Unsupervised neural decoding for concurrent and continuous multi-finger force prediction
Unsupervised neural decoding for concurrent and continuous multi-finger force prediction
复制标题
用于并发和连续多手指力预测的无监督神经解码
DOI：
10.1016/j.compbiomed.2024.108384
复制
发表时间：
2024
影响因子：
7.7
通讯作者：
Hu, Xiaogang
中科院分区：
文献类型：
--
作者：
Meng, Long;Hu, Xiaogang
多指力的可靠预测对于神经-机器接口至关重要。为了准确预测运动输出，各种神经解码方法取得了长足的进步。然而，大多数神经解码方法都是以监督的方式进行的，即需要手指的力量来进行模型训练，这可能不适合在某些情况下，特别是在涉及截肢个体的场景中。为了解决这个问题，我们开发了一种无监督的神经解码方法，利用脊髓运动神经元放电信息来预测多指力。当受试者进行单指和多指等距伸展任务时，我们获得了手指伸肌的高密度表面肌电图（sEMG）信号。我们首先从单指任务的表面肌电信号中提取运动单元（mu）。由于不可避免的手指肌肉共同激活，控制非目标手指的mu也可以被招募。为了确保准确的手指力量预测，这些mu需要被梳理出来。为此，我们基于动态时间弯曲技术测量的mu间距离对分解后的mu进行聚类，然后使用平均发射速率或发射速率相位幅度对mu进行标记。我们合并了与同一目标手指相关的聚类最小值，并根据保留的最小值的一致性分配权重。结果表明，与监督神经解码方法和传统的表面肌电信号振幅方法相比，我们的方法可以实现更高的r2（0.77±0.036 vs 0.71±0.11 vs 0.61±0.09）和更低的均方根误差（5.16±0.58% vs 5.88±1.34% vs 7.56±1.60% MVC）。我们的研究结果可以为开发准确和鲁棒的神经机器接口铺平道路，这可以显着增强在不同环境下人机交互的体验。
Reliable prediction of multi-finger forces is crucial for neural-machine interfaces. Various neural decoding methods have progressed substantially for accurate motor output predictions. However, most neural decoding methods are performed in a supervised manner, ie, the finger forces are needed for model training, which may not be suitable in certain contexts, especially in scenarios involving individuals with an arm amputation. To address this issue, we developed an unsupervised neural decoding approach to predict multi-finger forces using spinal motoneuron firing information. We acquired high-density surface electromyogram (sEMG) signals of the finger extensor muscle when subjects performed single-finger and multi-finger tasks of isometric extensions. We first extracted motor units (MUs) from sEMG signals of the single-finger tasks. Because of inevitable finger muscle co-activation, MUs controlling the non-targeted fingers can also be recruited. To ensure an accurate finger force prediction, these MUs need to be teased out. To this end, we clustered the decomposed MUs based on inter-MU distances measured by the dynamic time warping technique, and we then labeled the MUs using the mean firing rate or the firing rate phase amplitude. We merged the clustered MUs related to the same target finger and assigned weights based on the consistency of the MUs being retained. As a result, compared with the supervised neural decoding approach and the conventional sEMG amplitude approach, our new approach can achieve a higher R 2 (0.77±0.036 vs. 0.71±0.11 vs. 0.61±0.09) and a lower root mean square error (5.16±0.58% MVC vs. 5.88±1.34% MVC vs. 7.56±1.60% MVC). Our findings can pave the way for the development of accurate and robust neural-machine interfaces, which can significantly enhance the experience during human-robotic hand interactions in diverse contexts.
登录
查看更多内容
影响因子：
4
作者：
Zheng，Yang;Hu，Xiaogang
通讯作者：
Hu，Xiaogang
影响因子：
4.6
作者：
Downey JE;Brane L;Gaunt RA;Tyler-Kabara EC;Boninger ML;Collinger JL
通讯作者：
Collinger JL
影响因子：
2.5
作者：
Aggarwal, Vikram;Mollazadeh, Mohsen;Thakor, Nitish V.
通讯作者：
Thakor, Nitish V.
DOI：
10.1007/s12541-012-0107-2
发表时间：
2012-05-01
影响因子：
1.9
作者：
Heo, Piwon;Gu, Gwang Min;Kim, Jung
通讯作者：
Kim, Jung
影响因子：
7.7
作者：
Dai, Chenyun;Hu, Xiaogang
通讯作者：
Hu, Xiaogang