[ad_1]
Michigan Technological University researchers are applying neuromorphic computing to improve the efficiency and efficacy of deep brain stimulation systems used to treat Parkinson’s disease.
Currently incurable, Parkinson’s disease is a neurodegenerative disorder that affects millions of people worldwide. Deep brain stimulation (DBS) is an alternative to effective drugs, but loses effectiveness when patients develop drug resistance. Over time, larger doses of medication are necessary to control the condition, and with them can come serious side effects. DBS is one option.
Making deep brain stimulation systems better for patients
DBS systems act as pacemakers for the brain. Parkinson’s disease causes motor symptoms, including slow or slowed movements (called bradykinesia), tremors, and rigidity. An electrode implanted at a specific target in the brain generates electrical impulses using a battery-operated device in the chest.
For people with Parkinson’s disease, DBS systems can be life-changing. But battery life is a challenge. Current devices use a surgically implanted pulse generator (IPG) in the chest or abdomen to continuously send stimulation signals to the brain regardless of the patient’s clinical condition. This causes the IPG battery to drain quickly. Battery replacement can be disruptive to patients; It requires a surgical procedure. And there may be unwanted side effects due to continuous stimulation of IPG.
Chunxiu (Traci) Yu, assistant professor of biomedical engineering, and Hongyu An, assistant professor of electrical and computer engineering, are working with their research team to develop strategies using a different tool: neuromorphic computing.
“Neuromorphic computing, also called brain-inspired computing or neuroscience-powered artificial intelligence, mimics the nervous system using microchips and algorithms. It is also highly energy efficient,” Yu said.
“Using neuromorphic computing to improve deep brain stimulation for Parkinson’s disease is very new. To our knowledge, this is the first effort in the field.”
The closed-loop smart system provides intelligent adjustments
Both the U Brain-Inspired Engineering Lab in the Department of Biomedical Engineering, and the An Brain-Inspired AI Lab in the Department of Electrical and Computer Engineering are designing strategies to improve DBS systems.
The collaborative project is focused on a closed-loop DBS system that can intelligently adjust stimulus signals based on patient symptoms.
“Most DBS systems are open-loop. Open-loop DBS is available 24 hours a day, 365 days a year,” Yu said. Open-loop systems are high in energy consumption, they provide constant stimulation to the brain, because the real-time signals are unknown to the device. “Using a closed loop system allows us to optimize the energy efficiency of DBS equipment,” Yu explained. “The patient’s brain signals are used to generate a therapeutic signal – a stimulus – in real time as needed.”
Using spiking neural networks to detect signals
The cornerstone of U and N’s closed-loop DBS is neural networks, or SNNs, artificial neural networks. SNNs can detect Parkinson’s symptoms and generate optimized electrical stimulation pulses.
“The communication signals in SNNs are represented by small spikes of electrical impulses in volts,” Ann explained. “In digital systems, information is represented by high and low voltages. For example, a high voltage level represents a logic one and a low voltage level represents a logic zero. In this way, digital systems represent information as binary numbers.
Information in SNNs can be captured in time, such as the interval between spikes, in N. “As a result, SNN systems are much more energy-efficient compared to other artificial neural networks,” he said.
The researchers’ new closed-loop DBS system can assess the severity of Parkinson’s symptoms by measuring nerve activity in specific brain waves, or the amplitude of oscillations. Areas of the brain that control movement produce beta oscillations.
“We use beta oscillatory activity as a biomarker because it can be detected faster than other methods, such as tremor symptoms,” Ann said. “If the neural activity detected is unusually strong, it indicates that the symptoms of Parkinson’s disease are more severe.”
The SNNs in N Labs are powered by one of the most advanced neuromorphic chips around: the Intel Loihi. In collaboration with Intel, the lab is actively exploring ways to use the chip’s super-efficient intelligence to help patients with Parkinson’s disease.
We found that neuromorphic chips, including Intel Lohi, are 109 times more energy efficient.
Another innovation: Ann and Yu replaced SNN’s traditional electronic memory with memster, the electrical power used in next-generation computers and electronics. A memristor can both store data as a memory chip and resist the flow of electrical current, just like a resistor in an electrical circuit.
It looks like a memristor resistor. The difference is that the resistance is variable. “With carefully designed signals, the resistance of the memory can be changed to many or thousands of different resistances. This feature greatly increases the amount of information that can be stored by individual memories,” said Ann.
In images, DBS systems use memristors for smaller chips, faster transmission of signals, and lower power consumption.
“This result is very promising,” Ann said.
Designing a custom DBS chip is the next step.
Ann and Yu plan to collaboratively design their own memristive neuromorphic chip specifically for closed-loop DBS systems.
“Even wearable medical devices are the best they can be these days,” Ann said.
For their students at Michigan Tech, the ongoing joint research will provide a unique learning experience that comes from working on chip design, AI, neuromorphic computing and brain-computer interface.
“The opportunity to discover new deep brain stimulation technologies that will help people suffering from neurological diseases in the future will allow me to continue working in the laboratory and help advance knowledge in this area,” said Jacob Jackson, a biomedical engineering major who conducts research in the U lab. He plans to begin graduate work at Michigan Tech in the fall. “I’m really enjoying neural engineering research so I knew it was the right path for me,” he said.
Michigan Technological University is a public research university founded in 1885 in Houghton, Michigan, with more than 7,000 students from 55 countries around the world. Consistently ranked among the best universities in the country for return on investment, Michigan Technological University offers more than 120 undergraduate and graduate programs in science and technology, engineering, computing, forestry, business and economics, health professions, humanities, mathematics, social sciences, and Art. The rural campus is located miles from Lake Superior in Michigan’s Upper Peninsula, offering year-round opportunities for outdoor adventure.
[ad_2]
Source link
