Traditional methods to evaluate balance often face issues with objectivity, lack comprehensiveness, and are not suitable for remote assessment. They are also costly, requiring specialized equipment and expertise. By utilizing wearable sensors and advanced machine learning algorithms, researchers have introduced a practical and cost-effective solution that captures detailed movement data crucial for assessing balance. This method is more accessible and allows for remote administration, potentially impacting areas such as healthcare, rehabilitation, sports science, and other fields where balance assessment is vital.
Balance can be affected by various factors such as diseases like Parkinson’s disease, nervous system injuries, and aging. Accurate balance assessment is crucial in identifying and managing conditions that affect coordination and stability. It also plays a significant role in fall prevention, understanding movement disorders, and designing suitable therapeutic interventions for various age groups and medical conditions.
However, traditional methods for balance assessment can be subjective, lack comprehensiveness, and are unable to be conducted remotely. Furthermore, these methods depend on expensive specialized equipment that may not be available in all clinical settings and are influenced by the expertise of the clinician, leading to variability in results. There is a critical need for more objective and comprehensive tools for evaluating balance.
Researchers from Florida Atlantic University’s College of Engineering and Computer Science have developed an innovative approach that utilizes wearable sensors and advanced machine learning algorithms to address gaps in balance assessment, establishing a new standard in applying wearable technology and machine learning in healthcare. This method represents a significant advancement in objective balance assessment, particularly for remote monitoring in home-based or nursing care settings, potentially transforming the management of balance disorders.
In their study, researchers used the Modified Clinical Test of Sensory Interaction on Balance (m-CTSIB), a commonly used method in healthcare to assess an individual’s ability to maintain balance under different sensory conditions. Wearable sensors were placed on participants’ ankle, lower back, sternum, wrist, and arm.
Data collected from participants under different sensory conditions of m-CTSIB was processed, and various features were extracted for analysis. Multiple machine learning algorithms were applied including Multiple Linear Regression, Support Vector Regression, and XGBOOST to estimate the m-CTSIB scores based on wearable sensor data. Results published in the journal Frontiers in Digital Health showcased high accuracy and strong correlation with ground truth balance scores, indicating the method’s effectiveness in estimating balance.
“Wearable sensors provide an economical and practical means to capture detailed movement data, essential for balance analysis,” said Behnaz Ghoraani, Ph.D., senior author and associate professor at FAU. “Placed on areas like the lower back and lower limbs, these sensors offer insights into 3D movement dynamics crucial for applications such as assessing fall risk in diverse populations.”
Results of the study highlighted the importance of sensor placement and specific movements in estimating balance, with the XGBOOST model utilizing lumbar sensor data exhibiting exceptional performance. This research suggests that the novel method has the potential to revolutionize balance assessment practices, particularly in situations where traditional methods are impractical or unavailable.
Recognizing the need for advanced tools to capture the intricate effects of different sensory inputs on balance, researchers designed this study to address the gaps in current balance assessments to better understand and manage balance impairments. Utilizing wearables for remote monitoring enables healthcare professionals to assess patients’ balance remotely, offering significant benefits in various healthcare scenarios.
