about> going inside

Here is a chapter by chapter description of Going Inside, my big book about how the brain works and how to think about consciousness as a property of brains.


The 1990s were meant to be the decade of great discovery in neuroscience. There were new brain scanning techniques like PET and f-MRI and a new theoretical boldness from the likes of Francis Crick. At last, it seemed, the secrets of consciousness would be laid bare. But it turned out that what had to be unlearnt became as important as what was learnt. Researchers found they were entering the hunt for the secrets of the mind with too many wrong assumptions.


One of these key assumptions was that the brain was much like a computer. The brain had a hierarchical design so that sensory input entered at the bottom and was transformed by stages into conscious output. Cognitive scientists went even further in believing that particular brain modules would "do" specific functions such as memory recall, speech generation and perhaps self-awareness itself.

But as soon as researcher began using PET to scan real brains it became obvious that their circuits reacted in a much more globally coherent way. There was something structured about the brain - it did have a hierarchical organisation of mapping areas. But equally, it was dynamic and adaptive in its processing flows.

Experiments by Harvard's Stephen Kosslyn to discover how humans generate mental images illustrated the problem. When his subjects were scanned, individual modules did not light up. Instead, as he says, the brain's response was more like watching pebbles tossed into a pond. Ripples spread across the brain spinning complex feedback patterns.

[features Stephen Kosslyn, Roger Tootell, cognitive science woes, brain-scanning hopes]


The dynamism that Kosslyn and other neuro-imagers were seeing at the whole brain level was just as troublesome at the level of individual brain cells. Researchers had assumed that like a computer, the brain's circuitry would deal in a stream of bits. Neurons would signal to one another to create an accumulating pattern of processing. But it was being found that the output of a cell could be wildly variable. A spike might or might not cross a particular junction depending on whether the brain was interested in the message.

It was as if consciousness were controlling the spikes rather than the spikes adding up to produce a state of consciousness! This sounded spooky but the mystery was easily explained by a dynamical view of the brain. If neural states of representation had to evolve, then the firing of any cell needed to strike some sort of feedback-based balance with activity in the rest of the brain. The spiking of a neuron did not have meaning until it had developed to fit a context. The problem was that neuroscientists found it hard to accept just how deeply this changed the traditional conception of the brain.

[features Donald Hebb, Warren McCulloch, neural networks bandwagon, discomforting implications of chaos theory]


The finding that really set the cat among the pigeons was Robert Desimone's 1985 report of attention effects in V4, the "colour filter" of the visual cortex. If a monkey had to pay attention to a green bar and ignore a red bar, then red-coding neurons would be damped within half a second - a conscious-level intention could feed back to control the firing response of an individual brain cell. The idea that neurons coded pixel-like for discrete bits of information could no longer be sustained.

And yet apart from this, the computational view seemed to be telling the story so well. Neuroscientists were finding that neurons lined up to form topographically organised mappings - images were literally painted across the visual cortex in twinkling neural outlines. And these mappings then stacked up to form complex processing pathways with each map refining the information captured in the level below. With new discoveries like population voting and synchronised firing, neuroscientists even seemed close to cracking the neural code - an exact understanding of how brain cells encoded messages to each other in their complex packets of spikes.

[features Robert Desimone, David Hubel and Torsten Wiesel, Wolf Singer, cortex maps, population coding, synchronous cell firing]


Despite the continuing attempts to see neurons as simple input-output units, by the mid-1990s this position was crumbling fast. Not only was the behaviour of cells dynamic on a sub-second timescale - they responded to the attentional needs of the moment - but other work was showing how they could fluidly adapt their receptive fields over minutes, days and months. A hand coding cell could become a face coding cell if the local balance of feedback changed sufficiently.

Eventually, younger neuroscientists like Karl Friston - who knew about non-linear maths and complex systems - began to see how the actions of individual cells could not be divorced from the activity of the brain as a whole. Patterns of firing would settle as different levels of brain processing fell into temporary agreement. This quite different view of brain activity demanded new methods of measurement such as Friston's idea of the neural transient.

[features Francis Crick, Michael Merzenich, Karl Friston, the paradox of fluid structure]


Reluctantly, neuroscientists were being forced to give up their simple computational view of the brain. But one startling finding of the 1960s should already have given pause for thought. Indeed, even the work of the very first psychologists - 19th Century researchers like Helmhotz and Wundt - should have been a warning. Because consciousness seems instant and effortless, it is tempting to assume it is actually so. But careful reaction time experiments had already suggested that awareness develops slowly and in stages.

Then a San Francisco researcher, Benjamin Libet, stuck stimulating electrodes in people's heads during brain surgery and showed that it seemed to take a full half second for a person to evolve a state of consciousness for a new experience. The time it takes to settle a fully-tuned spread of neural representation means that we must constantly run half a second behind reality - although for some reason we never notice the fact.

The critical response to Libet's experiments over the following 30 years tells much about why science has struggled so hard to find the right path to an understanding of the human mind.

[features Wilhelm Wundt, Wilder Penfield, Benjamin Libet, Bernard Baars, mental chronometry, pre-consciousness and automatic actions, the freewill issue]


The big sticking point with Libet's results was that half a second seemed such a long time. If consciousness was the result of activity in the brain, then everyone knew it had to lag reality simply because of the time it took for signals to travel across its maze of billions of connections. But while most researchers could stomach a "barely noticeable" delay of, say, a tenth of a second, Libet's half a second posed too many uncomfortable questions - especially for the standard cognitive science view of brain processing.

Again, the answer had already been discovered by 19th Century psychologists. But the rise of sports psychology in the 1980s brought the matter into sharp focus. A tennis player or baseball batter has to contact the ball within a window measured in milliseconds and millimetres. The only thing that makes such accuracy possible is anticipation. And as a few cognitive psychologists such as Ulric Neisser and Bernard Baars realised, anticipation must lead the way into every moment of consciousness.

We begin each "perceptual cycle" with a set of plans and expectations that allow us to deal with the moment smoothly and skillfully. Consciousness does not lag. Instead it grades from strong prediction to settled resolution. But the question was how the brain might actually generate states of expectation? A dynamic view of the brain's pathways gave an obvious answer - and again, experiments carried out in Robert Desimone's lab held vital clues.

[features sports psychology studies, Ulric Neisser, Bernard Baars, Robert Desimone, Keiji Tanaka]


The dynamism of the neuron and the stretched-out cycle of processing needed to arrive at a state of settled awareness were two key realisations. But there were a number of other essential ingredients for a new view of consciousness.

For instance, there was the idea that the brain's purpose would be clear in its design. Early in the 1990s, it was recognised that the sensory cortex was divided by a what-where logic - activity headed for the temporal lobe was most concerned with questions of object identity and object meaning while activity headed for the parietal lobe focused on the question of location and spatial relationships.

There also seemed to be a similar division of the cerebral hemispheres with the left half of the brain having a "narrow" attentional focusing style and the right processing the same information in a broad or holistic way. Another point that became obvious was that the brain had to have a pathway for evaluating the moment - for deciding what particular aspect to promote to the eye of consciousness - and that the need to squeeze in this valuing step might be one reason for the delays observed by Libet. A large body of psychophysics research - especially Evgeny Sokolov's work on the orientation response and a EEG wave known as the P300 - seemed now to tell a story.

Then the first really important discovery to come out of PET scanning began to hit home. Research at Washington University and Hammersmith Hospital was revealing the "practice effect" - what happened in the brain as a mental skill went from being in the eye of consciousness to become a pre-conscious, unthinking habit.

[features Gerald Edelman, Evgeny Sokolov, hidden logic of the cortex sheet, aha! feelings, the P300 oddball response, the practice effect]


The idea of a dynamic perceptual cycle, with anticipations paving the way for the quick and graded assimilation of sensation, made sense of a lot of puzzling data. But the more astute realised that there was a step further still. Sensation could not be meaningful until it had begun to inspire a response - until the motor cortex had become woven into the evolving state of mental representation to provide at the least the inklings of an output intention.

The frontal lobes were discovered to have the same kind of hierarchical organisation as the sensory cortex, except the logic worked top-down with high-level thoughts being translated by stages into a concrete set of muscle commands. The essence of the moment - whatever aspect was turning out to matter most and so forming the centre of attention - would be "loaded" into the working memory buffers of the prefrontal lobe.

From there, associative connections would direct traffic down through the rungs of motor mapping to rouse a suitable state of response. Seeing a coffee cup would automatically inspire thoughts about what could or should be done with a coffee cup.

The motor hierarchy was actually only one leg of the brain's output machinery. Equally substantial was a hierarchy for orientation - for thinking about where to look next - and in humans, a language hierarchy. The focus of each cycle of processing also automatically became the potential jumping off point for some voiced, or merely thought, comment. The need to wait for the brain to become fully orientated to the moment - to reach a state poised for meaningful output - seemed another reason why the full perceptual cycle might take as long as half a second or more.

[features frontal motor hierarchy, prefrontal organisation, the complex role of the hippocampus]


Having described the general cycle of processing that takes place during every moment of consciousness, and then how this cycle develops across the cortex, this chapter looks at the role played by two sub-cortical organs, the thalamus and basal ganglia. The thalamus is used by the cortex as a lens to focus its own activity while the basal ganglia allow for the controlled (willed) development of thoughts and reactions.

[features David LaBerge, Chris Passingham, Patricia Goldman-Rakic, Ann Graybiel]


This chapter finishes the "systems level" review of the brain by zeroing in on the key decision the brain has to make during every moment - what to escalate to focal consciousness? The brain has to have a mechanism for choosing whether to stick with its planned and anticipated point of focus, or whether to break to take notice of an interruption or a surprise.

[features Jeffrey Gray, Joseph LeDoux, Michael Posner, the cingulate cortex, amygdala, and nucleus accumbens]


The conclusion begins. This chapter draws together the many strands of evidence to show what it means to see the mind as a dynamic brain process - an adaptation that takes place in the blink of an eye, but also one that is continuous with a recent history of working memories and expectations, a lifetime's history of habits and learnt perceptual skills, and even a genetic legacy of evolved neural structure.


But there is still the puzzle of exactly how the human mind is different from an animals. This chapter looks at how language allows for the extra mental abilities of humans. In particular, it shows how language-based habits of thought sink in to become part of the very structure of the brain and also how anticipation solves the traditional riddle of how we know what we are going to say or think before we the words are uttered (either audibly, or to ourselves using our inner voice).

[features evidence of Hominid evolution, Lev Vygotsky, scanning studies of language areas, learning the skills of recollective memory and self-awareness].


The sign-off chapter considers how close we can come to an intellectually-satisfying account of consciousness - one that balances the computational and dynamic approaches - and why some deeply rooted assumptions have led many to expect completely the wrong kind of answer.

Going Inside - a tour around a single moment of consciousness, by John McCrone, published March 1999 by Faber & Faber in London (ISBN 0 571 17319 5, price £17.99).

home> about> going inside