In recent years, deep learning algorithms have become increasingly more prominent for their unparalleled ability to automatically learn discriminant features from large amounts of data. However, within the field of electromyographybased gesture recognition, deep learning algorithms are seldom employed as they require an unreasonable amount of effort from a single person, to generate tens of thousands of examples.This work's hypothesis is that general, informative features can be learned from the large amounts of data generated by aggregating the signals of multiple users, thus reducing the recording burden while enhancing gesture recognition. Consequently, this paper proposes applying transfer learning on aggregated data from multiple users, while leveraging the capacity of deep learning algorithms to learn discriminant features from large datasets. Two datasets comprised of 19 and 17 able-bodied participants respectively (the first one is employed for pre-training) were recorded for this work, using the Myo Armband. A third Myo Armband dataset was taken from the NinaPro database and is comprised of 10 able-bodied participants. Three different deep learning networks employing three different modalities as input (raw EMG, Spectrograms and Continuous Wavelet Transform (CWT)) are tested on the second and third dataset. The proposed transfer learning scheme is shown to systematically and significantly enhance the performance for all three networks on the two datasets, achieving an offline accuracy of 98.31% for 7 gestures over 17 participants for the CWT-based ConvNet and 68.98% for 18 gestures over 10 participants for the raw EMG-based ConvNet. Finally, a use-case study employing eight able-bodied participants suggests that real-time feedback allows users to adapt their muscle activation strategy which reduces the degradation in accuracy normally experienced over time.

In the realm of surface electromyography (sEMG) gesture recognition, deep learning algorithms are seldom employed. This is due in part to the large quantity of data required for them to train on. Consequently, it would be prohibitively time consuming for a single user to generate a sufficient amount of data for training such algorithms. In this paper, two datasets of 18 and 17 able-bodied participants respectively are recorded using a low-cost, low-sampling rate (200Hz), 8-channel, consumer-grade, dry electrode sEMG device named Myo armband (Thalmic Labs). A convolutional neural network (CNN) is augmented using transfer learning techniques to leverage inter-user data from the first dataset and alleviate the data generation burden imposed on a single individual. The results show that the proposed classifier is robust and precise enough to guide a 6DoF robotic arm (in conjunction with orientation data) with the same speed and precision as with a joystick. Furthermore, the proposed CNN achieves an average accuracy of 97.81% on seven hand/wrist gestures on the 17 participants of the second dataset.

The research in myoelectric control systems primarily focuses on extracting discriminative representations from the electromyographic (EMG) signal by designing handcrafted features. Recently, deep learning techniques have been applied to the challenging task of EMG-based gesture recognition. The adoption of these techniques slowly shifts the focus from feature engineering to feature learning. However, the black-box nature of deep learning makes it hard to understand the type of information learned by the network and how it relates to handcrafted features. Additionally, due to the high variability in EMG recordings between participants, deep features tend to generalize poorly across subjects using standard training methods. Consequently, this work introduces a new multi-domain learning algorithm, named ADANN, which significantly enhances (p = 0.00004) inter-subject classification accuracy by an average of 19.40% compared to standard training.Using ADANN-generated features, the main contribution of this work is to provide the first topological data analysis of EMG-based gesture recognition for the characterisation of the information encoded within a deep network, using handcrafted features as landmarks. This analysis reveals that handcrafted features and the learned features (in the earlier layers) both try to discriminate between all gestures, but do not encode the same information to do so. Furthermore, using convolutional network visualization techniques reveal that learned features tend to ignore the most activated channel during gesture contraction, which is in stark contrast with the prevalence of handcrafted features designed to capture amplitude information. Overall, this work paves the way for hybrid feature sets by providing a clear guideline of complementary information encoded within learned and handcrafted features.

Wearable technology can be employed to elevate the abilities of humans to perform demanding and complex tasks more efficiently. Armbands capable of surface electromyography (sEMG) are attractive and noninvasive devices from which human intent can be derived by leveraging machine learning. However, the sEMG acquisition systems currently available tend to be prohibitively costly for personal use or sacrifice wearability or signal quality to be more affordable. This work introduces the 3DC Armband designed by the Biomedical Microsystems Laboratory in Laval University; a wireless, 10-channel, 1000 sps, dry-electrode, low-cost (∼150 USD) myoelectric armband that also includes a 9-axis inertial measurement unit. The proposed system is compared with the Myo Armband by Thalmic Labs, one of the most popular sEMG acquisition systems. The comparison is made by employing a new offline dataset featuring 22 able-bodied participants performing eleven hand/wrist gestures while wearing the two armbands simultaneously. The 3DC Armband systematically and significantly ( p < 0.05 ) outperforms the Myo Armband, with three different classifiers employing three different input modalities when using ten seconds or more of training data per gesture. This new dataset, alongside the source code, Altium project and 3-D models are made readily available for download within a Github repository.

Surface electromyography (sEMG) provides an intuitive and non-invasive interface from which to control machines. However, preserving the myoelectric control system's performance over multiple days is challenging, due to the transient nature of the signals obtained with this recording technique. In practice, if the system is to remain usable, a time-consuming and periodic recalibration is necessary. In the case where the sEMG interface is employed every few days, the user might need to do this recalibration before every use. Thus, severely limiting the practicality of such a control method. Consequently, this paper proposes tackling the especially challenging task of unsupervised adaptation of sEMG signals, when multiple days have elapsed between each recording, by introducing Self-Calibrating Asynchronous Domain Adversarial Neural Network (SCADANN). SCADANN is compared with two state-of-the-art selfcalibrating algorithms developed specifically for deep learning within the context of EMG-based gesture recognition and three state-of-the-art domain adversarial algorithms. The comparison is made both on an offline and a dynamic dataset (20 participants per dataset), using two different deep network architectures with two different input modalities (temporal-spatial descriptors and spectrograms). Overall, SCADANN is shown to substantially and systematically improves classification performances over no recalibration and obtains the highest average accuracy for all tested cases across all methods.

Within the field of electromyography-based (EMG) gesture recognition, disparities exist between the offline accuracy reported in the literature and the real-time usability of a classifier. This gap mainly stems from two factors: 1) The absence of a controller, making the data collected dissimilar to actual control. 2) The difficulty of including the four main dynamic factors (gesture intensity, limb position, electrode shift, and transient changes in the signal), as including their permutations drastically increases the amount of data to be recorded. Contrarily, online datasets are limited to the exact EMG-based controller used to record them, necessitating the recording of a new dataset for each control method or variant to be tested. Consequently, this paper proposes a new type of dataset to serve as an intermediate between offline and online datasets, by recording the data using a real-time experimental protocol. The protocol, performed in virtual reality, includes the four main dynamic factors and uses an EMG-independent controller to guide movements. This EMG-independent feedback ensures that the user is in-the-loop during recording, while enabling the resulting dynamic dataset to be used as an EMG-based benchmark. The dataset is comprised of 20 able-bodied participants completing three to four sessions over a period of 14 to 21 days. The ability of the dynamic dataset to serve as a benchmark is leveraged to evaluate the impact of different

Changes in motor activity are core symptoms of mood episodes in bipolar disorder. The manic state is characterized by increased variance, augmented complexity and irregular circadian rhythmicity when compared to healthy controls. No previous studies have compared mania to euthymia intra-individually in motor activity. The aim of this study was to characterize differences in motor activity when comparing manic patients to their euthymic selves. Motor activity was collected from 16 bipolar inpatients in mania and remission. 24-h recordings and 2-h time series in the morning and evening were analyzed for mean activity, variability and complexity. Lastly, the recordings were analyzed with the similarity graph algorithm and graph theory concepts such as edges, bridges, connected components and cliques. The similarity graph measures fluctuations in activity reasonably comparable to both variability and complexity measures. However, direct comparisons are difficult as most graph measures reveal variability in constricted time windows. Compared to sample entropy, the similarity graph is less sensitive to outliers. The little-understood estimate Bridges is possibly revealing underlying dynamics in the time series. When compared to euthymia, over the duration of approximately one circadian cycle, the manic state presented reduced variability, displayed by decreased standard deviation (p = 0.013) and augmented complexity shown by increased sample entropy (p = 0.025). During mania there were also fewer edges (p = 0.039) and more bridges (p = 0.026). Similar significant changes in variability and complexity were observed in the 2-h morning and evening sequences, mainly in the estimates of the similarity graph algorithm. Finally, augmented complexity was present in morning samples during mania, displayed by increased sample entropy (p = 0.015). In conclusion, the motor activity of mania is characterized by altered complexity and variability when compared within-subject to euthymia.

No abstract