We are closer to the ultimate neurology experiment: building a brain, writes Drew Turney.
The 17th-century philosopher Rene Descartes claimed there was a disembodied driver in the brain, a kernel of intelligence that viewed sensory input and wielded consciousness to act upon it.
Though we're no closer to discovering the soul today, we know about dendrites, axons (cell components) and synapses (empty, electro-conductive space).
But what is still a mystery is how even though the brain comprises little more than these simple structures, it has somehow given rise to everything from language to love, from Beethoven to Big Brother.
Now our increasing knowledge about the brain's building blocks is bringing us closer to the ultimate neurology experiment: building one. The Blue Brain Project is one effort to create an in situ brain, using software instead of proteins in a supercomputer called Blue Gene to model the brain of a two-week-old rat. It does so by creating a 3D computer model of neurons in the neocortex (the 'intelligent' sector of the brain) where scientists can simulate the sizes, shapes, densities and electro-receptiveness of different neurons in the biological version and watch their behaviour under certain conditions.
Phase one of the Swiss Ecole Polytechnique Federale de Lausanne project (evaluation of the ''consistency and relevance of neurobiological data'') is complete, and last July the project director, Henry Markram, announced that they were 10 years away from a functional artificial human brain.
The Blue Brain Project's goal is to model the brain's response behaviour down to the nano-level. If we've programmed the way each neuron operates we can impose specific conditions like injuries or diseases that will tell us how a biological brain will behave without risk to a living patient.
Though recreating the brain is the stated aim, the Blue Brain project isn't the first neurological architecture model. As long ago as 2005 the futurist and writer Kevin Kelly claimed the internet was essentially a brain. Every computer and device can be thought of as neurons, every bit of information like the bio-electrical sparks across synapses that constantly shift, connect and disconnect to form memories, thoughts, emotions and sensory data.
Jeff Stibel, author of last year's Wired for Thought, thinks the similarities go deeper. ''The internet already has the parallel processing capability of the human mind,'' he says, ''neuroscientists call it distributed computing, where the brain's functions are distributed all over the place to happen simultaneously. The internet is a massive storage and retrieval system, and the brain's fundamental structure is roughly the same.''
This interconnectedness is also the secret to that elusive quality we call intelligence. Older scientific understanding might have convinced us to try and replicate the brain by creating a super-driver, but the true smarts might be in the network integration of a huge number of simple, low-powered processors. ''A brain is really a massive composition of mini-brains or hives,'' Stibel says. ''There's no such thing as a central decision-maker in the brain; intelligence emerges from complex parallel processing of information.''
Bill Lytton, a Blue Brain Project technical adviser, is a professor of physiology, pharmacology, biomedical engineering and neurology at New York State University. He thinks if we can perfect the mechanism to generate those hundreds of millions of base-level computations in a simulated brain, the mental commands we know as decisions, emotions or thoughts might arise as spontaneously and magically in it as they do in us.
''The simultaneous processing going on will achieve its objectives without ever bothering to reach higher levels of integration,'' he says, ''and it's not just the brain that does this - consider chickens or mice running around with their heads cut off.''
The world view of the brain as a machine packed with incredible densities of deceptively simple parts isn't new. In his 1995 book Are We Alone? cosmologist Paul Davies suggests there is no life force. The only difference between us and rocks, air or plastic might be the complexity of the structure; make something complex enough - like the proteins that form DNA - and it can be termed alive, he suggests.
So how do even start creating something so intricate it generates normal function spontaneously from the basic engineering, like expecting a 747 to take off and fly by itself just because we've built it? While a single neuron is comparatively simple it's the sheer number of types and behaviours involved that makes the computational task so big.
Another way to gain a closer understanding of the brain is to examine how it interacts with itself, as Srini Pillay does. The chief of the coaching and organisational psychology company NeuroBusiness Group and a brain imaging researcher, Pillay uses brain scanning technology to watch for changes in blood flow relative to emotion. When he prompts an emotional response in a patient or subject, the positioning and quantity of blood flow offers the possibility of regulating it to generate the mental model or mind state we want to study.
The applications of brain modelling could transform neuroscience. We can already model chemicals - most of us did so in year 7 science - if we have a virtual mock-up of how they interact at the synaptic level it would let us design better treatment for a huge range of conditions from depression to stroke.
But as we have seen, simply arranging the parts won't cause the virtual brain to just switch on. External stimulus is the key to all theoretical biology. ''We can grow neurons and support cells in culture and make them grow,'' says Richard Senelick, a neuro-rehabilitation specialist and medical director of HealthSouth Rehabilitation Institute of San Antonio in Texas, ''but we can't effectively direct them to make the correct connections consistently to reproduce function.''
The limitations of a disembodied, software-based brain don't stop there. As Bill Lytton reminds us, a real brain doesn't exist in a vacuum. ''A full brain wouldn't be all that interesting without being connected,'' he says, ''and I doubt we'd ever have the resources to build a full simulated nervous system.''
''Recent studies on heart-brain pathways illustrate that we'd need to include feedback from the heart to understand the brain more closely,'' adds NeuroBusiness Group's Srini Pillay. ''Even in the brain itself, many of the functions of brain chemicals and pathways are U-shaped. Stress, for example, may impact the brain positively up to a point, but then starts to reverse its effects. We're still trying to understand these kinds of thresholds.''
But the biggest constraint is that we simply don't know what language the brain uses to work. ''We're missing the code,'' says Lytton, who worked with the DNA pioneer Francis Crick at the Salk Institute, where the latter devoted the remainder of his career to decoding the brain.
''The brain uses different codes simultaneously to deliver information at different rates for different purposes. It's difficult to record from many neurons at once and then analyse what you've recorded. Right now we can record up to 100 but we might need to do up to 10,000.''
Of course, if we can grasp and deploy all the theory and technology we need to build a simulated brain from the ground up, there's an ever bigger philosophical question.
In her 2008 book ID, the Oxford University pharmacology professor Susan Greenfield outlines how our personalities, our hopes and dreams and our sense of self are no more than the sum total of our own neurons.
''The mind'', she writes, ''far from being some airy-fairy philosophical alternative to the biological squalor of the physical brain, is the physical brain - more specifically the personalised connectivity of that otherwise generic brain.''
If that's true and we build a human brain from scratch, what will it mean for the spiritual dimension of human life if a fully formed, sentient consciousness wakes up, shakes it computerised head and declares: ''I think, therefore I am''?
No comments:
Post a Comment