It may be a good idea to start monitoring your steps, at least their duration. The distance between your steps can reveal early signs of various neurological diseases and allow you to monitor their worsening. Step length usually decreases with age and also in people with neurological disorders.
Researchers from Tel Aviv University (TAU) and Tel Aviv Sourasky Medical Center (TASMC) conducted an international multidisciplinary study in which an innovative machine learning-based model was developed to accurately estimate step length.
Four times more accurate than conventional methods
Their algorithm converts data from a small, lightweight, waterproof wearable sensor attached to the lower back, which provides an accurate estimate of the length of each step. It is about four times more accurate than conventional biomechanical models.
Previous studies have looked at wearable devices based on sensors called inertial measurement units (IMUs) to assess step length, but these experiments were conducted with devices that were uncomfortable to wear, sometimes requiring the use of multiple sensors simultaneously. They were also conducted only with healthy people who had no difficulty walking, and were based on a small sample size that made it difficult to generalize.
“Step length is a sensitive, noninvasive measure of a wide range of conditions associated with aging, cognitive decline, many neurological diseases, multiple sclerosis, Parkinson’s disease, and Alzheimer’s disease, as well as acute cardiovascular disease and stroke. Our model enables continuous monitoring of this key aspect of a patient’s condition,” the researchers wrote in their paper, which was just published in the journal Digital Medicine under the title “A wearable sensor and machine learning estimate step length in older adults and patients with neurological disorders.” An accurate and diverse database of 83,569 steps was collected.
The new model can be integrated into a wearable device attached to the lower back with an adhesive tape and allows continuous monitoring of the patient’s steps in everyday life. Existing conventional measuring devices are stationary and bulky and are only found in specialized clinics and laboratories. The new model allows for precise measurements in the patient’s natural environment throughout the day, using a wearable sensor, the researchers said.
The study was conducted by Assaf Zadka, a graduate student at the Department of Biomedical Engineering at Tau University, Professor Jeffrey Hausdorff from the Department of Physiotherapy at the Faculty of Medical and Health Sciences, Sagol School of Neuroscience at Tau University and the Department of Neurology at TASMC, and Professor Neta Rabin from the Department of Industrial Engineering at the Fleischman Faculty of Engineering at Tau University. Also participating in the study were Eran Gazit from TASMC and Professor Anat Mirelman from Tau University and TASMC, as well as researchers from Belgium, England, Italy, the Netherlands and the United States.
Hausdorff, a neurologist specializing in gait, aging and falls, explained that “step length is a highly sensitive, noninvasive measure for assessing a wide variety of conditions and diseases, including aging, deterioration resulting from neurological and neurodegenerative diseases, and cognitive decline.
“Today, it is common to measure stride length using camera-based devices and measuring devices such as force-sensitive treadmills found only in specialized laboratories and clinics,” he said. “While these tests are accurate, they provide only a snapshot of a person’s gait that likely does not fully reflect actual real-world functioning. Daily walking can be influenced by a patient’s fatigue level, mood, and medications. Continuous 24/7 monitoring, such as that made possible by this new stride length model, can capture this real-world walking behavior.”
Translating steps into disease detection
Rabin, a machine learning expert, added: “We wanted to solve the problem by exploiting IMU systems, lightweight and relatively cheap sensors that are currently installed in all phones and smartwatches and that measure parameters associated with walking. The goal was to develop an algorithm capable of translating IMU data into an accurate assessment of step length and integrating it into a wearable and comfortable device.”
The team successfully used IMU-based gait and step length data, conventionally measured in a previous study of 472 people with Parkinson’s disease, mild cognitive impairment, or multiple sclerosis, as well as healthy older and younger people. They used this data to train a number of computer models that translated the IMU data into an estimate of step length. To test the accuracy of the models, they determined how well the different models could analyze new data that had not been used in the training process.
“We found that the model called XGBoost is the most accurate and is 3.5 times more accurate than the most advanced biomechanical model currently used to estimate step length,” Zadka said.
“For a single step, the average error of our model was six centimeters, compared to 21 centimeters predicted by the conventional model. When we evaluated an average of 10 steps, we arrived at an error of less than five centimeters – a threshold known in the professional literature as “the minimal difference of clinical importance” – which allows to identify a significant improvement or decrease in the subject’s condition,” he said.
“Our model is powerful and reliable, and can be used to analyze sensor data from patients, some of whom have difficulty walking, who were not included in the original training set.”
Hausdorff concluded that “in our research, we collaborated with researchers from various fields around the world and the multidisciplinary effort led to promising results. We developed a machine learning model that can be integrated into a wearable, easy-to-use sensor and provides an accurate estimate of the patient’s step length in everyday life.
“The data collected in this way allows for continuous, remote and long-term monitoring of the patient’s health status and can also be used in clinical trials to examine the effectiveness of drugs,” he said. “Based on our encouraging results, we are investigating the possibility of developing similar models based on data from smartwatch sensors, which would further improve the subject’s comfort.”