Objective. Understanding the neurophysiological signals underlying voluntary motor control and decoding them for prosthesis control are among the major challenges in applied neuroscience and bioengineering. Usually, information from the electrical activity of residual forearm muscles (i.e. the electromyogram, EMG) is used to control different functions of a prosthesis. Noteworthy, forearm EMG patterns at the onset of a contraction (transient phase) have shown to contain predictive information about upcoming grasps. However, decoding this information for the estimation of grasp force (GF) was so far overlooked. Approach. High density-EMG signals (192 channels) were recorded from twelve participants performing a pick-and-lift task. The final GF was estimated offline using linear regressors, with four subsets of channels and ten features obtained using three channels-features selection methods. Two different evaluation metrics (absolute error and R 2), complemented with statistical analysis, were used to select the optimal configuration of the parameters. Different windows of data starting at the GF onset were compared to determine the time at which the GF can be ascertained from the EMG signals. Main results. The prediction accuracy improved by increasing the window length from the moment of the onset and kept improving until the steady state at which a plateau of performances was reached. With our methodology, estimations of the GF through 16 EMG channels reached an absolute error of 2.52% the maximum voluntary force using only transient information and 1.99% with the first 500 ms of data following the onset. Significance. The final GF estimation from transient EMG was comparable to the one obtained using steady state data, confirming our hypothesis that the transient phase contains information about the final GF. This result paves the way to fast online myoelectric controllers capable of decoding grasp strength from the very early portion of the EMG signal.

No abstract

We present the SurfacE Electromyographic with hanD kinematicS (SEEDS) database. It contains electromyographic (EMG) signals and hand kinematics recorded from the forearm muscles of 25 non-disabled subjects while performing 13 different movements at normal and slow-paced speeds. EMG signals were recorded with a high-density 126-channel array centered on the extrinsic flexors of the fingers and 8 further electrodes placed on the extrinsic extensor muscles. A data-glove was used to record 18 angles from the joints of the wrist and fingers. The correct synchronisation of the data-glove and the EMG was ascertained and the resulting data were further validated by implementing a simple classification of the movements. These data can be used to test experimental hypotheses regarding EMG and hand kinematics. Our database allows for the extraction of the neural drive as well as performing electrode selection from the high-density EMG signals. Moreover, the hand kinematic signals allow the development of proportional methods of control of the hand in addition to the more traditional movement classification approaches.

The force applied with a prosthetic device is fundamental for the correct handling of objects in daily tasks. However, it is also a factor that normally gets relegated to a secondary plane, as researchers mainly focus on decoding the users intent in terms of movements to be performed. Continuous estimates of the grasp force from the electromyographic (EMG) signals were proposed in the past. As motor actions are preplanned in humans, we hypothesized that it would be possible to decode the intended grasp force from the transient state of the EMG signal. We tested this hypothesis by using features extracted from surface HD-EMG recordings from forearm muscles, classified using artificial neural networks. Data from 6 able-bodied subjects were collected. They were trained and tested at segments of 120 ms with 20 ms overlap, starting 1 s before and ending 0.5 s after the detection of the onset with different subsets of channels. The results obtained showed that the transient phase contains information about the target grasp force, achieving predictions of 2.62 % MVC average absolute errors within 430 ms from the onset of the EMG.