May 17, 2012

How complex is a mouse brain?

Microsoft co-founder Paul Allen founded the Allen Institute for Brain Science in 2003 with the aim of unlocking the secrets of the human brain. A human brain is 1,000 times as complex as a mouse brain. The neurons in a rat brain are in many respects functionally similar to those in a human brain. But given the daunting challenge of trying to understand something as complex as a human brain, brain researchers are starting with the more tractable issues involved in the operation of a mouse brain. The Allen institute is focusing its efforts on understanding how a mouse sees the world, in order to gain insights into how the brains mice and humans interpret vision. In an interview with Sander Olson for Next Big Future, Senior Scientific Director Dr. Hongkui Zeng explains why studying the brains of rodents may be the most effective method of discovering how the human brain operates.


Hongkui Zeng

Question: The Paul Allen institute has just launched three separate brain initiatives. What do these initiatives entail?

The three initiatives pertain to neural coding, cell types, and cell networks. I am overseeing the cell types program. The general goal of these programs is to attain a substantial understanding of the neural networks that operate in a human or mouse brain. Each program deals with a different level. the network level, the neuron level, and the intracellular level.




Question: You have done extensive research on mouse brains. Is the cell morphology of a mouse neuron similar to that of a human neuron?

Yes, the neurons in a mouse brain are highly similar to those found in a human brain at a general level. There are many differences at a detailed level. The human brain contains 1,000 times as many neurons as a mouse brain. The underlying principles are similar, but the human brain is vastly more complex.

Question: Aside from morphology, how does one study the activity of groups of neurons?

Studying the activity of a brain is a bit like studying a google map from the satellite-level view to the street-level view. There are levels and levels of complexity and activity. Collecting and simulating all this information requires prodigious quantities of computing power.

Question: It seems like you would need a supercomputer to simulate the workings of a neuron?

Yes, in the next couple of years the supercomputer will become an indispensable tool. There is so much data that we need to process and synthesize. There are 100 million neurons in a mouse brain, and 100 billion in a human brain. And the morphology of each is unique.

Question: Other than the Paul Allen institute and the Blue Brain project, what other labs are working on understanding the human brain?

There are many research labs throughout the world, including the Howard Hughes Medical Institute, several Centers at MIT, and the Gatsby foundation in the UK. But in Europe, Henry Markram is leading the Human Brain Project initiative. If funded, that project should last for a decade and receive possibly 100 million euros of funding.

Question: The goal of fully understanding the human brain by 2022 seems highly ambitious.

Yes, we set more realistic goals. We hope to derive general principles of brain operation by first studying specific capabilities of mice, such as vision. For instance, how does the brain differentiate the subtle difference in images? Both animals and humans have this capability.

Question: Can brain activity be simulated by a classical computer?

I don't believe that classical computers can simulate brain activity. The brain is the most complicated object or machine in the universe. Every adult human brain contains 100 billion neurons, and every neuron is different. How many possibilities for interaction between different neurons are there? We don't have a full understanding of how a brain works yet, but I cannot see any digital computer ever performing a fine-grained simulation of a human brain.

Question: What role do axons, dendrites, and glial cells perform in thinking?

Glial cells are not neurons, they perform support functions for neurons. Axons and dendrites are part of the neurons. Every neuron contains 3 parts, the cell body, known as the soma, the dendrite, and the axon. Neurons receive information from the dendrites, and send information out via the axons. A synapse forms between one point on the axon of neuron one and one point on the dendrite of neuron 2. Every neuron has a unique morphology of dendrites and axons which harbors thousands of synapses.

Question: So the brain is continually rewiring itself?

Yes, the brain continually rewires itself based on external stimuli. The basic wiring diagrams are determined largely by genetics. But perhaps 1% of the synapses continue to rewire after adulthood. That is all related to experiences and external stimuli. So on a macro scale, your brain is the same today as yesterday, but on a cellular level your brain is different today than yesterday.

Question: Paul Allen has just committed $300 million to the institute. How will this money be spent?

The money will be spent on infrastructure, such as new building, lab space, microscopes and computers, and salaries. Much of the equipment needed to pioneer new brain research is quite sophisticated and expensive, and we want to ensure that we have the best tools available.

Question: What sort of computing resources does the institute have at its disposal?

We don't currently own a supercomputer, but we might acquire one. Or we may collaborate with other institutions, such as Microsoft and Google. That would allow us to take advantage of their server farms. Or we may collaborate with a supercomputer center. So attaining the necessary computing resources will not be a bottleneck.

Question: How has our knowledge of the human brain changed during the past two decades?

Our knowledge of the human brain has exploded during the past two decades. Our study of animal brains has enhanced our knowledge of human brains considerably. We now have a much better understanding of how neurons function, and how they interact with each other. We have also benefited from increased computing power and more sophisticated equipment.

Question: Ray Kurzweil has argued that our knowledge of the brain is increasing exponentially. Do you agree?

Although he is correct that our knowledge of the brain is increasing exponentially, there are bottlenecks. In order to overcome these bottlenecks, we will require new technologies, such as high-resolution, non-invasive imaging. We need to find a way to monitor the simultaneous activities of large number of functioning neurons in a live animal at very high temporal and spatial resolutions.

Question: How feasible is completely non-invasive scanning of large numbers of neurons?

Completely non-invasive scanning is not possible for the foreseeable future. There are functional MRIs which are widely used, but the spatial and temporal resolution is very poor. We are open to any concepts which would increase the spatial and temporal resolution, but we aren't currently aware of any. More invasive technologies that are being used in animals are better at detecting neural activities.

Question: Henry Markram's Blue Brain project is focusing on rat brains. How much similarity exists between the two projects?

The Blue Brain project is focusing on rat brains, and we are focusing on mouse brains. Rat and mouse brains are very similar to each other. So the two projects have many similarities as well. However, there are many experimental approaches that can be taken, so at a detailed level there are many differences between the two projects. Rat and mouse brains contain 100 million neurons, and human brains contain 100 billion, so understanding rat/mouse brain function will not immediately lead to a comprehensive understanding of human brains. But the human brain is so complex that it makes more sense to start with exploring rodent brains and work our way up.

Question: What specific goals does the institute have for the next decade?

Our goal is to further brain research in general, and to gain a comprehensive understanding of the structural and functional basis of one model system - the visual system. So we are not focused on a particular disease. But the knowledge that we obtain should be applicable to brain diseases, such as Parkinsons and Alzheimers. There are many researchers working to try to understand the mysteries of the brain. I am confident that we could find effective treatments for many neurological diseases within the next decade.




If you liked this article, please give it a quick review on ycombinator or StumbleUpon. Thanks

Congratulations! Now you can use SolidOpinion commenting system in all its magnificence! Click the link to get your password.

Форма для связи

Name

Email *

Message *