Posts Tagged algorithm
“The hope is to make seizures less like earthquakes, which can strike without warning, and more like hurricanes, where you have enough advance warning to seek safety,” Dr. Levin Kuhlmann from the University of Melbourne’s Graeme Clarke Institute and St. Vincent’s Hospital said.
“Accurate seizure prediction will transform epilepsy management by offering early warnings to patients or triggering interventions.”
Published on Thursday, the research began with a world-wide mathematical data science challenge in 2016.
Contestants were tasked with designing algorithms that could effectively distinguish between a pre-seizure and an inter-seizure.
With more than 646 participants and 478 teams, the most accurate algorithms were tested on patients with the lowest seizure prediction rates.
“Our evaluation revealed on average a 90-percent improvement in seizure prediction performance, compared to previous results,” Kuhlmann said.
Effecting over 65 million people around the world, epilepsy can be “highly different” among individual sufferers.
“Results showed different algorithms performed best for different patients, supporting the use of patient-specific algorithms and long-term monitoring,” Kuhlmann said.
Encouraged by the positive findings, researchers have now developed an algorithm and data sharing website called Epilepsy Ecosystem, to encourage others to share their work and help build on the project.
“It’s about bringing together the world’s best data scientists and pooling the greatest algorithms to advance epilepsy research,” Kuhlmann said.
“Our results highlight the benefit of crowdsourcing an army of algorithms that can be trained for each patient and the best algorithm chosen for prospective, real-time seizure prediction.”
In the United States, there are approximately 17,000 new cases of spinal cord injury (SCI) every year. Of these, 20 percent result in complete paraplegia (paralysis of the legs and lower half of body) and over 13 percent result in tetraplegia (paralysis of all four limbs).
But SCI is not the only reason that people experience this type of disability. Stroke, multiple sclerosis, cerebral palsy, and a range of other neurological disorders can all lead to paralysis. In fact, a recent survey estimated that in the U.S., almost 5.4 million people live with paralysis, with stroke being the leading cause of this disability.
Now, researchers from the National Centre of Competence in Research Robotics at École Polytechnique Fédérale de Lausanne (EPFL), and at the Lausanne University Hospital in Switzerland, have come up with a groundbreaking technology that may help these patients to regain their locomotor skills.
The scientists came up with an algorithm that helps a robotic harness to facilitate the movements of the patients, thus enabling them to move naturally.
The new research has been published in the journal Science Translational Medicine, and the first author of the study is Jean-Baptiste Mignardot.
Helping people to walk again
Current rehabilitation technologies for people with motor disabilities as a result of SCI or stroke involve walking on a treadmill, with the upper torso being supported by an apparatus. But existing technologies are either too rigid or do not allow the patients to move naturally in all directions.
As the authors of the new study explain, the challenge of locomotor rehabilitation resides in helping the nervous system to “relearn” the right movements. This is difficult due to the loss of muscle mass in the patients, as well as to the neurological wiring that has “forgotten” correct posture.
In order to overcome these obstacles and promote natural walking, Mignardot and colleagues designed an algorithm that coordinates with a robotic rehabilitation harness. The team tested the algorithm in more than 30 patients. The “smart walk assist” markedly and immediately improved the patients’ locomotor abilities.
This mobile harness, which is attached to the ceiling, enables patients to walk. This video shows how it works:
Additionally, after only 1 hour of training with the harness and algorithm, the “unsupported walking ability” of five of the patients improved considerably. By contrast, 1 hour on a conventional treadmill did not improve gait.
The researchers developed the so-called gravity-assist algorithm after carefully monitoring the movements of the patients and considering parameters such as “leg movement, length of stride, and muscle activity.”
As the authors explain, based on these measurements, the algorithm identifies the forces that must be applied to the upper half of the body in order to allow for natural walking.
The smart walk assist is an innovative body-weight support system because it manages to resist the force of gravity and push the patient back and forth, to the left and to the right, or in more of these directions at once, which recreates a natural gait and movement that the patients need in their day to day lives.
Grégoire Courtine, a neuroscientist at EPFL and the Lausanne University Hospital, comments on the significance of the findings, saying, “I expect that this platform will play a critical role in the rehabilitation of walking for people with neurological disorders.”
“This is a smart, discreet, and efficient assistance that will aid rehabilitation of many persons with neurological disorders.”
Prof. Jocelyne Bloch, Department of Neurosurgery, Lausanne University Hospital
[ARTICLE] Smart control for functional electrical stimulation with optimal pulse intensity – Full Text
Transcutaneous electrical stimulation is a common treatment option for patients suffering from spinal cord injury or stroke. Two major difficulties arise when employing electrical stimulation in patients: Accurate stimulation electrode placement and configuration of optimal stimulation parameters. Optimizing the stimulation parameters has the advantage to reduce muscle fatigue after repetitive stimulation. Here we present a newly developed system which is able to automatically find the optimal individual stimulation intensity by varying the pulse length. The effectiveness is measured with flex sensors. By adapting the stimulation parameters, the effect of muscle fatigue can be compensated, allowing for a more stable movement upon stimulation over time.
Functional electrical stimulation (FES) has been used to help patients who suffer from stroke or spinal chord injury for many years now . FES is able to support patients in activities of daily living, like walking or grasping . Performing electrical stimulation in most cases arises two core questions: Where the optimal placement of the stimulation electrodes is and which stimulation intensity should be applied . Addressing the former question, prior studies have investigated the optimal electrode placement by using electrode arrays . Addressing the latter is an equivalently complex question, as parameters vary from patient to patient. Providing neither too strong, nor too weak stimulation intensity is crucial, as too weak stimulation leads to an insufficiently opened hand or incorrect step while walking. Too strong stimulation exhaust the muscles quickly without any further benefit, leading to muscle fatigue .
Here we present a newly developed system which is able to increase the intensity of stimulation in a stepwise manner until the optimal point is reached. We demonstrate it’s use in a chronic stroke patient with hand paresis, focusing the opening of the hand. In order to cancel out fatigue, we regulate the intensity, allowing a stable opening of the affected hand.