Mind's valleys

On the wall of the Montreal Neurological Institute are engraved the words of the institute's founder, Wilder Penfield, whose pioneering studies of the effects of direct electrical stimulation of the cortex of neurosurgical patients laid the foundations for modern functional neuroanatomy: "The problem of Neurology is to understand man himself". Ironically, Penfield was unable ultimately to accept that an understanding of the nature of mind arises directly from knowledge of brain functions, despite the obvious associations that he (and generations before and since) have observed. The relationship between mind and brain - between the conscious, purposeful sense of self and the physical substance of the brain - remains to be understood. Increasingly better definition of interactions between neuronal elements is the focus for most of modern neuroscience. The fundamental questions concerning the relationship of mind and brain remain the province of philosophy.

It is therefore refreshing to read how boldly a neuroscientist, J. A.Scott Kelso, director of the Center for Complex Systems at Florida Atlantic University, has taken on the challenge to define and test a concept of mind in this new book. Kelso has a clearly defined hypothesis: "The brain is a self-organised, pattern-forming dynamical system. And its coherent, but unpredictable spatio-temporal trajectories - brain behaviour - is the result."

The key to Kelso's concept is the daunting complexity of the brain. He argues that, although reductionist analyses would always fail to find characteristics of mind in any individual neuronal or neuronal network interactions, we should not be surprised at the way in which mind could arise from overall integration of the individual elements of the brain as an emergent property. The way in which this might occur is plausibly presented as a consequence of the tendency of (unquestionably inanimate) complex systems that are open (ie, exchanging energy with their environment) to become self-organising in ways that can lead to purposeful-appearing behaviour.

In developing this idea Kelso describes several experiments that strikingly illuminate the evolution of his own thought. The Rayleigh-Benard experiment, for example, is perfect "kitchen-sink" science. If a layer of oil is evenly heated with increasing flame in a frying pan, when a critical rate of heating is reached, a sufficient temperature gradient can be generated between the surface and the bottom of the oil to change suddenly the random motion of molecules of oil into an orderly pattern of convection rolls.

The mathematical tools of nonlinear dynamics can be used to describe how such convection patterns can appear so spontaneously. The effect of changing the rate of heat transfer is to move the system to a new stable state. The relative likelihood that any particular state can be achieved is determined by characteristics of the system as a whole. Under some conditions, the system may be so delicately equipoised between alternative states that precise prediction of behaviour may be impossible.

Kelso argues that the characteristics of an open, complex system of this sort, such as self-organisation into simple patterns, shifting between patterns of activity without a programme or other "master" switching mechanism, apparently nondeterministic behaviour, relational effects between control parameters, and nonlinear behaviour, also are those of the mind.

This is meant to be more than a metaphor. To test his concept directly, Kelso proposes that behaviour can be used as a mirror of mind and describes a series of investigations in which simple behaviours (such as the spontaneous shifting of rhythmical oscillations of the index fingers from antiparallel to parallel motions with increasing rates of movement) are very elegantly modelled using principles of non-linear dynamics. The potential generality of this approach is demonstrated by applications to phenomena as diverse as the way in which a horse's gait changes with running speed, and the almost unsuppressable shift in pronunciation of "eep" to "pee" at high repetition rates (try this experiment yourself!).

But what does all of this have to do with mind? A fundamental concept in Kelso's analysis is that principles of coordination in complex systems are independent of the level at which they are analysed. He argues that these simple experiments therefore provide a window onto the mechanisms of much more complex phenomena.

To oversimplify a subtle argument, Kelso proposes that we regard mind as a shifting landscape of great peaks and valleys of potential behaviours. The substance of the landscape is information. One can use the position of a (figurative) rolling ball to define the behaviour of a moment. Preferred stable behaviours are those driven by occupation of valleys, where the motion of the ball is arrested at least as long as the local landscape contours are fixed. Behavioural changes follow not from preprogrammed changes, but from movement (driven by shifts in the dynamics of this landscape of potentialities) between valleys corresponding to the drives for alternative behaviours.

Kelso boldly addresses major issues of brain and behaviour. For example, what is the meaning of "intention"? We all know only too well that "intention" is not always a programme for action. Kelso argues that we may consider "intention" as information acting on the dynamics of mind. The distinguishing quality of an intention is that it defines an "attractor" (a downward slope or valley in the landscape metaphor) favouring a new behaviour. But the behaviour may or may not be elicited, depending on the overall dynamic landscape of the mind.

This is not an easy book, as the non-linear dynamic models underlying Kelso's concept of mind cannot be appreciated without understanding at least the rudiments of the mathematical principles used in their description. Nonetheless, I found the arguments compelling and would hope it is widely read by educators, clinicians, neuroscientists and professional philosophers. None of these groups may be entirely comfortable with his formulation of the problem, but this may make the ideas all the more stimulating. Kelso's language is direct (and often amusing). The ideas that he tackles are big. Although ultimately he has not "solved" the problem, this chronicle of a career in cognitive neuroscience wonderfully illustrates the way in which theoretical biologists work and offers intriguing insights into the meaning of mind.

Paul M. Matthews is director, Centre for Functional Magnetic Imaging of the Brain, University of Oxford.