Enhanced Deep Transfer Learning Model based on Spatial-Temporal driven Scalograms for Precise Decoding of Motor Intent in Stroke Survivors.

Intelligent rehabilitation robotics (RR) have been proposed in recent years to aid post-stroke survivors recover their lost limb functions. However, a large proportion of these robotic systems operate in a passive mode that restricts users to predefined trajectories that rarely align with their intended limb movements, precluding full functional recovery. To address this issue, an efficient Transfer Learning based Convolutional Neural Network (TL-CNN) model is proposed to decode post-stroke patients' motion intentions toward realizing dexterously active robotic training during rehabilitation. For the first time, we use Spatial-Temporal Descriptor based Continuous Wavelet Transform (STD-CWT) as input to TL-CNN to optimally decode limb movement intent patterns. We evaluated the STD-CWT method on three distinct wavelets including the Morse, Amor, and Bump, and compared their decoding outcomes with those of the commonly adopted CWT technique under similar experimental conditions. We then validated the method using electromyogram signals of five stroke survivors who performed twenty-one distinct motor tasks. The results showed that the proposed technique recorded a significantly higher (p<0.05) decoding accuracy and faster convergence compared to the common method. Our method equally recorded obvious class separability for individual motor tasks across subjects. The findings suggest that the STD-CWT Scalograms have the potential for robust decoding of motor intention and could facilitate intuitive and active motor training in stroke RR.Clinical Relevance- The study demonstrated the potential of Spatial Temporal based Scalograms in aiding precise and robust decoding of multi-class motor tasks, upon which dexterously active rehabilitation robotic training for full motor function restoration could be realized.

An Attention-based Bidirectional LSTM Model for Continuous Cross-Subject Estimation of Knee Joint Angle during Running from sEMG Signals.

Accurate and robust estimation of joint kinematics via surface electromyogram (sEMG) signals provides a human-machine interaction (HMI)-based method that can be used to adequately control rehabilitation robots while performing complex movements, such as running, for motor function restoration in affected individuals. To this end, this paper proposes a deep learning-based model (AM-BiLSTM) that integrates a bidirectional long short-term memory (BiLSTM) network and an attention mechanism (AM) for robust estimation of joint kinematics. The proposed model was appraised using knee joint kinematic and sEMG signals collected from fourteen subjects who performed running at the speed of 2 m/s. The proposed model's generalizability was tested for both within- and cross-subject scenarios and compared with long short-term memory (LSTM) and multi-layer perceptron (MLP) networks in terms of normalized root-mean-square error and correlation coefficient metrics. Based on the statistical tests, the proposed AM-BiLSTM model significantly outperformed the LSTM and MLP methods in both within- and cross-subject scenarios (p<0.05) and achieved state-of-the-art performance.Clinical Relevance- The promising results of this study suggest that the AM-BiLSTM model has the potential for continuous cross-subject estimation of lower limb kinematics during running, which can be used to control sEMG-driven exoskeleton robots oriented towards rehabilitation training.

Estimation of Lower Limb Kinematics during Squat Task in Different Loading Using sEMG Activity and Deep Recurrent Neural Networks.

The aim of the present study was to predict the kinematics of the knee and the ankle joints during a squat training task of different intensities. Lower limb surface electromyographic (sEMG) signals and the 3-D kinematics of lower extremity joints were recorded from 19 body builders during squat training at four loading conditions. A long-short term memory (LSTM) was used to estimate the kinematics of the knee and the ankle joints. The accuracy, in terms root-mean-square error (RMSE) metric, of the LSTM network for the knee and ankle joints were 6.774 ± 1.197 and 6.961 ± 1.200, respectively. The LSTM network with inputs processed by cross-correlation (CC) method showed 3.8% and 4.7% better performance in the knee and ankle joints, respectively, compared to when the CC method was not used. Our results showed that in the prediction, regardless of the intensity of movement and inter-subject variability, an off-the-shelf LSTM decoder outperforms conventional fully connected neural networks.