The brain is the most complex part of the human body. This is something we’ve been taught since the early days of science class. The hypothalamus controls body temperature, sleep, hunger, thirst, and emotions. The cerebellum controls movements and balance. There’s a lot more to the brain than that and there’s more that can be done with just a little bit of biomedical engineering.
John A. Paulson of the Harvard School of Engineering and Applied Sciences wrote, “A computer and human brain are, in many ways, more alike than they are different. A computer’s central processing unit contains billions of integrated transistors; the brain is comprised of billions of neurons.” (John A. Paulson seas.harvard.edu.) Using this idea of brains matching computers, scientists have made attempts to mimic the brain to help computers run as efficiently as possible.
This kind of engineering is called neuromorphic computing. Neuromorphic systems are designed and created to mirror the brain’s neural and synaptic structures to process information more efficiently and at lower energy costs and levels.
Integrated memory processing is a way of integrating memory storage and data processing into a specific area, this mimics the brain’s functioning. Integrated memory processing significantly increases energy efficiency, speed, and parallel processing capabilities. This technology also is a way to overcome the “von Neumann bottleneck” — von Neumann was an American-Hungarian scientist who is credited with designing the modern computer — which is the performance limitation created by the shared bus in the von Neumann architecture.
Stanford’s neurogrid is a specialized energy-efficient neuromorphic hardware system created to simulate large scale biological neural networks in real time in order to provide neuroscientists with an energy and cost efficient tool to better understand and study brain functions. The neuroscientists “interconnected neural arrays in a tree network; this choice maximized throughput. These three choices made it possible to simulate a million neurons with billions of synaptic connections in real time.” (Stanford University departments of Bio Engineering and Electrical Engineering)
This wouldn’t be one of my columns without talking about prosthetics at least once. Researchers from University of California at Berkeley and UC San Francisco have made a breakthrough in creating prosthetics controlled by the user’s brain to restore naturalistic speech for people with paralysis. What they did works around a flaw in brain-to-voice prosthetics in that there was a considerable amount of time lag between when the user attempted to speak and when they actually would. “Our streaming approach brings the same rapid speech decoding capacity of devices like Alexa and Siri to neuroprostheses,” (Gopala Anumanchipalli, Robert E. and Beverly A. Brooks UC Berkeley engineering) “Using a similar type of algorithm, we found that we could decode neural data and, for the first time, enable near-synchronous voice streaming. The result is more naturalistic, fluent speech synthesis.”