Neural Engineering With Brain-Machine-Interfaces Holds Much Promise

We are approaching the day when people paralyzed by disease or other circumstance will easily and efficiently regain the experience of movement, thanks to revolutionary deep brain stimulation and its life changing effect on human health and quality of life.

Imagine using thoughts alone to control and play video games on your computer or even smartphone. It sounds futuristic, but innovative companies like NeuroSky are already marketing the technology. A wrap-around headset with sensors (electrodes) records electrical activity within the brain, and a chip converts the recorded biological activity into a digital signal that can travel wirelessly to a computer.

Known as brain-computer or brain-machine interfaces (BCI/BMI), their application goes beyond video games and smart phones; the same technology is used to diagnose seizures and to restore lost motor function in paralyzed individuals. The BMI can restore the lost connection between the nervous system and muscles, digitizing the electrical activity of the brain. A patient’s thoughts can then be used to restore movement, by operating a robotic arm or leg, or to restore communication, by moving a cursor on a computer screen to select words.

Like technologies that record electrical activity in the brain, there are technologies that stimulate the brain. These have tremendous potential for increasing the quality of life in patients with debilitating neurological conditions, such as Parkinson’s disease, chronic pain, epilepsy, and drug-resistant psychiatric disorders, including depression. Neural engineers are working to develop better and more effective stimulation devices that interact with the nervous system.

“In our lab one of the therapies that we work with is deep brain stimulation (DBS),” says Bryan Howell, a PhD candidate in the department of biomedical engineering at Duke University in Durham, North Carolina, US. “After first line drug treatments are no longer effective in treating Parkinson’s disease, DBS provides patients with another option.”

Parkinson’s disease (PD) is a neurodegenerative disease in which dopamine-producing neurons in small, movement-associated regions of the brain slowly die over time. These dopamine neurons are crucial to movement regulation, thus those afflicted with PD have difficulty initiating and terminating voluntary movements. With this loss comes a predictable set of symptoms, including involuntary trembling limbs (tremor), muscle stiffness (rigidity), lack of or slowness in initiating movements (akinesia), and postural instability.

“DBS involves implanting an electrode within a target region of the brain and using generated electrical fields to modulate abnormal activity of neurons associated with symptoms,” Howell says. “The electrode is sending current pulses at a constant frequency that can transform someone from uncontrollable tremor to fluid movement. Within a couple of days, 90-95% of visible symptoms are eliminated.”

DBS batteries last only three to five years. However, PD is a chronic, progressive disease and patients are looking not only at significant initial startup costs of around $25,000-$30,000 – for the device and the surgery to implant it – but also additional battery costs throughout the therapy. Battery replacement requires additional surgery, which carries its own costs and risks, such as infection.

Implantation is not always perfect, either; the electrode can move throughout the course of therapy, and, if a surgeon is not precise in the placement of the electrode, side effects can develop – pain in certain areas or uncontrollable movement in one arm or leg – which may or may not be tolerable.

Although DBS research is still in its infancy, neural engineers at Duke are working toward improving DBS power efficiency and selectivity by optimizing certain parts of the device such as the electrode used to deliver current, and the waveform and pattern of current pulse. “Instead of just implanting these devices and hoping for the best, we can model the device and the brain regions in which it is used,” says Howell.

He spoke about the computational needs for the growing field of neural engineering at the grid computing and campus infrastructure workshop held at Duke. “We want to model the portion of the brain where the electrode is implanted and the neural elements that are responding to stimulation,” Howell explains. “These are complex models that don’t have analytic solutions; they require a lot of compute time to solve.”

Howell’s work, showing the power of model-based optimization of electrodes, will be published in late November following the 6^th International IEEE EMBS Neural Engineering Conference. Compared with clinical electrodes, those optimized consumed 48-67% less power. The optimized electrodes also showed similar gains in selectivity, reducing non-targeted elements from 34-71% to 1-36% while activating 100% of the targeted elements.

“Distributed computing is critical to DBS research. With a combination of campus grid and Open Science Grid resources, simulations that would take on the order of months to years to solve using a single desktop computer can be solved in a matter of days,” Howell says. “Now we’re able to submit hundreds of thousands of jobs at a time – but, even so, the models we are using are still relatively simple. Going forward we will be paying more attention to the complex geometry of, and the neural elements in, each brain region – all of which will require even greater compute times.”

As a new discipline, neural engineering has neither direction nor definition, but its future holds much promise. Only imagination and compute resources limit research into design problems at the interface of living neural tissue and non-living constructs. We are approaching the day when people paralyzed by disease or other circumstances will easily and efficiently regain the experience of movement, improved health and quality of life. -- by Amber Harmon, © i SGTW