With a massive influx of multimodality data, the role of data analytics in health informatics has grown rapidly in the last decade. This has also prompted increasing interests in the generation of analytical, data driven models based on machine learning in health informatics. Deep learning, a technique with its foundation in artificial neural networks, is emerging in recent years as a powerful tool for machine learning, promising to reshape the future of artificial intelligence. Rapid improvements in computational power, fast data storage, and parallelization have also contributed to the rapid uptake of the technology in addition to its predictive power and ability to generate automatically optimized high-level features and semantic interpretation from the input data. This article presents a comprehensive up-to-date review of research employing deep learning in health informatics, providing a critical analysis of the relative merit, and potential pitfalls of the technique as well as its future outlook. The paper mainly focuses on key applications of deep learning in the fields of translational bioinformatics, medical imaging, pervasive sensing, medical informatics, and public health.

This study examined the brain bases of early human social cognitive abilities. Specifically, we investigated whether cortical regions implicated in adults' perception of facial communication signals are functionally active in early human development. Four-month-old infants watched two kinds of dynamic scenarios in which a face either established mutual gaze or averted its gaze, both of which were followed by an eyebrow raise with accompanying smile. Haemodynamic responses were measured by near-infrared spectroscopy, permitting spatial localization of brain activation (experiment 1), and gamma-band oscillatory brain activity was analysed from electroencephalography to provide temporal information about the underlying cortical processes (experiment 2). The results revealed that perceiving facial communication signals activates areas in the infant temporal and prefrontal cortex that correspond to the brain regions implicated in these processes in adults. In addition, mutual gaze itself, and the eyebrow raise with accompanying smile in the context of mutual gaze, produce similar cortical activations. This pattern of results suggests an early specialization of the cortical network involved in the perception of facial communication cues, which is essential for infants' interactions with, and learning from, others.

Whole brain functional connectomes hold promise for understanding human brain activity across a range of cognitive, developmental and pathological states. So called resting-state (rs) functional MRI studies have contributed to the brain being considered at a macroscopic scale as a set of interacting regions. Interactions are defined as correlation-based signal measurements driven by blood oxygenation level dependent (BOLD) contrast. Understanding the neurophysiological basis of these measurements is important in conveying useful information about brain function. Local coupling between BOLD fMRI and neurophysiological measurements is relatively well defined, with evidence that gamma (range) frequency EEG signals are the closest correlate of BOLD fMRI changes during cognitive processing. However, it is less clear how whole-brain network interactions relate during rest where lower frequency signals have been suggested to play a key role. Simultaneous EEG-fMRI offers the opportunity to observe brain network dynamics with high spatio-temporal resolution. We utilize these measurements to compare the connectomes derived from rs-fMRI and EEG band limited power (BLP). Merging this multi-modal information requires the development of an appropriate statistical framework. We relate the covariance matrices of the Hilbert envelope of the source localized EEG signal across bands to the covariance matrices derived from rs-fMRI with the means of statistical prediction based on sparse Canonical Correlation Analysis (sCCA). Subsequently, we identify the most prominent connections that contribute to this relationship. We compare whole-brain functional connectomes based on their geodesic distance to reliably estimate the performance of the prediction. The performance of predicting fMRI from EEG connectomes is considerably better than predicting EEG from fMRI across all bands, whereas the connectomes derived in low frequency EEG bands resemble best rs-fMRI connectivity.

Abstract. In robotically assisted laparoscopic surgery, soft-tissue motion tracking and structure recovery are important for intraoperative surgical guidance, motion compensation and delivering active constraints. In this paper, we present a novel method for feature based motion tracking of deformable soft-tissue surfaces in totally endoscopic coronary artery bypass graft (TECAB) surgery. We combine two feature detectors to recover distinct regions on the epicardial surface for which the sparse 3D surface geometry may be computed using a pre-calibrated stereo laparoscope. The movement of the 3D points is then tracked in the stereo images with stereo-temporal constrains by using an iterative registration algorithm. The practical value of the technique is demonstrated on both a deformable phantom model with tomographically derived surface geometry and in vivo robotic assisted minimally invasive surgery (MIS) image sequences.

The current study tested whether the purely amodal cue of contingency elicit orientation following behaviour in 8-months-old infants. We presented 8-month-old infants with automated objects without human features that did or did not react contingently to the infants' fixations recorded by an eye-tracker. We found that an object's occasional orientation towards peripheral targets was reciprocated by a congruent visual orientation following response by infants only when it had displayed gaze-contingent interactivity. Our finding demonstrates that infants' gaze following behaviour does not depend on the presence of a human being. The results are consistent with the idea that the detection of contingent reactivity, like other communicative signals, can itself elicit the illusion of being addressed in 8-months-old infants.

In Diffusion Weighted MR Imaging (DWI), the signal is affected by the biophysical properties of neuronal cells and their relative placement, as well as extra-cellular tissue compartments. Typically, microstructural indices, such as fractional anisotropy (FA) and mean diffusivity (MD), are based on a tensor model that cannot disentangle the influence of these parameters. Recently, Neurite Orientation Dispersion and Density Imaging (NODDI) has exploited multi-shell acquisition protocols to model the diffusion signal as the contribution of three tissue compartments. NODDI microstructural indices, such as intra-cellular volume fraction (ICVF) and orientation dispersion index (ODI) are directly related to neuronal density and orientation dispersion, respectively. One way of examining the neurophysiological role of these microstructural indices across neuronal fibres is to look into how they relate to brain function. Here we exploit a statistical framework based on sparse Canonical Correlation Analysis (sCCA) and randomised Lasso to identify structural connections that are highly correlated with resting-state functional connectivity measured with simultaneous EEG-fMRI. Our results reveal distinct structural fingerprints for each microstructural index that also reflect their inter-relationships.

Assistive robots play an important role in improving the quality of life of patients at home. Among all the monitoring tasks, gait disorders are prevalent in elderly and people with neurological conditions, which increases the risk of fall. Therefore, the development of mobile systems for gait monitoring at home in normal living conditions is important. Here we present a mobile system that is able to track humans and analyze their gait in canonical coordinates based on a single RGB-D camera. Firstly, view-invariant 3D lower limb pose estimation is achieved by fusing information from depth images along with 2D joints derived in RGB images. Next, both the 6D camera pose and the 3D lower limb skeleton are real-time tracked in a canonical coordinate system based on Simultaneously Localization and Mapping (SLAM). A maskbased strategy is exploited to improve the re-localization of the SLAM in dynamic environments. Abnormal gait is detected by using the Support Vector Machine (SVM) and the Bidirectional Long-Short Term Memory (BiLSTM) network with respect to a set of extracted gait features. To evaluate the robustness of the system, we collected multi-camera, ground truth data from sixteen healthy volunteers performing six gait patterns that mimic common gait abnormalities. The experiment results demonstrate that our proposed system can achieve good lower limb pose estimation and superior recognition accuracy compared to previous abnormal gait detection methods.

Functional connections between brain regions are supported by structural connectivity. Both functional and structural connectivity are estimated from in vivo magnetic resonance imaging and offer complementary information on brain organization and function. However, imaging only provides noisy measures, and we lack a good neuroscientific understanding of the links between structure and function. Therefore, inter-subject joint modeling of structural and functional connectivity, the key to multimodal biomarkers, is an open challenge. We present a probabilistic framework to learn across subjects a mapping from structural to functional brain connectivity. Expanding on our previous work [1], our approach is based on a predictive framework with multiple sparse linear regression. We rely on the randomized LASSO to identify relevant anatomo-functional links with some confidence interval. In addition, we describe resting-state functional magnetic resonance imaging in the setting of Gaussian graphical models, on the one hand imposing conditional independences from structural connectivity and on the other hand parameterizing the problem in terms of multivariate autoregressive models. We introduce an intrinsic measure of prediction error for functional connectivity that is independent of the parameterization chosen and provides the means for robust model selection. We demonstrate our methodology with regions within the default mode and the salience network as well as, atlas-based cortical parcellation.