How Do You Build A Synthetic Brain?

Parker and co-principal investigator Chongwu Zhou, both of the USC Viterbi School’s Ming Hsieh Department of Electrical Engineering, have teamed up on the “BioRC (Biomimetic Real-Time Cortex) Project,” which has set out to create nanocarbon brain neurons that can talk to each other.

The research team includes USC Viterbi School electrical engineering graduate students Jonathan Joshi, Chih-Chieh Hsu, Adi Azar, Matthew Walker, Ko-Chung Tseng, Ben Raskob, Chuan Wang, Yoon Sik Cho, Changsoo Jeong and Jason Mahvash.

The team is studying the behavior of cortical neurons – what makes them fire and send signals through synaptic connectors to other neurons in the human cortex – as well as the neurons’ “plasticity,” or ability to learn and remember.

Each time a neuron fires, it sends an electro-chemical spark through thousands of other neurons at speeds of up to 200 miles per hour. But with approximately 100 billion neurons in the human cortex and approximately 60 trillion synaptic connections, the brain is massively interconnected, Parker said. That makes the task of unraveling a neuron’s electrical circuitry quite complicated.

“The brain is kind of like a biochemical factory, operating in a sphere that you can’t stretch out on integrated circuits and circuit boards in order to emulate all of its electrical activity,” she said. “The connectivity is too great and too many delays are introduced. We had to turn to nanotechnology to build something three-dimensionally, so that eventually we’ll be able to emulate how the neurons fire and activate others along a specific path within that sphere.”

According to Joshi, who has engineered the circuit design for artificial synapses that learn, “This is a big departure from some previous synthetic brain projects, which attempted to emulate neural behavior with electrical signals using conventional multiprocessors.

“Nanocarbon modeling solves problems such as the sheer physical size in building a section of synthetic cortex, the cost of expensive electronics that have been required in the past to build these structures and then the cost of powering them, since the brain never shuts off.”

Until quite recently, the size and cost of available electronics made construction of complex brain-like structures totally impractical, Parker said.

The team already has designed and simulated the transistor circuits for a single synapse, said Hsu, a senior member of the team and Ph.D. student in electrical engineering. In addition, a complementary metal oxide semiconductor chip that will be used to validate the concepts is about to be fabricated. Now it’s time to connect the structure to another synapse and study neural interconnectivity. By the end of the semester, she hopes to have “several synthetic neurons talking to each other.”

Ultimately, the researchers hope to answer one question: Will science ever be able to construct an artificial brain of reasonable size and cost that exhibits almost real-time behavior?

“We really don’t know if we can yet, despite all of the press that you’ve seen claiming how close we are to that,” Parker said. “The human cortex is massively interconnected and the connections are always changing. That’s always been one of the biggest hurdles in trying to simulate neural functioning. But as technologies become smaller and less expensive, there is a possibility of constructing neural structures on the scale of the human brain.”

A lot is riding on it, she added. Autonomous vehicle navigation, identity determination, robotic manufacturing and medical diagnostics are engineering challenges that could benefit from technological solutions that involve artificial neural structures.

And in medicine, the stakes are even higher.

“Researchers have already built experimental cochlear implants that are able to restore some hearing in the deaf and new vision systems that can restore some sight to the blind, but what we’re working on now is what you’ll see 30 years in the future,” Parker said. “This is work that could revolutionize neural prosthetics, for one thing, and give us some pretty amazing biomimetic devices.”

The project also involves collaborators Kang Wang, Alex Khitun and Mary Eshaghian-Wilner at UCLA, Philip Wong at Stanford University and Jie Deng at IBM, as well as neuroscience faculty members at USC.