How sensors can prevent dangerous freezing of gait in Parkinson’s patients
Imagine you’re walking along and suddenly your feet are stuck to the ground and won’t budge — as if the unconscious rhythm in your walk that normally is always there suddenly vanished. This is the frightening reality people with advanced Parkinson’s often find themselves in.
But, there is hope. Sensors combined with machine learning algorithms and audio feedback have the potential to detect this dangerous freezing of gait and unfreeze these patients, so they can live more independent lives.
If you’re not familiar, Parkinson’s is a neurodegenerative disorder. People with Parkinson’s typically have trouble with movement. In earlier stages of a the disease, a person may exhibit a slow, shuffling gait. But in more advanced stages, a person might suddenly freeze in the middle of a walk.
Freezing of gait occurs most often in walking transitions, like when a person is starting, stopping or changing directions. It’s terrifying, because even though your feet come to an absolute halt, the upper part of the body continues to lunge forward, causing the person to lose balance or even topple over.
With an increased risk of falls, these patients may end up spending more time in wheelchairs or ultimately wind up in nursing homes.
Fortunately, there are ways to release a person from a freeze. If patient hears a rhythmic signal, they can borrow that rhythm to continue walking again. But since continuous rhythmic cueing wears off over time, it is important to use it only when a disruption in gait occurs.
Several studies now underway are looking at how sensors (specifically gyroscopes and accelerometers), worn on various parts of the body, such as the feet, arms, or wrists — can be used to detect freezing of gait, so that perfectly time auditory signal can unfreeze the person.
CuPid, a Tel Aviv company is testing a smartphone app that harnesses sensors worn on the shoes to bring Parkinson’s patients out of a freeze. When the app detects deviations from a pre-set norm, an audio message alerts the patient to change his or her walking pattern.
While the aim of the CuPid project is to “rehabilitate” the patient, so they learn how to change their behavior, other studies like this one focus on providing a rhythmic signal to stimulate patients once they come into a freeze to start walking again.
Whatever the approach, the key to the success of these devices are machine learning algorithms sophisticated enough to pick up subtle changes in motion on a regular, reliable basis. With the right machine algorithms in place, systems like these may one day become standard assist devices for people living with Parkinson’s.