousness—but when awake we are always conscious of something. From an introspective perspective we can say that consciousness is more like a stream than individual percepts that follow each other consecutively. However this could be an illusion. Although the isolation is necessary it makes the problem considerable simpler than it actually is. Whether or not the signatures are truly casually related to consciousness remains an open question. Perhaps if we can develop paradigms to measure consciousness in animals we can manipulate the signatures to see what happens to the conscious percept. Unfortunately we cannot ask animals what they are experiencing. We cannot us the signatures as the definition of consciousness since that we be circular reasoning.

The neuronal GWT needs considerable refinement before we can begin to speak about a complete theory of consciousness. Nevertheless I am inclined to think that the GWT is a (huge) step in the right direction. However, have we

(15)

explained the phenomenal aspect of consciousness? Dissociative theories argue that GWT ignore the phenomenal aspect of consciousness.

Dissociative Theories

When we watch the world we experience a rich and detailed world. However if a thing captures our attention, we might become completely oblivious of other occurrences, even moonwalking gorillas. Simons and Chabris (1999) showed subjects a video of a group of basketball players passing the ball. Subjects were told to keep the count of the total amount of passes, or separate counts of bounce and aerial passes (which is a considerable harder task). During the passing game a gorilla walked through the scene and later on subjects were asked whether they noticed something unusual. Surprisingly a great deal of the subjects did not notice the gorilla (37.5% in the easy condition; 46% in the hard condition). How is it possible that the subjects missed a gorilla? The subject’s top-down attention was engaged by a demanding task. Remember, the GWT states that top-down attention is a necessary condition for conscious percept. Without the ability to use our top-down attention we are unable to become conscious of something. This is the reason subjects missed the gorilla. However, when the stimulus is significantly salient bottom-up attention processes the stimulus thoroughly and enables top-down attention, hence we still perceive the stimulus. Thus we would not miss a pink gorilla in black and white scene. This seems all the more evidence for the

(16)

GWT—that states that only attended objects can be accessed consciously. The gorilla is not perceived because our top-down attention is engaged by another task.

Surprisingly some people think we consciously saw the gorilla. We just cannot access that piece of information. According to Lamme and Block we still experience the phenomenological aspect of the scene; we however do not posses the knowledge of the experience, which is dependent on the access of information. We see but we do not know that we see. Introspection is the limiting aspect: by the time we have to report the experience we already forgot most of the content. As previously mentioned Block makes the distinction between access and phenomenal consciousness (1995). A phenomenal conscious state entails the sensation of the content. On the other hand the access state is propositional attitude about the content in the form of thoughts, believes, desires et cetera. Our conscious perceptual system overflows, i.e. has a higher capacity, than the system that cognitively accesses the percept (Block, 2011). Phenomenal consciousness comes into existence by a process called recurrent processing.

Lamme and colleagues laid the groundwork for understanding recurrent processing with their work on the visual system. Recurrent processing refers to repeated processing of the same visual input by a neuron. Initially an image causes successive activation in the hierarchical levels of the visual cortex through feedforward connections, called the feedforward sweep (Lamme & Roelfsema,

(17)

2000). After the feedforward sweep the neurons remain active. Information from both horizontal and feedback connections are incorporated into the responses. Individual neuron initially tune for a single aspect, for example whether the stimulus is a face or not. However, after a delay, there is a change in tuning in the neurons, and the neurons start to convey fine information, such as the nature of the facial expression (Sugase et al., 1999). Furthermore neurons start to respond to contextual information outside its receptive field (Lamme & Roelfsema, 2000). Transcranial magnetic stimulation (TMS) studies supply important evidence that recurrent processing is a necessary condition for conscious perception. The first part of the feedforward sweep is completed at about 100ms after the presentation of the stimulus. When pulses are applied over the early visual cortex 30ms until 50ms after the onset of the stimulus perception is blocked (Corthout et al., 1999). This is no surprise; the feedforward sweep cannot be finished so a disruption will surely disrupt the processing. However this disruption also takes place when pulses are applied between 80 and 120ms after the onset on the stimulus (Walsh & Cowey, 1998). The feedforward sweep has moved on the higher visual areas—yet visual perception is still absent. Pulses therefore disrupt the recurrent processing, and not the feedforward sweep. This is a clear indication that recurrent processing is a necessary condition for conscious perception.

How can you prove the subjects saw the gorilla but failed to report? Dissociative theorists have been using Sperling’s (1960) partial report paradigm as

(18)

evidence. In these studies the subjects looked at a screen with three rows of four letters each; when the screen disappeared the subjects were asked to report as many letters as possible. Even though the subjects thought they could see all the letters they could only reliably report 3 to 4 letters. However, when a cue after the stimulus offset indicated which row the subjects had to report the same number was reported from that row alone. This suggests that the subjects had a persistent image of all the letters. This type of memory has been dubbed as iconic memory. Lamme and colleagues combined the partial report with the change blindness paradigm (Lamme, 2010). Change blindness is failing to notice a change in the environment—even as big as an engine of the airplane. In the modification of Lamme et al. an array of objects is presented followed by a blank interval, which again is followed by an array of objects. The new array may include a change in one of the objects. When a cue after the reappearance of the objects asked the subjects whether a change has occurred at that location, the subject’s performance was low. However, when the cue was presented at the interval, the performance was nearly as good as presenting the cue during the first array of objects. The more objects presented the more remembered (Sligte et al., 2008). The iconic memory of an image has no phenomenality itself. Nevertheless Lamme and Block argue that this is the closest evidence of what we truly see when we are looking at the image. Obviously this is no direct evidence that we see more than we know. It has been argued that thinking that we see all the objects is an illusion.

(19)

Dehaene (2006) calls this the refrigerator light illusion; when you open the door the light is always there, but in reality, when the door is closed, it’s not. So what is it—do we have separate phenomenal experiences or is consciousness explained by access?

Settling the Debate

Is recurrent processing alone enough for consciousness—or are the other signatures of consciousness necessary? This question has proven to be quite difficult to answer. I shall introduce the arguments and subsequently supply my own opinion about the matter. To clarify recurrent processing is certainly necessary for consciousness—but is it sufficient?

The whole argument that consciousness can exist without access seems to be based on the notion of the richness of our visual experience (Block, 2011; Lamme, 2010). When we see an array of objects we get the impression that we see all the objects. However when we avert our gaze we seem to forget the majority of the content. To quote Lamme: What is evident from this small exercise is that

introspection is a poor guide to conscious visual sensation (Lamme, 2010). Yet the

whole claim to the richness of the visual sensation rests on exactly the same method: the act of introspection. Why should we trust our introspection when judging whether our visual sensation is rich—but not when reporting the contents?

(20)

We simply cannot know that we see more than we know, otherwise we would contradict ourselves.

To be fair, there are more arguments in favor of the dissociative theories. Lamme uses his neural arguments to make a case for phenomenal consciousness. The feedforward sweep is known to affect distant areas and subsequently behavior even though conscious percept is absent. If the feedforward sweep can affect the areas involved in the global workspace, such as the prefrontal cortex, then the convergence towards those areas is not sufficient for conscious sensation to arise (Lamme, 2010). The only difference between the feedforward processing stages and the conscious access of the percept is recurrent processing. This however is a misrepresentation of the GWT. It is indeed true, beyond doubt, that unconscious perception of visual stimuli can influence other areas in the brain, even the prefrontal cortex. These connections are functional, i.e. they can bring about behavior, such as inhibition of motor response (van Gaal et al., 2010). However the fact that certain areas that participate in the global workspace are (functionally) active but do not produce consciousness is no evidence for that recurrent processing is the only difference between feedforward processing and access consciousness. For example synchronized long-distance oscillations are necessary for the global workspace. Only if these Lamme can prove that the activity in the fronto-parietal is identical in both conscious and unconscious perception then we are allowed to conclude that the only difference indeed is recurrent processing.

(21)

Lamme continues with the ontological issues of the GWT. In the dissociative theory consciousness is orthogonal from other cognitive functions (Lamme, 2010). Consciousness and attention can exist independently. In the GWT this is not the case. The two are highly confounded in the GWT—according to Lamme. This grants the dissociative theory a better ontological status than the ontological flawed GWT. Science should always demarcate concepts as clearly as possible (Lamme, 2010). Hence recurrent processing is the more scientific explanation of the data. It seems to me that Lamme is stating a normative rather than descriptive account of what a theory of consciousness should be. Merely pointing to the fact that two concepts are orthogonal is not a support for a theory, and there is nothing “ontological” wrong with a theory that states that one process can only occur on the basis of another. Surely we need to explain how attention makes conscious perception possible. Attention and consciousness might seem confounded at this time but untwining the two will resolve the problem. I find this no reason whatsoever to disqualify the theory.

In fact, it seems that the dissociative theory is unscientific. By separating phenomenality from access consciousness we are absent of a functional explanation. To be clear, the GWT states that phenomenal consciousness is the same as access consciousness. In this case phenomenality has a clear function. However in the dissociative theory phenomenal consciousness has no function. We could easily imagine a situation where the absence of phenomenality will produce

(22)

exactly the same behavior. Let us assume that the subjects of Sperling’s studies were indeed unconscious of the majority of the letters, however the visual cortex did process the objects to some extent. We have mental representation of all the objects, but not all the features are extracted. Only the consciously perceived objects are fully organized. After the cue the subjects can elaborate on their mental representation by using top-down attention. The objects in the iconic memory are now organized and accessed. This will result in exactly the same behavior: the correct report of the letters. The iconic representation of the objects is thus enough to explain Sperling’s findings. We have no reason to grant phenomenality to an earlier stage of processing (recurrent processing).

This puts us in a rather awkward position. A theory that does not provide a function of the explained phenomena is not a scientific theory. Cohen and Dennett (2011) illustrate this matter by a thought experiment they call the “the perfect experiment”. Cohen and Dennett sketch a future were scientist are able to isolate the V4 area that is responsible for color perception. The isolation refrains the color area from projecting to higher areas but leaves everything else intact. If we then present the subject with a red apple what will he say? He surely won’t be able to report the color of the apple, since there is no access to this information (we can assume everybody agrees on this part). Other features, such as form, are extracted since the areas responsible are still intact. If we identified the pattern of activation beforehand and see the same kind of activation we now know the redness of the

(23)

apple is processed. However the subject will—no matter how hard he tries—not report the color of the apple. Not only that, other functions such as an affective judgment will fail, since even those functions require the information that pertains the redness of the apple. Despite the absent of any kind of behavioral reaction of the color red, dissociative theorists still have to assume that the subject is indeed conscious of the color. After all, phenomenal consciousness does not require the connections to other parts of the brain. Here exactly lies the trouble of the dissociative theory. Even a perfect experiment cannot falsify the theory (Cohen & Dennett, 2011). All behavioral responses are gone—yet the proponents of the dissociative view still have to claim that there is phenomenal consciousness.

Separating consciousness from other cognitive functions, which according to Lamme (2010) should be the goal of a science of consciousness, is exactly the opposite route we should pursue. Consciousness should be explained in the grand scheme of cognition. Without other cognitions we simply have no way to test the hypothesis. Dissociative theories only seem to account for unconscious visual processing.

To present date no study has been conducted—at least to my knowledge— that falsifies the GWT. Some studies claimed that there is a double dissociation between top-down attention and consciousness—meaning that there is consciousness without top-down attention (these studies are discussed in later on for other purposes). However, even the proponents of dissociative theories admit

(24)

that top-down attention is not fully absent in these studies. Which is problematic for both theories, since we cannot fully control top-down attention at this point. Furthermore there is no evidence that the four signatures of consciousness are not casually related with conscious access. For example, a disruption of the dorsal lateral prefrontal cortex—which is part of the fronto-parietal network—by TMS bilateral theta-bursts results in degraded visual awareness of stimuli (Rounis, 2010). If the areas associated with the fronto-parietal network would be knocked out and consciousness would still be present then the GWT would be unsound. However this is not the case. Considering the empirical data and the philosophical arguments I am therefore inclined to back GWT as the most parsimonious account of consciousness. If GWT is true, what can we say about the functions of consciousness?

Flexible Consciousness

Separating access and phenomenal consciousness from each other seems unfounded. With the help of GWT we can now define the functions of consciousness. In order to do so we need to dissociate between top-down and bottom-up attention. Both processes can support functional behavior. Bottom-up processing works via the feedforward sweep and is thus able to react swiftly. This can be shown in dual-task studies, where the subject’s attention is drawn to an attentional demanding task, while in the periphery a second stimulus is flashed.

(25)

The subjects then have to determine the identity of the stimulus discriminate (Koch, 2006). Even though the subjects are using their top-down attention on the main task we cannot assume that the top-down attention is fully disengaged from the periphery. However we can say that there is little top-down amplification. Surprisingly subjects were very adept at discriminating between animals and vehicles (Li et al., 2002); famous/non-famous faces and gender (Reddy et al., 2004; 2006). However the subjects were unable to discriminate between a rotated L and T (Koch, 2006). This seems like an indication that bottom-up processing can perceptually organize a range of stimuli without much help of top-down attention. Recognizing nature scenes and the identity of faces are evolutionary adaptive, therefore we can assume that there are specific feedforward connections for these discrimination. However reading is a relative new skill we possess and is not innate. Therefore only some of the perceptual organizing can happen with little top-down attention. Bottom-up processing is swift and useful to identify and act upon our environment. Kirchner and Thorpe (2006) showed that subjects were able to discriminate between images with and without animals in 120ms by the use of forced-choice saccades (eye movements). Evidently the subjects were not conscious of the content, and top-down attention was absent since the response was only after 120ms. From these studies we can conclude that bottom-up attention is functional for the sensory-motor behavior. However the behavior is automatic and not flexible.

(26)

Consciousness does not suffer this problem. When top-down attention is directed to the input the processing becomes deeper and in most cases (but not in blindsight) the content accesses the fronto-parietal network thus making the content conscious. Multiple brain areas with different functions can use the information in the global workspace. Not surprisingly the global workspace is tightly related to working memory. We can keep content online if we wish to elaborate on it. Imagine for example the exercise of reading. If you would forget the content as soon as it enters the consciousness, then a sentence would make no sense. In order to understand our environment we need to infer relationships between the elements. This can only done by keeping information online in order to relate it to new incoming information. Learning the relationship between a cue and stimulus when the stimulus does not immediately follow the cue is impossible without consciousness (Clark & Squire, 1998). Novel behaviors can be learned by sequential executing the steps for that that behavior. The beauty of this system is that once practiced enough consciousness is not necessary anymore and the behavior is executed flawless without top-down attention. Novel problems are also ideally suited for consciousness. Unconscious processing can bias a solution as can seen the study of Dijksterhuis et al. (2006), where subjects who unconsciously deliberated about a problem made a better decision than subjects who deliberated consciously. The unconscious mind is a great statistician—it can capitalize on repeated exposure in order to infer regularities. However this does

(27)

not result in a crystal clear decision. Forming a strategy—a decision to do the same thing when something happens—can only happen when we consciously deliberate. In the Bechara et al. (1997) study the chance that subjects unconsciously picked out the correct response in a gambling task started to improve after some time. However, only when subjects became conscious of the best strategy, the subjects made the correct decision all of the time. At this point all the areas now what the correct decision is by virtue of the global workspace. Only when we are conscious we can reason with logic, remember knowledge, represent and manipulate information, and plan our actions. The availability of content to many processors exponentially increases the possibilities and complexity of our behavior. Consciousness is a highly adaptive solution to the inflexibility of the unconscious mind.

Demystifying the Hard Problem of Consciousness

The GWT argues that the global availability of information is what we subjectively experience as consciousness (Gaillard et al., 2009). Thus the phenomenological aspect we all know so well is merely the global availability of that content. Have we not explained away the hard problem of consciousness? I have on purpose avoided the so-called hard problem of consciousness in this thesis. Is the question solvable or as Chalmers puts it beyond the reach of science?

(28)

In order to answer the question we need to return to the philosophical zombies. Philosophical zombies are biological identical to us and yet lack the phenomenal aspect of consciousness. Even if some brilliant scientists have the ingenuity to make such a zombie they will face an even harder problem. That is, to prove that the zombies are not conscious. The zombies will pass every test that men can throw at them. Why is that? Well obviously the zombies have to have the functional complexity we have—otherwise they will not produce the same behavior. The zombies are indistinguishable from humans exactly because they are functionally the same as we are. Just as us their fronto-parietal networks will grant access to the thing they perceive. Subsequently they can base further behavior on the processing of the accessed content. However when we assume they have the same functional complexity they too will have qualia. There is no way around it. The philosophical zombie cannot both be functionally the same and not be conscious. Stripping away the phenomenal aspect is impossible. The contrast between the imagined zombies and conscious beings is an illusionary one (Dennett, 1995). The hard problem of consciousness is not beyond the reach of science. If we understand how conscious access works, the phenomenal aspect of conscious will fall right in its place (Dehaene, 2014).

(29)

The raison d’être is making information globally available for the use of a wide range of processors. We have examined the neuronal signatures of the GWT, which seem to be a reliable indication of consciousness. Furthermore we have discounted the dissociative theories of consciousness. Separating consciousness from function is unfounded, impossible and above all unscientific. Making information globally available gives rises to new flexible behavior, which would have been impossible to produce by the unconscious mind. Access consciousness accounts for the phenomenal aspect of consciousness. The philosophical zombie without phenomenal experience is thus impossible. The hard problem of consciousness is a non-existent one.

Nevertheless, to say that we have explained consciousness in detail would be downright a lie. There is still immense amount of work to be done to finish the project. For example we need to have better understanding what the unique properties of the fronto-parietal networks are and how conscious percepts are encoded in those networks (Cohen & Dennett, 2011). Furthermore the understanding how neural synchrony contributes to consciousness and how it exactly relates to (working)memory and decisions needs to be improved (Dehaene, 2014). Moreover we are in need of new paradigms, such as a paradigm that can dissociate top-down attention from consciousness and a paradigm that can investigate animal consciousness. At the least we have built a solid foundation to

(30)

start discovering the fine points of consciousness. The upcoming years will truly be very interesting for the science of consciousness.

Literature

Baars, B. (1988). A cognitive theory of consciousness. New York, NY: Cambridge University Press.

Baars, B. J. (2005). Global workspace theory of consciousness: Toward a cognitive neuroscience of human experience. Progress in Brain Research, 150, 45-53.

Bechara, A., Damasio, H., Tranel, D., & Damasio, A. R. (1997). Deciding advantageously before knowing the advantageous strategy. Science, 275(5304), 1293-1295.

Block, N. (1995). On a confusion about a function of consciousness. Brain and

Behavioral Sciences,18(2), 227-247.

Block, N. (2011). Perceptual consciousness overflows cognitive access. Trends in

Cognitive Sciences, 15(12), 567-575.

(31)

Clark, R. E., & Squire, L. R. (1998). Classical conditioning and brain systems: the role of awareness. Science, 280(5360), 77-81.

Cohen, M. A., & Dennett, D. C. (2011). Consciousness cannot be separated from function. Trends in Cognitive Sciences, 15(8), 358-364.

Corthout, E., Uttl, B., Walsh, V., Hallett, M., & Cowey, A. (1999). Timing of activity in early visual cortex as revealed by transcranial magnetic stimulation.

Neuroreport, 10(12), 2631-2634.

Dehaene, S. (2014). Consciousness and the brain: Deciphering how the brain codes our thoughts. New York, NY: Penguin Group.

Dehaene, S., Changeux, J. P., Naccache, L., Sackur, J., & Sergent, C. (2006). Conscious, preconscious, and subliminal processing: a testable taxonomy. Trends in

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

(32)

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.

Dennett, D. C. (1995). The unimagined preposterousness of zombies. Retrieved from http://ase.tufts.edu/cogstud/dennett/papers/unzombie.htm.

Dijksterhuis, A., Bos, M. W., Nordgren, L. F., & Van Baaren, R. B. (2006). On making the right choice: The deliberation-without-attention effect. Science, 311(5763), 1005-1007.

Fisch, L., Privman, E., Ramot, M., Harel, M., Nir, Y., Kipervasser, S., Andelman, F., Neufeld, M. Y., Kramer, U., Fried, I. & Malach, R. (2009). Neural “ignition”: enhanced activation linked to perceptual awareness in human ventral stream visual cortex. Neuron, 64(4), 562-574.

Flanagan, O. J. (1992). Consciousness reconsidered. Cambridge, MA: MIT press.

Gaillard, R., Dehaene, S., Adam, C., Clémenceau, S., Hasboun, D., Baulac, M., Cohen, L. & Naccache, L. (2009). Converging intracranial markers of conscious

(33)

James, W. (1890). The principles of psychology. New York, NY: Holt and company.

James, W. (1904). Does consciousness exist? The Journal of Philosophy,

Psychology and Scientific Methods, 1(18), 477-491.

Kirchner, H., & Thorpe, S. J. (2006). Ultra-rapid object detection with saccadic eye movements: Visual processing speed revisited. Vision Research, 46(11), 1762-1776.

Lamme, V. A. (2003). Why visual attention and awareness are different. Trends in

Lamme, V. A. (2006). Towards a true neural stance on consciousness. Trends in

Lamme, V. A. (2010). How neuroscience will change our view on consciousness. Cognitive Neuroscience, 1(3), 204-220.

Lamme, V. A., & Roelfsema, P. R. (2000). The distinct modes of vision offered by

(34)

Li, F. F., VanRullen, R., Koch, C., & Perona, P. (2002). Rapid natural scene categorization in the near absence of attention. Proceedings of the National Academy of

Sciences, 99(14), 9596-9601.

Libet, B., Gleason, C. A., Wright, E. W., & Pearl, D. K. (1983). Time of conscious

intention to act in relation to onset of cerebral activity (readiness-potential): The unconscious initiation of a freely voluntary act. Brain, 106(3), 623-642.

Monti, M. M., Vanhaudenhuyse, A., Coleman, M. R., Boly, M., Pickard, J. D., Tshibanda, L., Owen, A. M., & Laureys, S. (2010). Willful modulation of brain activity in

disorders of consciousness. New England Journal of Medicine, 362(7), 579-589.

Naccache, L., & Dehaene, S. (2001). Unconscious semantic priming extends to novel unseen stimuli. Cognition, 80(3), 215-229.

Reddy, L., Reddy, L., & Koch, C. (2006). Face identification in the near-absence of focal attention. Vision Research, 46(15), 2336-2343.

Reddy, L., Wilken, P., & Koch, C. (2004). Face-gender discrimination is possible in the near-absence of attention. Journal of Vision, 4(2), 4.

(35)

Rounis, E., Maniscalco, B., Rothwell, J. C., Passingham, R. E., & Lau, H. (2010). Theta-burst transcranial magnetic stimulation to the prefrontal cortex impairs

metacognitive visual awareness. Cognitive Neuroscience, 1(3), 165-175.

Russell, B. (1945). History of Western Philosophy. London: George Allen & Unwin Ltd.

Sergent, C., Baillet, S., & Dehaene, S. (2005). Timing of the brain events underlying access to consciousness during the attentional blink. Nature Neuroscience, 8(10), 1391-1400.

Seth, A. K. (2009). Functions of consciousness. Encyclopedia of Consciousness, 1, 279-293.

Simons, D. J., & Chabris, C. F. (1999). Gorillas in our midst: Sustained inattentional blindness for dynamic events. Perception-London, 28(9), 1059-1074.

Sligte, I. G., Scholte, H. S., & Lamme, V. A. (2008). Are there multiple visual short-term memory stores? PLOS One, 3(2), e1699.

(36)

Soon, C. S., Brass, M., Heinze, H. J., & Haynes, J. D. (2008). Unconscious determinants of free decisions in the human brain. Nature Neuroscience, 11(5), 543-545.

Sperling, G. (1960). The information available in brief visual presentations. Psychological

Monographs: General and Applied, 74(11), 1.

Sugase, Y., Yamane, S., Ueno, S., & Kawano, K. (1999). Global and fine information coded by single neurons in the temporal visual cortex. Nature, 400(6747), 869-873.

van Gaal, S., Naccache, L., Meuwese, J. D., van Loon, A. M., Leighton, A. H., Cohen, L., & Dehaene, S. (2014). Can the meaning of multiple words be integrated

unconsciously? Philosophical Transactions of the Royal Society B: Biological

Sciences, 369(1641), 20130212.

van Gaal, S., Ridderinkhof, K. R., Scholte, H. S., & Lamme, V. A. (2010). Unconscious activation of the prefrontal no-go network. The Journal of Neuroscience, 30(11), 4143-4150.

Walsh, V., & Cowey, A. (1998). Magnetic stimulation studies of visual cognition. Trends

(37)

Weiskrantz, L. (1986). Blindsight: A case study and implications. Oxford: Oxford University Press.

A functional approach to consciousness : why access explains consciousness

A Functional

Approach to

Consciousness

Why access explains consciousness

Table of Contents