Saturday 26 April 2008

Neurofunkers

In the last posts in this mini-series on Consciousness we were philosogeeks, the one sitting on the edge of the party worrying about things like whether girls actually exist, whether human minds can ever understand girls and how the hell to chat them up.

We are now going to be neurofunkers and just assume that girls exist and we can understand them. Neurofunkers aren't practising their chat up lines by themselves in the corner; they're out there using them, seeing what happens, throwing away the duds and refining the good ones (if I have lost you in this weird club, substitute 'girls' with consciousness and 'chat-up lines' with 'theories'.)

As neurofunkers you create and test theories. But what should a good theory do? At its most basic I would say that it should predict the most with the least. That is, it should make predictions and it should be parsimonious (Hawking, 1996; Popper, 1963). It should explain.

But what should it explain? Broadly speaking, the answer is experience, in sickness and in health. So, any decent account of consciousness should have an answer for sleep and wakefulness, how anesthesia works, how brain damage has the effects it does, how attention and arousal work, how normal functioning operates and possibly even the point of consciousness, its function.

I won't go into too much detail but here is a list of the theories you should all have a good knowledge of:

Multiple Drafts Theory -
This is the theory of Dennett and Kinsbourne (1992), which suggests are brains are running multiples processes concurrently. Does all this information collect in some centre? They argue that this is not the case, “…the Multiple Drafts model avoids the tempting mistake of supposing that there must be a single narrative (the "final" or "published" draft) that is canonical - that represents the actual stream of consciousness of the subject” (p. 185).

Micro-Consciousnesses -
This position states that the brain is a parallel processor with many different consciousnesses arising from modularity. The champions of this idea are Zeki and Bartels (1998)


Consciousness from Language -
According to some, language allows outputs from other brain areas to be commonly understood and integrated to form conscious experience. The main advocate of this position is Gazzaniga (1995).

Consciousness from meta-representation -
It is entirely possible that rather than language (like this is written in) being the overall interpreter of brain wide functions some other system is involved in re-representing information from disparate brain areas into one unifying ‘language’. It might be this representation that is responsible for consciousness. Such is the view of Singer (1998) who says, “Brains capable of processing signals at a conscious level seem to have the ability to represent the outcome of their distributed computational operations in a common format. These metarepresentations comprise protocols not only of sensory and motor processes but also of the state of value-assigning systems. Thus brains that have consciousness possess a representational metalevel at which internal states are explicitly represented: they have what one might call an inner eye function....." (p.1829)

Representational Theories of Consciousness -
According to this position the brain is a reality emulator, representing the world in neural code, which once complex enough can create consciousness. Damasio (1998) says "[l]iving creatures such as we are, produce core consciousness when [they are] caught in the act of representing themselves when they represent other things." For Domasio the brain uses structures designed to map both the organism and external objects to create a fresh, second-order representation. Other champions of this position are Pare and Llinas (1995) and Llinas and Pare (1996). Sensory input, according to these authors, amends the current model in the brain. It is worth restating this point because their argument is subtle. Sensory input does not lead to internal images in the brain but alters the pre-existing model. In particular, they suggest that 40Hz resetting allows for the model to be changed. The whole position is quite well captured by Humphrey (2002), “Now imagine that a new form of sense organ evolves, an ‘inner eye’ whose field of view is not the outside world but the brain itself" (p.75).

Graded representation -
According this position the quality of information is critical to consciousness. The proponents of this position are Farah and Feinberg (1997) and broadly speaking the higher the quality of information the greater probability there is of it being associated with consciousness. Note similarities with Chalmer's panprotopsychism 'information theory' and how this view is essentially non-eliminative materialism.

Multiple Consciousnesses -
The suggestion with this position is that there may be qualitatively different types of consciousness. For instance, Block (1995) suggests there is access consciousness (A) and phenomenal consciousness (P). Another position put forward by Damasio (1998) suggests the existence of core and extended consciousnesses. He says, "[b]oth are internal phenomena of the mind but core consciousness is more basic than the extended variety. Extended consciousness depends on core consciousness. Both occur automatically....no amount of willpower can either make them happen or prevent them from happening" (p.1880). Yet another position advanced by Bogen (1997) suggests there is a central core or ‘me-ness’ to experience, which he links to subcortical areas, in particular the intralaminar nuclei of the thalamus. The content of this experience, he argues, is linked to activity in the neocortex.

Binding -
If there are indeed multiple consciousnesses or a host of parallel systems working away to solve their specific problems and consciousness experience is holistic then either this is an illusion or somehow these disparate sources of information are bound. Thus functional specialization must be overcome by functional integration. Ignoring the possibility that it is an illusion for the moment, there are a number of different versions of this binding idea:


Cartesian Theatre/Localized Processes -
First, binding could occur at convergence zones in the brain either locally or globally. This is the idea that information is integrated in space. Schacter et al (1988) is a supporter. According to these authors there is a single 'Conscious Awareness System' in the brain. This system is separate from perceptual, cognitive and action systems and has the special function of generating conscious experience. A similar stance is taken by Posner and Rothbart’s (1998) who suggest that the cingulate gyrus is the organ of focal awareness and consciousness. Milner (1995) makes the claim that the dorsal pathway is unconscious whereas the ventral pathway is conscious. Crick and Koch (1998) make the claim that visual consciousness (although not visuomotor abilities) may reside in the frontal lobes. Baars (1998) who theorizes that converging information in special areas are lit up by attentional processes. Dehaene & Naccache (2003) also advance a similar theory called the Global Workspace Model, this is a good theory and you should know about it.

Distributed consciousness -
Instead of a localised area in the brain for consciousness this position states that consciousness might be more distributed throughout the brain. This is the position of Kinsbourne (1992).

Temporally bound information flow -
Second, binding could be synchronized. That is, information could be integrated in time. This is the view of Pare and Llinas (1995)Llinas and Pare (1996). According to these authors the sensory input updates the various parallel system of the brain which is bound together using a 40Hz temporal oscillation. They suggest the thalamus is crucial in mediating this oscillation.

Function -
Epiphenomenalism -
This is the idea that consciousness serves no function but is rather an emergent property of a system as complex as the brain. Consciousness in this view has no function to speak of. The often-used analogy is that consciousness is like smoke over the factory. See also Velmans (1991), ‘Is Human Information Processing Conscious?’ He finds a lot of it isn't, so why the need for consciousness?

Adaptive -
As Humphrey (1987) (slightly inaccurately) put it, “either we throw away the idea that consciousness evolved by natural selection or else we have to find a function for it” (p.378) (It is slightly inaccurate because of course it doesn't need a function, it could be epiphenomenal.) To my mind, asking this question is of fundamental importance. Historically, psychology was satisfied to describe behavioural phenomena – it was a descriptive science. All that changed when the gene-centred view came into the arena. Instead behavioural scientists were apt to ask, “What function a certain behaviour served?” Only really by the 1990s had scientists begun to accept that humans were not exempt from this logic. The idea that consciousness may have some function is a very recent one although I believe it is one of the most crucial questions we can ask. Recognizing the function served by consciousness could be the key to understanding it. A number of theories abound about the function that consciousness serves.

Gray (2004), for instance, suggests that the brain is a massive parallel processor, although these servoprocesses, as he calls them, are not infallible ones. For this reason the brain must have some sort of error detection. By modelling the world including its enduring features such error detection can occur by comparing expected and actual models.

Dehaene and Naccache (2001) have suggested that consciousness is for i) durable and explicit information maintenance, ii) novel combinations of operations, and iii) intentional behaviour all constitutes such mental operations that require consciousness.

Ramachandran & Hirstein (1997) say qualia states allow information to be kept ‘alive’. This occurs because qualia states endure in short-term memory.

According to Gregory (1998) qualia may allow the present to be flagged in contrast to past knowledge.

***

The job your revision should take into this delightful subject is to match up the evidence to these theories not just positively, as in 'this supports this', but negatively so that you can say 'this evidence puts a spanner in the works for...' I say this last bit because you'll see in the past papers quite a lot of the questions ask you to rip another theory apart. This will set you up to do just that.

My advice is to set out all the theories on big bit of paper and then go through the course notes adding data that supports or rejects. Have a couple of favorite theories and preferably create your own.

Also have a couple of bad ones to talk about (psst - Gazzaniga [1995] is flimsy. Also panprotopsychism can be taken down a notch with a crisp argument for functionalism). There is always a question asking you to support or reject a theory with evidence or discuss the scientific progress into the problem of consciousness. Knowing how theories and data dance with each other will set you up nicely for this.


Also keeping theories in the forefront of your mind will hopefully keep a lid on the feeling of being overwhelmed by the thousands of papers you could read in this area. Just like I said in an earlier post, without a course blueprint you can't properly be selective. Well, without theories to constrain your choice of evidence you'll be lost in Consciousness.

No comments: