一种深度学习算法从行动受限疾病的现实生活步行片段中检测步态事件的生态效度。
Ecological validity of a deep learning algorithm to detect gait events from real-life walking bouts in mobility-limiting diseases.
作者信息
Romijnders Robbin, Salis Francesca, Hansen Clint, Küderle Arne, Paraschiv-Ionescu Anisoara, Cereatti Andrea, Alcock Lisa, Aminian Kamiar, Becker Clemens, Bertuletti Stefano, Bonci Tecla, Brown Philip, Buckley Ellen, Cantu Alma, Carsin Anne-Elie, Caruso Marco, Caulfield Brian, Chiari Lorenzo, D'Ascanio Ilaria, Del Din Silvia, Eskofier Björn, Fernstad Sara Johansson, Fröhlich Marceli Stanislaw, Garcia Aymerich Judith, Gazit Eran, Hausdorff Jeffrey M, Hiden Hugo, Hume Emily, Keogh Alison, Kirk Cameron, Kluge Felix, Koch Sarah, Mazzà Claudia, Megaritis Dimitrios, Micó-Amigo Encarna, Müller Arne, Palmerini Luca, Rochester Lynn, Schwickert Lars, Scott Kirsty, Sharrack Basil, Singleton David, Soltani Abolfazl, Ullrich Martin, Vereijken Beatrix, Vogiatzis Ioannis, Yarnall Alison, Schmidt Gerhard, Maetzler Walter
机构信息
Digital Signal Processing and System Theory, Electrical and Information Engineering, Faculty of Engineering, Kiel University, Kiel, Germany.
Arbeitsgruppe Neurogeriatrie, Department of Neurology, Universitätsklinikum Schleswig-Holstein, Kiel, Germany.
出版信息
Front Neurol. 2023 Oct 16;14:1247532. doi: 10.3389/fneur.2023.1247532. eCollection 2023.
INTRODUCTION
The clinical assessment of mobility, and walking specifically, is still mainly based on functional tests that lack ecological validity. Thanks to inertial measurement units (IMUs), gait analysis is shifting to unsupervised monitoring in naturalistic and unconstrained settings. However, the extraction of clinically relevant gait parameters from IMU data often depends on heuristics-based algorithms that rely on empirically determined thresholds. These were mainly validated on small cohorts in supervised settings.
METHODS
Here, a deep learning (DL) algorithm was developed and validated for gait event detection in a heterogeneous population of different mobility-limiting disease cohorts and a cohort of healthy adults. Participants wore pressure insoles and IMUs on both feet for 2.5 h in their habitual environment. The raw accelerometer and gyroscope data from both feet were used as input to a deep convolutional neural network, while reference timings for gait events were based on the combined IMU and pressure insoles data.
RESULTS AND DISCUSSION
The results showed a high-detection performance for initial contacts (ICs) (recall: 98%, precision: 96%) and final contacts (FCs) (recall: 99%, precision: 94%) and a maximum median time error of -0.02 s for ICs and 0.03 s for FCs. Subsequently derived temporal gait parameters were in good agreement with a pressure insoles-based reference with a maximum mean difference of 0.07, -0.07, and <0.01 s for stance, swing, and stride time, respectively. Thus, the DL algorithm is considered successful in detecting gait events in ecologically valid environments across different mobility-limiting diseases.
引言
对活动能力，特别是步行能力的临床评估，目前仍主要基于缺乏生态效度的功能测试。得益于惯性测量单元（IMU），步态分析正转向在自然且无限制的环境中进行无监督监测。然而，从IMU数据中提取临床相关的步态参数通常依赖于基于启发式的算法，这些算法依靠经验确定的阈值。这些算法主要在有监督环境下的小样本队列中得到验证。
方法
在此，开发了一种深度学习（DL）算法，并在不同行动受限疾病队列的异质人群以及一组健康成年人中对步态事件检测进行了验证。参与者在其日常环境中双脚佩戴压力鞋垫和IMU，持续2.5小时。来自双脚的原始加速度计和陀螺仪数据被用作深度卷积神经网络的输入，而步态事件的参考时间则基于IMU和压力鞋垫数据的组合。
结果与讨论
结果显示，初始接触（IC）的检测性能很高（召回率：98%，精确率：96%），最终接触（FC）的检测性能也很高（召回率：99%，精确率：94%），IC的最大中位时间误差为-0.02秒，FC的最大中位时间误差为0.03秒。随后得出的时间步态参数与基于压力鞋垫的参考值高度一致，站立、摆动和步幅时间的最大平均差异分别为0.07、-0.07和<0.01秒。因此，该DL算法被认为在检测不同行动受限疾病的生态有效环境中的步态事件方面是成功的。
Zotero 插件