Improving Tongue Command Accuracy: Unlocking the Power of Electrotactile Feedback Training.

People with spinal cord injury or neurological disorders frequently require aid in performing daily tasks. Utilizing hand-free assistive technologies (ATs), particularly tongue-controlled ATs, may offer a feasible solution as the tongue is controlled by a cranial nerve and remains functional in the presence of spinal cord injury. However, existing intra-oral ATs require a significant level of training to accurately issuing these commands. To minimize the training process, we have designed intuitive tongue commands for our Multifunctional intraORal Assistive technology (MORA). Our prior works demonstrated that electrotactile feedback outperformed visual feedback in tasks involving tongue motor learning. In this study, we implement electrical stimulation (E-stim) as electrotactile feedback on the tongue to teach new tongue commands of MORA, and quantitatively analyze the efficacy of the electrotactile feedback in command accuracy and precision. The random command task was adopted to evaluate tongue command accuracy with 14 healthy participants. The average sensors contacted per trial dropped significantly from 1.57 ± 0.15 to 1.16 ± 0.05 with electrotactile feedback. After training with electrotactile feedback, 83% of the trials were completed with only one command having been activated. These results suggest that E-stim enhanced both the accuracy and precision of subjects' tongue command training. The results of this study pave the way for the implementation of electrotactile feedback as an accurate and precise command training technique for MORA.

A CNN-based method to reconstruct 3-D spine surfaces from US images in vivo.

Three-dimensional (3-D) reconstruction of the spine surface is of strong clinical relevance for the diagnosis and prognosis of spine disorders and intra-operative image guidance. In this paper, we report a new technique to reconstruct lumbar spine surfaces in 3-D from non-invasive ultrasound (US) images acquired in free-hand mode. US images randomly sampled from in vivo scans of 9 rabbits were used to train a U-net convolutional neural network (CNN). More specifically, a late fusion (LF)-based U-net trained jointly on B-mode and shadow-enhanced B-mode images was generated by fusing two individual U-nets and expanding the set of trainable parameters to around twice the capacity of a basic U-net. This U-net was then applied to predict spine surface labels in in vivo images obtained from another rabbit, which were then used for 3-D spine surface reconstruction. The underlying pose of the transducer during the scan was estimated by registering stacks of US images to a geometrical model derived from corresponding CT data and used to align detected surface points. Final performance of the reconstruction method was assessed by computing the mean absolute error (MAE) between pairs of spine surface points detected from US and CT and by counting the total number of surface points detected from US. Comparison was made between the LF-based U-net and a previously developed phase symmetry (PS)-based method. Using the LF-based U-net, the averaged number of US surface points across the lumbar region increased by 21.61% and MAE reduced by 26.28% relative to the PS-based method. The overall MAE (in mm) was 0.24±0.29. Based on these results, we conclude that: 1) the proposed U-net can detect the spine posterior arch with low MAE and large number of US surface points and 2) the newly proposed reconstruction framework may complement and, under certain circumstances, be used without the aid of an external tracking system in intra-operative spine applications.

A Computational Internal Model to Quantify the Effect of Sensorimotor Augmentation on Motor Output.

The aging process, as well as neurological disorders, causes a decline in sensorimotor functions, which can often bring degraded motor output. As a means of compensation for such sensorimotor deficiencies, sensorimotor augmentation has been actively investigated. Consequently, exoskeleton devices or functional electrical stimulation could augment the muscle activity, while textured surfaces or electrical nerve stimulations could augment the sensory feedback. However, it is not easy to precisely anticipate the effects of specific augmentation because sensory feedback and motor output interact with each other as a closed-loop operation via the central and peripheral nervous systems. A computational internal model can play a crucial role in anticipating such an effect of augmentation therapy on the motor outcome. Still, no existing internal sensorimotor loop model has been represented in a complete computational form facilitating the anticipation. This paper presents such a computational internal model, including numerical values representing the effect of sensorimotor augmentation. With the existing experimental results, the model performance was evaluated indirectly. The change of sensory gain affects motor output inversely, while the change of motor gain did not change or minimally affects the motor output.Clinical Relevance- The presented computational internal model will provide a simple and easy tool for clinicians to design therapeutic intervention using sensorimotor augmentation.