BACKGROUND: Personal sensing, leveraging data passively and near-continuously collected with wearables from patients in their ecological environment, is a promising paradigm to monitor mood disorders (MDs), a major determinant of the worldwide disease burden. However, collecting and annotating wearable data is resource intensive. Studies of this kind can thus typically afford to recruit only a few dozen patients. This constitutes one of the major obstacles to applying modern supervised machine learning techniques to MD detection.
OBJECTIVE: In this paper, we overcame this data bottleneck and advanced the detection of acute MD episodes from wearables\' data on the back of recent advances in self-supervised learning (SSL). This approach leverages unlabeled data to learn representations during pretraining, subsequently exploited for a supervised task.
METHODS: We collected open access data sets recording with the Empatica E4 wristband spanning different, unrelated to MD monitoring, personal sensing tasks-from emotion recognition in Super Mario players to stress detection in undergraduates-and devised a preprocessing pipeline performing on-/off-body detection, sleep/wake detection, segmentation, and (optionally) feature extraction. With 161 E4-recorded subjects, we introduced E4SelfLearning, the largest-to-date open access collection, and its preprocessing pipeline. We developed a novel E4-tailored transformer (E4mer) architecture, serving as the blueprint for both SSL and fully supervised learning; we assessed whether and under which conditions self-supervised pretraining led to an improvement over fully supervised baselines (ie, the fully supervised E4mer and pre-deep learning algorithms) in detecting acute MD episodes from recording segments taken in 64 (n=32, 50%, acute, n=32, 50%, stable) patients.
RESULTS: SSL significantly outperformed fully supervised pipelines using either our novel E4mer or extreme gradient boosting (XGBoost): n=3353 (81.23%) against n=3110 (75.35%; E4mer) and n=2973 (72.02%; XGBoost) correctly classified recording segments from a total of 4128 segments. SSL performance was strongly associated with the specific surrogate task used for pretraining, as well as with unlabeled data availability.
CONCLUSIONS: We showed that SSL, a paradigm where a model is pretrained on unlabeled data with no need for human annotations before deployment on the supervised target task of interest, helps overcome the annotation bottleneck; the choice of the pretraining surrogate task and the size of unlabeled data for pretraining are key determinants of SSL success. We introduced E4mer, which can be used for SSL, and shared the E4SelfLearning collection, along with its preprocessing pipeline, which can foster and expedite future research into SSL for personal sensing.

BACKGROUND: To develop and validate a deep learning model for automated assessment of endodontic case difficulty from periapical radiographs.
METHODS: A dataset of 1,386 periapical radiographs was compiled from two clinical sites. Two dentists and two endodontists annotated the radiographs for difficulty using the \"simple assessment\" criteria from the American Association of Endodontists\' case difficulty assessment form in the Endocase application. A classification task labeled cases as \"easy\" or \"hard\", while regression predicted overall difficulty scores. Convolutional neural networks (i.e. VGG16, ResNet18, ResNet50, ResNext50, and Inception v2) were used, with a baseline model trained via transfer learning from ImageNet weights. Other models was pre-trained using self-supervised contrastive learning (i.e. BYOL, SimCLR, MoCo, and DINO) on 20,295 unlabeled dental radiographs to learn representation without manual labels. Both models were evaluated using 10-fold cross-validation, with performance compared to seven human examiners (three general dentists and four endodontists) on a hold-out test set.
RESULTS: The baseline VGG16 model attained 87.62% accuracy in classifying difficulty. Self-supervised pretraining did not improve performance. Regression predicted scores with ± 3.21 score error. All models outperformed human raters, with poor inter-examiner reliability.
CONCLUSIONS: This pilot study demonstrated the feasibility of automated endodontic difficulty assessment via deep learning models.

UNASSIGNED: Smartwatches enable continuous and noninvasive time series monitoring of cardiovascular biomarkers like heart rate (from photoplethysmograms), step counter, skin temperature, et cetera; as such, they have promise in assisting in early detection and prevention of cardiovascular disease. Although these biomarkers may not be directly useful to physicians, a machine learning (ML) model could find clinically relevant patterns. Unfortunately, ML models typically need supervised (ie, annotated) data, and labeling of large amounts of continuous data is very labor intensive. Therefore, ML methods that are data efficient, ie, needing a low number of labels, are required to detect potential clinical value in patterns found in wearable data.
UNASSIGNED: The primary study objective of the ME-TIME (Machine Learning Enabled Time Series Analysis in Medicine) study is to design an ML model that can detect atrial fibrillation (AF) and heart failure (HF) from wearable data in a data-efficient manner. To achieve this, self-supervised and weakly supervised learning techniques are used.
UNASSIGNED: Two hundred subjects (100 reference, 50 AF, and 50 HF) are being invited to participate in wearing a Fitbit fitness tracker for 3 months. Interested volunteers are sent a questionnaire to determine their health, in particular cardiovascular health. Volunteers without any (history of) serious illness are assigned to the reference group. Participants with AF and HF are recruited in the Haga teaching hospital in The Hague, The Netherlands.
UNASSIGNED: Enrollment commenced on May 1, 2022, and as of the time of this report, 62 subjects have been included in the study. Preliminary analysis of the data reveals significant inter-subject variability. Notably, we identified heart rate recovery curves and time-delayed correlations between heart rate and step count as potential strong indicators for heart disease.
UNASSIGNED: Using self-supervised and multiple-instance learning techniques, we hypothesize that patterns specific to AF and HF can be found in continuous data obtained from smartwatches.

UNASSIGNED: Deep learning models have shown great success in automating tasks in sleep medicine by learning from carefully annotated electroencephalogram (EEG) data. However, effectively using a large amount of raw EEG data remains a challenge.
UNASSIGNED: In this study, we aim to learn robust vector representations from massive unlabeled EEG signals, such that the learned vectorized features (1) are expressive enough to replace the raw signals in the sleep staging task, and (2) provide better predictive performance than supervised models in scenarios involving fewer labels and noisy samples.
UNASSIGNED: We propose a self-supervised model, Contrast with the World Representation (ContraWR), for EEG signal representation learning. Unlike previous models that use a set of negative samples, our model uses global statistics (ie, the average representation) from the data set to distinguish signals associated with different sleep stages. The ContraWR model is evaluated on 3 real-world EEG data sets that include both settings: at-home and in-laboratory EEG recording.
UNASSIGNED: ContraWR outperforms 4 recently reported self-supervised learning methods on the sleep staging task across 3 large EEG data sets. ContraWR also supersedes supervised learning when fewer training labels are available (eg, 4% accuracy improvement when less than 2% of data are labeled on the Sleep EDF data set). Moreover, the model provides informative, representative feature structures in 2D projection.
UNASSIGNED: We show that ContraWR is robust to noise and can provide high-quality EEG representations for downstream prediction tasks. The proposed model can be generalized to other unsupervised physiological signal learning tasks. Future directions include exploring task-specific data augmentations and combining self-supervised methods with supervised methods, building upon the initial success of self-supervised learning reported in this study.

Deep learning (DL) can predict microsatellite instability (MSI) from routine histopathology slides of colorectal cancer (CRC). However, it is unclear whether DL can also predict other biomarkers with high performance and whether DL predictions generalize to external patient populations. Here, we acquire CRC tissue samples from two large multi-centric studies. We systematically compare six different state-of-the-art DL architectures to predict biomarkers from pathology slides, including MSI and mutations in BRAF, KRAS, NRAS, and PIK3CA. Using a large external validation cohort to provide a realistic evaluation setting, we show that models using self-supervised, attention-based multiple-instance learning consistently outperform previous approaches while offering explainable visualizations of the indicative regions and morphologies. While the prediction of MSI and BRAF mutations reaches a clinical-grade performance, mutation prediction of PIK3CA, KRAS, and NRAS was clinically insufficient.

Recently, deep learning has become more and more extensive in the field of fault diagnosis. However, most deep learning methods rely on large amounts of labeled data to train the model, which leads to their poor generalized ability in the application of different scenarios. To overcome this deficiency, this paper proposes a novel generalized model based on self-supervised learning and sparse filtering (GSLSF). The proposed method includes two stages. Firstly (1), considering the representation of samples on fault and working condition information, designing self-supervised learning pretext tasks and pseudo-labels, and establishing a pre-trained model based on sparse filtering. Secondly (2), a knowledge transfer mechanism from the pre-training model to the target task is established, the fault features of the deep representation are extracted based on the sparse filtering model, and softmax regression is applied to distinguish the type of failure. This method can observably enhance the model\'s diagnostic performance and generalization ability with limited training data. The validity of the method is proved by the fault diagnosis results of two bearing datasets.

Quantifying emotional aspects of animal behavior (e.g., anxiety, social interactions, reward, and stress responses) is a major focus of neuroscience research. Because manual scoring of emotion-related behaviors is time-consuming and subjective, classical methods rely on easily quantified measures such as lever pressing or time spent in different zones of an apparatus (e.g., open vs. closed arms of an elevated plus maze). Recent advancements have made it easier to extract pose information from videos, and multiple approaches for extracting nuanced information about behavioral states from pose estimation data have been proposed. These include supervised, unsupervised, and self-supervised approaches, employing a variety of different model types. Representations of behavioral states derived from these methods can be correlated with recordings of neural activity to increase the scope of connections that can be drawn between the brain and behavior. In this mini review, we will discuss how deep learning techniques can be used in behavioral experiments and how different model architectures and training paradigms influence the type of representation that can be obtained.

Imaging is commonly used as a characterization method in the pharmaceuticals industry, including for quantifying subvisible particles in solid and liquid formulations. Extracting information beyond particle size, such as classifying morphological subpopulations, requires some type of image analysis method. Suggested methods to classify particles have been based on pre-determined morphological features or use supervised training of convolutional neural networks to learn image representations in relation to ground truth labels. Complications arising from highly complex morphologies, unforeseen classes, and time-consuming preparation of ground truth labels, are some of the challenges faced by these methods. In this work, we evaluate the application of a self-supervised contrastive learning method in studying particle images from therapeutic solutions. Unlike with supervised training, this approach does not require ground truth labels and representations are learned by comparing particle images and their augmentations. This method provides a fast and easily implementable tool of coarse screening for morphological attribute assessment. Furthermore, our analysis shows that in cases with relatively balanced datasets, a small subset of an image dataset is sufficient to train a convolutional neural network encoder capable of extracting useful image representations. It is also demonstrated that particle classes typically observed in protein solutions administered by pre-filled syringes emerge as separated clusters in the encoder\'s embedding space, facilitating performing tasks such as training weakly-supervised classifiers or identifying the presence of new subpopulations.

Existing electronic cleansing (EC) methods for computed tomographic colonography (CTC) are generally based on image segmentation, which limits their accuracy to that of the underlying voxels. Because of the limitations of the available CTC datasets for training, traditional deep learning is of limited use in EC. The purpose of this study was to evaluate the technical feasibility of using a novel self-supervised adversarial learning scheme to perform EC with a limited training dataset with subvoxel accuracy. A three-dimensional (3D) generative adversarial network (3D GAN) was pre-trained to perform EC on CTC datasets of an anthropomorphic phantom. The 3D GAN was then fine-tuned to each input case by use of the self-supervised scheme. The architecture of the 3D GAN was optimized by use of a phantom study. The visually perceived quality of the virtual cleansing by the resulting 3D GAN compared favorably to that of commercial EC software on the virtual 3D fly-through examinations of 18 clinical CTC cases. Thus, the proposed self-supervised 3D GAN, which can be trained to perform EC on a small dataset without image annotations with subvoxel accuracy, is a potentially effective approach for addressing the remaining technical problems of EC in CTC.