[Help] [Aide] [Up]

Science Tribune - Introduction to article - May 1997

http://www.tribunes.com/tribune/art97/peru2.htm

Introductory comments to the article on "Neural networks, quantum systems, and consciousness" by Mitja Perus


Tiiu Ojasoo

E-mail : ojasoo@tribunes.com.


Consciousness and the man in the street

To the man in the street, the word consciousness probably conjures up, first and foremost, a negative image, i.e., a lack of consciousness (fainting and death) which he may well attribute to neurophysiological causes. However, if asked for a positive definition, he may describe consciousness as the subjective inner life of the mind : self-awareness, feeling the things he perceives in the outside world, attributing meaning and value .... and refer to the disciplines of neurosciences and psychology and, in particular, to Freud's subdivision of the mind in terms of different levels of consciousness. This understanding, largely derived from sciences of the late 19th century and first half of the 20th century, now needs to integrate much more recent methods of attempting to approach the study of consciousness. Our introductory comments will try and explain, in very simple terms, some of these methods, which will then be developed further in the article by Mitja Perus.


Objectivity

For over two hundred years now, science has been largely confined to the study of the objective external world and has ignored the subjective because it is not readily amenable to the hypothetic--inductive approach : formulating a hypothesis, testing it 'objectively' and, if possible, disproving it. We have placed ourselves in the position of independent non-participatory observers as if we were not part of the world we are studying. The real world is outside; it includes our brain insofar as it is study material but it does not include ourselves and our mental experiences and models. For this reason, the workings of the mind have been studied so far mainly in terms of molecular and cellular machinery. The mind is seen as a highly complex assembly of little machine units, that send and receive information, for instance in the form of chemical messengers at synapses; taken together, these units form a higher-order dynamic working entity. This classic metaphor of a factory is only partly satisfactory because it does not really answer the question how all these physical processes located in different regions of the brain are linked so that subjective experiences can emerge. Clearly, additional lines of enquiry need to be followed.


New disciplines that can help to explain consciousness

Scientific disciplines do not progress at uniform rates. The second half of the 20th century has been dominated by advances in new disciplines such as for example molecular genetics, immunology, materials sciences, computer sciences, information technology and high energy particle physics. It is thus not surprising that explanations for the workings of the mind and for consciousness have been sought in the worlds of mathematics (programming) and physics. Mathematics because it provides models to explain functions and physics because the fundamental laws of physics are thought to govern the universe.


Neural networks

The comparison of the human brain to a factory - as if it were like any other organ - has thus evolved into the now commonplace comparison between the brain and the computer. If tissues, cells and molecules are the hardware, what is the software ? A type of software currently in vogue is based on neural network theory. What is a neural network process ? Basically, a training process. The computer is provided with a data-set and with the solutions to problems relating to this data-set. By a system of trial and error, the computer will construct a model of behaviour that gives the appropriate solution to a given problem. This model, based on practice (i.e. concrete examples) and not on theory, can then be applied to solve problems concerning much larger data-sets to which the solutions are not yet known. The synaptic connections of the brain could be likened to neural networks in which case the data would be patterns of impulses.

The structure of the brain is not a single neural network but is formed of multiple networks that relate to each other and interact. As explained by Perus, neural networks have been successful in explaining several aspects of the functions of the mind, even some of the background processes involved in self-awareness, because neural circuits can self-interact.

However, do neural networks explain consciousness ? Can they account for the feelings and emotions we experience when we perceive something ? When looking at this screen, readers will perceive black print against a cream background - a purely physical process - yet they will also undergo a subjective 'interpretative' experience possibly in the form of an individual thought. This internal consciousness arising from a multitude of external factors (colour, presentation, etc...) is highly personal; yet first-person experiences are common to us all and apparently very similar in quality. Can the subjective unity of experience, creativity, intuition, free will ... be accounted for solely by a neural connectionist approach ? Many scientists think not and believe that quantum mechanics may play a part. This is the essence of a fiery ongoing debate.


Quantum mechanics

Many of us will have spent much of their secondary education learning the mechanical physics of Newton and some may have moved on to Einstein's theory of relativity. Less people are familiar with the statistical laws of quantum physics which deal with the infinitely small, with particles (energy quanta) that exist in a multitude of states and whose precise location is only known when the wave functions that describe them collapse into a single state (undergo a quantum jump). A specific feature of quantum physics is that the wave function collapses when it is observed, i.e., the observer and observation are part of the same world. For this reason, a collapse of wave functions, where the state of the brain changes, has been correlated with the crystallisation of thoughts.

As pointed out by Perus, a crucial issue in studying consciousness is one of scale. What is the scale required to describe human consciousness ? If a computer were to reproduce all the neural nets of the brain, how large and powerful would it have to be ? Is all the information in the brain processed by chemical messengers, synapses or are there other, submolecular, routes of communication ?

According to critics of quantum states of mind, information travelling by submolecular routes would need an amplifier if it is not to be drowned by background noise. Some members of the pro-quantum camp have suggested that the microtubule (inner protein scaffolding of the cell) might fulfill this function by permitting the long-range alignment of dipoles that are created by trapped charges in proteins that undergo quantum flip-flops in position. This alignment would represent a coherent quantum state. Might not, they argue, clusters of quantum states be the make-up of memories, creative thought, intuition.... .


Bridging the gap between the pro-connectionist and pro-quantum camps

Perus argues that a higher-level abstract organization, involving a combination of both the neural network and quantum worlds, might provide an explanatory framework for consciousness :
(i) The system of neural-like nets could operate on different scales including a quantum scale,
(ii) In complex dynamic systems, the whole is greater than the sum of the parts and not just an aggregate of independent entities.
(iii) Large scale processes are built up from small-scale processes. A hierarchy of quantum neural -like nets might thus act as an amplification system and create a new quality.
(iv) The mathematics involved in neural nets and quantum physics have analogies. One can imagine associative hierarchies between neural and quantum nets where different types of information become intertwined,
(v) Inherent to quantum physics is a quality of profound unity that could be related to a unified subjective experience.

Perus' thesis is that the inclusion of quantum physics in a description of the structure and function of our brain system provides an added dimension, i.e., a more subtle level of organisation. Like others, however, he cannot answer the question why brain function is accompanied by conscious experience. If the right abstract organisations were one day to be imprinted on silicon chips and the system made to function at the necessary speed, will machines become conscious ? An engineering problem for the 21st century Materials Sciences graduate ?



A footnote on dimensions and scale

On reading Perus' article, the reader may feel that attempting to explain consciousness addresses some of the same problems as trying to explain the universe. Is there a single 'theory of everything' that will account for observations on all scales, from the infinitely small to the infinitely large ?

There are several different types of forces in nature. Weak and strong nuclear forces as well as electromagnetism are described by quantum theory, whereas gravity is described by general relativity (the space-time world). As everyone knows, Einstein introduced the fourth dimension of time. The German mathematician T. Kaluza (1919) added a fifth dimension by unifying space-time with Maxwell's equations for electromagnetism. But how many dimensions does one need to unify all the forces in nature ? According to the theories of the 1990's, 10 dimensions in 'superstring theory' (which can be compacted down to 6 dimensions) or 11 dimensions in the later more comprehensive M-theory. How many dimensions does one need to describe consciousness ?


[Up]