Prediction and risk stratification from hospital discharge records based on Hierarchical sLDA

Background The greatly accelerated development of information technology has conveniently provided adoption for risk stratification, which means more beneficial for both patients and clinicians. Risk stratification offers accurate individualized prevention and therapeutic decision making etc. Hospital discharge records (HDRs) routinely include accurate conclusions of diagnoses of the patients. For this reason, in this paper, we propose an improved model for risk stratification in a supervised fashion by exploring HDRs about coronary heart disease (CHD). Methods We introduced an improved four-layer supervised latent Dirichlet allocation (sLDA) approach called Hierarchical sLDA model, which categorized patient features in HDRs as patient feature-value pairs in one-hot way according to clinical guidelines for lab test of CHD. To address the data missing and imbalance problem, RFs and SMOTE methods are used respectively. After TF-IDF processing of datasets, variational Bayes expectation-maximization method and generalized linear model were used to recognize the latent clinical state of a patient, i.e., risk stratification, as well as to predict CHD. Accuracy, macro-F1, training and testing time performance were used to evaluate the performance of our model. Results According to the characteristics of our datasets, i.e., patient feature-value pairs, we construct a supervised topic model by adding one more Dirichlet distribution hyperparameter to sLDA. Compared with established supervised algorithm Multi-class sLDA model, we demonstrate that our proposed approach enhances training time by 59.74% and testing time by 25.58% but almost no loss of average prediction accuracy on our datasets. Conclusions A model for risk stratification and prediction of CHD based on sLDA model was proposed. Experimental results show that Hierarchical sLDA model we proposed is competitive in time performance and accuracy. Hierarchical processing of patient features can significantly improve the disadvantages of low efficiency and time-consuming Gibbs sampling of sLDA model.

Receptor architecture of macaque and human early visual areas: not equal, but comparable

Existing cytoarchitectonic maps of the human and macaque posterior occipital cortex differ in the number of areas they display, thus hampering identification of homolog structures. We applied quantitative in vitro receptor autoradiography to characterize the receptor architecture of the primary visual and early extrastriate cortex in macaque and human brains, using previously published cytoarchitectonic criteria as starting point of our analysis. We identified 8 receptor architectonically distinct areas in the macaque brain (mV1d, mV1v, mV2d, mV2v, mV3d, mV3v, mV3A, mV4v), and their respective counterpart areas in the human brain (hV1d, hV1v, hV2d, hV2v, hV3d, hV3v, hV3A, hV4v). Mean densities of 14 neurotransmitter receptors were quantified in each area, and ensuing receptor fingerprints used for multivariate analyses. The 1st principal component segregated macaque and human early visual areas differ. However, the 2nd principal component showed that within each species, area-specific differences in receptor fingerprints were associated with the hierarchical processing level of each area. Subdivisions of V2 and V3 were found to cluster together in both species and were segregated from subdivisions of V1 and from V4v. Thus, comparative studies like this provide valuable architectonic insights into how differences in underlying microstructure impact evolutionary changes in functional processing of the primate brain and, at the same time, provide strong arguments for use of macaque monkey brain as a suitable animal model for translational studies.

fMRI Evidence for Default Mode Network Deactivation Associated with Rapid Eye Movements in Sleep

System-specific brain responses—time-locked to rapid eye movements (REMs) in sleep—are characteristically widespread, with robust and clear activation in the primary visual cortex and other structures involved in multisensory integration. This pattern suggests that REMs underwrite hierarchical processing of visual information in a time-locked manner, where REMs index the generation and scanning of virtual-world models, through multisensory integration in dreaming—as in awake states. Default mode network (DMN) activity increases during rest and reduces during various tasks including visual perception. The implicit anticorrelation between the DMN and task-positive network (TPN)—that persists in REM sleep—prompted us to focus on DMN responses to temporally-precise REM events. We timed REMs during sleep from the video recordings and quantified the neural correlates of REMs—using functional MRI (fMRI)—in 24 independent studies of 11 healthy participants. A reanalysis of these data revealed that the cortical areas exempt from widespread REM-locked brain activation were restricted to the DMN. Furthermore, our analysis revealed a modest temporally-precise REM-locked decrease—phasic deactivation—in key DMN nodes, in a subset of independent studies. These results are consistent with hierarchical predictive coding; namely, permissive deactivation of DMN at the top of the hierarchy (leading to the widespread cortical activation at lower levels; especially the primary visual cortex). Additional findings indicate REM-locked cerebral vasodilation and suggest putative mechanisms for dream forgetting.

Speech Encryption Scheme Based on Ciphertext Policy Hierarchical Attribute in Cloud Storage

In order to ensure the confidentiality and secure sharing of speech data, and to solve the problems of slow deployment of attribute encryption systems and fine-grained access control in cloud storage, a speech encryption scheme based on ciphertext policy hierarchical attributes was proposed. First, perform hierarchical processing of the attributes of the speech data to reflect the hierarchical structure and integrate the hierarchical access structure into a single-access structure. Second, use the attribute fast encryption framework to construct the attribute encryption scheme of the speech data, and use the integrated access to the speech data; thus, the structure is encrypted and uploaded to the cloud for storage and sharing. Finally, use the hardness of decisional bilinear Diffie–Hellman (DBDH) assumption to prove that the proposed scheme is secure in the random oracle model. The theoretical security analysis and experimental results show that the proposed scheme can achieve efficient and fine-grained access control and is secure and extensible.

Processing Complex Events in Fog-Based Internet of Things Systems for Smart Agriculture

The recent growth of the Internet of Things’ services and applications has increased data processing and storage requirements. The Edge computing concept aims to leverage the processing capabilities of the IoT and other devices placed at the edge of the network. One embodiment of this paradigm is Fog computing, which provides an intermediate and often hierarchical processing tier between the data sources and the remote Cloud. Among the major benefits of this concept, the end-to-end latency can be decreased, thus favoring time-sensitive applications. Moreover, the data traffic at the network core and the Cloud computing workload can be reduced. Combining the Fog computing paradigm with Complex Event Processing (CEP) and data fusion techniques has excellent potential for generating valuable knowledge and aiding decision-making processes in the Internet of Things’ systems. In this context, we propose a multi-tier complex event processing approach (sensor node, Fog, and Cloud) that promotes fast decision making and is based on information with 98% accuracy. The experiments show a reduction of 77% in the average time of sending messages in the network. In addition, we achieved a reduction of 82% in data traffic.

Distinct neural sources underlying visual word form processing as revealed by steady state visual evoked potentials (SSVEP)

EEG has been central to investigations of the time course of various neural functions underpinning visual word recognition. Recently the steady-state visual evoked potential (SSVEP) paradigm has been increasingly adopted for word recognition studies due to its high signal-to-noise ratio. Such studies, however, have been typically framed around a single source in the left ventral occipitotemporal cortex (vOT). Here, we combine SSVEP recorded from 16 adult native English speakers with a data-driven spatial filtering approach—Reliable Components Analysis (RCA)—to elucidate distinct functional sources with overlapping yet separable time courses and topographies that emerge when contrasting words with pseudofont visual controls. The first component topography was maximal over left vOT regions with a shorter latency (approximately 180 ms). A second component was maximal over more dorsal parietal regions with a longer latency (approximately 260 ms). Both components consistently emerged across a range of parameter manipulations including changes in the spatial overlap between successive stimuli, and changes in both base and deviation frequency. We then contrasted word-in-nonword and word-in-pseudoword to test the hierarchical processing mechanisms underlying visual word recognition. Results suggest that these hierarchical contrasts fail to evoke a unitary component that might be reasonably associated with lexical access.

Human Texture Vision as Multi-Order Spectral Analysis

Texture information plays a critical role in the rapid perception of scenes, objects, and materials. Here, we propose a novel model in which visual texture perception is essentially determined by the 1st-order (2D-luminance) and 2nd-order (4D-energy) spectra. This model is an extension of the dimensionality of the Filter-Rectify-Filter (FRF) model, and it also corresponds to the frequency representation of the Portilla-Simoncelli (PS) statistics. We show that preserving two spectra and randomizing phases of a natural texture image result in a perceptually similar texture, strongly supporting the model. Based on only two single spectral spaces, this model provides a simpler framework to describe and predict texture representations in the primate visual system. The idea of multi-order spectral analysis is consistent with the hierarchical processing principle of the visual cortex, which is approximated by a multi-layer convolutional network.

Understanding the Prefrontal Cortex

The primate prefrontal cortex sits at the top of the sensory, motor, and outcome processing hierarchies of the neocortex. It transforms sensory inputs into motor outputs, determining the response that is appropriate given the current context and desired outcome. This transformation involves conditional rules. The dorsal prefrontal cortex supports the learning of behavioural sequences, where the next action is conditional on the previous one. The ventral prefrontal cortex supports associations between objects, where the choice of one object is conditional on the presence of another object. However, because hierarchical processing supports the extraction of abstract representations, the primate prefrontal cortex is able to represent conditional rules that are abstract, meaning that they apply irrespective of the specific inputs. The selective advantage is that by learning these rules, primates can solve new problems rapidly when they have the same conditional logic as prior problems. The human prefrontal cortex has the same fundamental organization as in other primates. The dorsal prefrontal cortex supports the understanding of sequences and the ventral prefrontal cortex supports the ability to learn semantic associations. Thus the human prefrontal cortex has co-opted and elaborated mechanisms that were present in ancestral primates. These mechanisms can be used for new ends. For example, words have been associated with objects so as to communicate with others. This means that to understand human intelligence it is necessary to take into account the fact that the abstract rules are transmitted verbally from one generation to another.