Creating an Artificial Brain

Dr. Henry Markram recently announced that he expects to have a computer model of the human brain in ten years. As part of the Blue Brain Project, he is part of a team trying to “reverse-engineer the mammalian brain.”

The human brain is exceedingly complex. There are about 100 billion neurons within the human central nervous system (brain and spinal cord) with an estimated 100 trillion synapses (connections between neurons). However, the brain grows and prunes synapses constantly, so that number is fluid. In addition there are probably 1 trillion glial cells that serve as support among other roles to the neurons. This web of densely-connected cells produces all behavior, thoughts, movements, and emotions. The human brain is the most complex organ on earth and one of the most complex things in the universe, especially gram for gram.

Because it is so complex, the brain does not always function properly. Sometimes various chemicals or substances act as teratogens, which interfere with the normal development of a fetus and often result in serious and permanent neural deficits, such as fetal alcohol syndrome. Other times, genetic abnormalities like trisomy 21 produce Down’s Syndrome. Other abnormal brain developmental pathways can result in anything from mild, even grossly unnoticeable deficits, to death. For the most part and for most people, however, the human brain develops normally and functions as nature intended.

The researchers at the Blue Brain Project are working towards simulating the entire human brain with its billions of neurons and trillions of synapses. How feasible is the project? Currently it takes the equivalent of one computer — one microprocessor — to model a single neuron. The goal of the project is to accurately model individual neocortical columns (a cylindrical volume with a diameter of 0.5 mm and height of 2 mm that contains about 10,000 neurons) in series and then extrapolate that model out to the entire cortex, which should simplify the overall model and reduce needed computing power. Currently to model a single neocortical column the researchers utilize a supercomputer with 10,000 processors. As you can imagine, modeling the entire brain of 100 billion neurons is a mind-boggling task. However, with advances in computer hardware and software, we are moving closer to such models, especially as researchers are able to simplify the overall brain model by modeling it at a more macro level than individual neurons.

One reason the researchers give for wanting to model the entire human brain is so they can hopefully better understand brain diseases and abnormal brain development. Imagine being able to simulate a brain of someone with Down’s Syndrome, or better yet, the development of a brain of someone with Down’s Syndrome! We could then hopefully understand exactly what goes wrong and when and try to correct it genetically or through some other means. Or, researchers could model the brain of an autistic child to try and understand how it functions. There are myriad possibilities.

However, is creating a complete model of the human brain ethical? Would the model develop into a self-aware and potentially sentient entity? If the model has self-awareness would it be ethical or moral to turn off the simulation? Would the model or simulated brain be considered alive? What are the potential pitfalls, if any, to creating a fully-functional brain simulation? What happens if treatments or policies are created based on the simulated brain and those treatments prove deleterious or the models miscalculated? Could we plug a simulated brain into a body and create a new “person”?

I think this research is exciting and ground-breaking should it come to full fruition. On the other hand, I do not believe we should proceed without serious ethical discussions. This is not just cloning a sheep or a rat, this is creating a full simulation of the human brain that would ostensibly grow, develop, feel, and mature.

Jared Tanner, PhD

Jared Tanner has a Ph.D. in clinical psychology with an emphasis in neuropsychology. His interests are mainly neuroimaging and neuroanatomy. He spends his research time looking at the structure of gray and white matter in the brains of people with Alzheimer's disease and Parkinson's disease. With a focus on neuropsychology, he is also interested in how normal and abnormal brain structure relates to cognitive and behavioral functioning.
See All Posts By The Author