My abiding interest is consciousness and the brain, and has been since I was very young. To set the scene properly as to how I have approached this difficult problem over the last 50 years, let me give a bit of my background. I was originally trained at Cambridge University, in England, as a theoretical physicist, and researched over the subsequent decades in quantum field theory, elementary particle physics, cosmology and quantum gravity, finishing up on supergravity and superstrings [1]. So I bring to my life-long interest in consciousness a scientist's hard-nosed belief that the world can only be explained by mathematically-based concepts. This approach has led to the amazing advance of our understanding of the physical world. Think of how our picture of the world has been transformed by the use of it over the last century. We now know about the quarks and gluons inside the protons and neutrons, which are themselves deep inside the atom. We even expect ultimately to discern superstrings when we can summon up enough energy to release them from their entrapment in our low energy, low dimensional surroundings. Further away, black holes and neutron stars are now familiar objects to us. Even more exotic objects are suspected to exist in the universe
This is the prior training I have at my disposal to tackle the problem of how the mind could be created from the brain. Over 33 years ago, I started to use neural networks in order to develop a neural model of consciousness. I began with an attack on real neurons, and constructed a mathematical model of how they can respond in a noisy manner to inputs [2], out of which later sprang (in collaboration with colleagues) a hardware chip with on-chip learning [3], mimicking this noisy neuron in some detail. But the gap to consciousness at that time was too big. So I changed tack to a higher level, and pursued a more functional approach.
To develop this functional approach, I realised the need for memory's guidance to give meaning to inputs to cortex. Through past experience, such meaning could be given expression. From this stemmed my 'Relational Mind' model of consciousness [4], which I then expanded and explored in brain terms in a variety of manners to lead to the 'Competitive Relational Mind' model [5]. Allied to that was a mathematical analysis of brain activity, leading to 'bubbles' of autonomous activity [6]. I suggested that these are important in creating suitably long-lived brain activity to sustain the conscious experience (there is now increasing experimental evidence for their existence) [7].
To get a closer grip on the brain I went to work at a brain imaging group in Germany from 1996-8, analysing visual illusions, memory, attention and motor control through MEG, fMRI and PET [8]. This led me to the parietal lobe as an important brain region for the creation of consciousness in the brain [9]. I realised that essential information is gathered there for overall control of response and experience; this information I termed the 'Central Representation'. However I was at that time unclear of its nature.
Over the last three years, I have found it possible to probe more deeply into the nature of the Central Representation. I have done this by developing an engineering control approach to what is arguably the highest control faculty of all that is occurring in the brain: attention. The notions of inverse and forward control models allow a picture of the movement of attention to be created which explains psychological data derived from studies on normal and brain-damaged people, both at a general [10] and more detailed simulation level [11]. The model can also be related to brain imaging results using various attention paradigms.
Attention is necessary for consciousness. So I was led to ask where in my control model of attention could consciousness reside? There are (at least) two aspects of consciousness relevant here:
The experience of the external world (including oneself and one's body as objects),
The feeling of 'what it is like to be conscious'.
The former - the contentful component - was one I felt comfortable with in terms of the Competitive Relational Mind model I mentioned earlier [5]. However I had to probe into the activity of the attention movement control model more closely. This would allow me to understand how the so-called 'pre-reflective self' - that of inner experience - could arise from the neural activity of the model. It was out of this that I derived the CODAM model [10]:
The pre-reflective self is created in brief spurts and flashes as short-lasting neural activity on a buffer holding a copy - the so-called corollary discharge signal - of the attention movement control signal.
This buffered activity is sandwiched between that with more persistence residing on the well-established buffer supporting working memory, for attended input from various cortical sources. The details of the CODAM model led naturally to the remarkable property that we each possess: immunity to error about our own experience. I cannot respond sensibly, when you tell 'I am in pain', by asking you 'Are you sure it is you who is in pain? You are certain it is you without any possibility of being mistaken. Such immunity to error may not be completely sacrosanct: it may be lost, for example, in schizophrenia (although this is yet controversial).
More recently I realised the importance of the CODAM model in giving a possible understanding of the strange experience of 'Pure Consciousness', described in many meditatory states across the world's religions: Buddhism, Islam (Sunni), Christianity, Hinduism. This state can be achieved, in the CODAM model, by meditatory effort to alter the corollary discharge buffer connectivity, so as to inhibit completely all incoming sensory input to its own buffer sites [12]. This is achieved by creation of a special goal, through various meditatory practices, to produce the required inhibition of awareness of any content. The result is the neural activity of the brain only attending to itself: the 'eye' of attention looking only at itself. I have conjectured that this is the source of the notion of the human soul.
Much work has to be done to amplify the CODAM model to work out how it fits with other brain activity. Presently I am investigating language, the most human of all activities needing consciousness. I have accordingly begun, with colleagues, to model how language might be based on the overall brain incorporating CODAM. I have started with a simplified neural model of the frontal lobes, to determine how it could support various syntactic components of language, in collaboration with Dr Neill Taylor here at King's [13]. This is now developing into a more complete model of language: the language acquisition device of Chomsky, or LAD [14].
The future of this work is along at least two main lines:
Development of a software system able to understand language (the LAD system) [14], and further scientific exploration of the overall LAD system's ability to support consciousness;
Exploration of the CODAM model, as inserted in the LAD brain model, to generate predictions of behaviour testable by suitable neuro-scientific techniques.
I hope I have indicated to you the excitement I feel about this development over the last 3 years. I feel I am close to the finishing post in the 'Race for Consciousness'. I may of course be on the wrong horse. But at least I have not been unseated by cranky scientific bye-roads. One such is the erroneous belief that quantum mechanics must be explicitly involved to enable consciousness to be created (my own earlier grounding in things quantum gives me a level of expertise that makes me confident in this assessment). These cranky approaches seem remote from inner experience and related brain activity and structure. I now have a mount that can be rigorously tested by such scientific methodology. But above all, my horse possesses, I claim, a structure able, for the first time, to generate from its neural activity its own soul. This is a large claim. Only time will tell if I am wrong. But time certainly will do so, since the tests can be made.
References
[1] Taylor JG, Restuccia A & Bressloff PC (1994) Finite Superstrings. Singapore: World Scientific.
[2] Taylor JG (1970) Spontaneous Behaviour in Neural Networks. J. Theor. Biol. 36:513-528
[3] Clarkson TG, Gorse D & Taylor JG (1992) From wet-ware to hardware: reverse engineering using probabilistic rams. J. Intelligent Systems 2:11-30 (www.freenetpages.co.uk/hp/tgclarkson/pram.html).
[4] Taylor JG (1973) The Relational Mind. Seminars given at the Universities of Tubingen, London, Trieste and Brunel (unpublished); ibid (1991) Can neural networks ever be made to think? Neural Network World 1:4-11
[5] Taylor JG (1999) The Race for Consciousness. Cambridge, MA: MIT Press.
[6] Taylor JG (1998) Neural 'bubble' dynamics in two dimensions: foundations. Biol Cybernetics 80:393-409
[7] Taylor JG (1998) Cortical Activity and the Explanatory Gap. Consciousness and Cognition 7:109-148 & 7:216-237
[8] Taylor JG, Schmitz N, Ziemons K, Grosse-Ruykin M-L, Gruber O, Mueller-Gaertner H-W and Shah J (2000) The Network of Brain Areas Involved in the Motion AfterEffect. NeuroImage 11:257-270 (together with other references on my web-site: address below)
[9] Taylor JG (2001) The Importance of the Parietal Lobes for consciousness. Consciousness and Cognition 10:379-417.
[10] Taylor JG (2000) Attentional Movement: The Control Basis for Consciousness. Neuroscience Abstracts 26:2231, #839.3; ibid (2000) The Central Representation: The Where, How and What of Consciousness. Pp 117-148 in The Emergence of Mind, ed KE White. Milan: Fondazione Carlo Erba.
[11] Taylor JG & Rogers M (2002) A Neural Control Model of the Movement of Attention. Neural Networks (in press)
[12] Taylor JG (2002) Paying Attention to Consciousness. Trends in Cognitive Sciences 6:206-210 (http://journals.bmn.com/journals/list/latest?jcode=tics); ibid (2002) From Matter to Mind. J Consciousness Studies (in press)
[13] Taylor JG & Taylor NR(2002) Language and the Brain: Constructing LAD. In Theoretical Models of the Cerebral Cortex, ed R Hecht-Nielsen, London: Springer (in press).
[14] Lobal Technologies (http://www.lobal tech.com)