Parameter estimation of the BOLD fMRI model within a general particle filter framework

This work demonstrates a novel Bayesian learning approach for model based analysis of Functional Magnetic Resonance (fMRI) data. We use a physiologically inspired hemodynamic model and investigate a method to simultaneously infer the neural activity together with hidden state and the physiological parameter of the model. This joint estimation problem is still an open topic. In our work we use a Particle Filter accompanied with a kernel smoothing approach to address this problem within a general filtering framework. Simulation results show that the proposed method is a consistent approach and has a good potential to be enhanced for further fMRI data analysis.

Home-range, diet and breeding of a Powerful Owl Ninox strenua in East Gippsland, Victoria

Between 2004 and 2008 the diet and breeding success of a pair of Powerful Owls Ninox strenua were studied near Lakes Entrance, Victoria. In early November 2006 the adult female Powerful Owl was captured and radio-tracked for a period of 7.5 months. During this time the Owl's location was recorded on 111 occasions, including 65 nocturnal locations over 29 nights. Her home-range was calculated as 1589 ha using the Minimum Convex Polygon (MCP) method, or 871 ha based on the 95% Adaptive Kernel method. The area of forested habitat within the MCP home-range was 896 ha (the remainder representing cleared land). Her activity was centred primarily on the nesting gully where two dependent juveniles roosted, but several long-distance foraging expeditions (including roosting) that occurred more than 2.5 km from the juveniles were recorded. Arboreal mammals and birds dominated the Owls' diet. Low prey availability is suggested as being responsible for the single successful breeding event recorded in four nesting seasons.

Fidelity to foraging sites, consistency of migration routes and habitat modulation of home range by sea turtles

Aim  Resources can shape patterns of habitat utilization. Recently a broad foraging dichotomy between oceanic and coastal sites has been revealed for loggerhead sea turtles (Caretta caretta). Since oceanic and coastal foraging sites differ in prey availability, we might expect a gross difference in home-range size across these habitats. We tested this hypothesis by equipping nine adult male loggerhead sea turtles with GPS tracking devices. Location  National Marine Park of Zakynthos (NMPZ) Greece, central and eastern Mediterranean (Adriatic, Ionian and Aegean seas). Methods  In 2007, 2008 and 2009, Fastloc GPS-Argos transmitters were attached to nine male loggerheads. In addition, a Sirtrack PTT unit was attached to one male in 2007. Four of the turtles were tracked on successive years. We filtered the GPS data to ensure comparable data volumes. Route consistency between breeding and foraging sites of the four re-tracked turtles was conducted. Foraging site home range areas and within site movement patterns were investigated by the fixed kernel density method. Results  Foraging home range size ranged between circa 10 km2 at neritic habitats (coastal and open-sea on the continental shelf) to circa 1000 km2 at oceanic sites (using 90% kernel estimates), the latter most probably reflecting sparsely distributed oceanic prey. Across different years individuals did not follow exactly the same migration routes, but did show fidelity to their previous foraging sites, whether oceanic or neritic, with accurate homing in the final stages of migration. Main conclusions  The broad distribution and diverse life-history strategies of this population could complicate the identification of priority marine protected areas beyond the core breeding site.

An information-theoretic approach to face recognition from face motion manifolds

In this work, we consider face recognition from face motion manifolds (FMMs). The use of the resistor-average distance (RAD) as a dissimilarity measure between densities confined to FMMs is motivated in the proposed information-theoretic approach to modelling face appearance. We introduce a kernel-based algorithm that makes use of the simplicity of the closed-form expression for RAD between two Gaussian densities, while allowing for modelling of complex and nonlinear, but intrinsically low-dimensional manifolds. Additionally, it is shown how geodesically local FMM structure can be modelled, naturally leading to a stochastic algorithm for generalizing to unseen modes of data variation. Recognition performance of our method is demonstrated experimentally and is shown to exceed that of state-of-the-art algorithms. Recognition rate of 98% was achieved on a database of 100 people under varying illumination

Sex-biased space-use response to urbanization in an endemic urban adapter

Urbanization impacts on the composition and distribution of wildlife. The tawny frogmouth (Podargus strigoides) is an endemic, nocturnal bird species widespread throughout Australia with recent research highlighting high densities within urban environments. The aim of this study was to investigate homerange size and land-use in response to a gradient of urbanization by determining (a) the key land-use types influencing home-range size and location in the urban landscape (b) whether urbanization impacts on home-range size; and (c) whether the response to urbanization is gender specific. Twelve birds, seven male and five female were radio-tracked within a study zone located in Melbourne, Australia. We used minimum convex polygons (MCP) 95% and 50% fixed-kernel isopleths to calculate home-range size and areas of core use within each home-range. In both the landscape and core areas of their home-range, birds positioned their home-range in areas with more trees, avoiding impervious surfaces and utilizing grassed areas. Male mean kernel home-range was 17.65 ± 4.35 ha and female 6.55 ± 1.40 ha. Male home-ranges contained higher levels of impervious surfaces than females. Modelling demonstrated that as urbanization intensified the home-range size of males increased whereas female home-ranges remained static in size. This research identifies land-use selection and highlights the possibility that spatial behaviour in the species is sex-biased in response to a gradient of urbanization. © 2014 Elsevier B.V.

An improved TIGER2 implementation for NAMD suitable for the Blue Gene architecture

Here we present an improved implementation of the TIGER2 Replica Exchange Molecular Dynamics (REMD) method, using the replica exchange Application Programming Interface (API) found in contemporary versions of the NAMD Molecular Dynamics Package. The implementation takes the form of a TCL script which is used in conjunction with the standard configuration file. This implementation is validated against a previous TIGER2 implementation, as well as data reported for the original TIGER2 simulations. Our implementation is compatible with a range of architectures; crucially it enables the use of this wrapper with the BlueGene/Q architecture, in addition to the x86 architecture. Program summary: Program title: TIGER2-NAMD. Catalogue identifier: AEWC_v1_0. Program summary URL: http://cpc.cs.qub.ac.uk/summaries/AEWC_v1_0.html Program obtainable from: CPC Program Library, Queen's University, Belfast, N. Ireland. Licensing provisions: Standard CPC licence, http://cpc.cs.qub.ac.uk/licence/licence.html No. of lines in distributed program, including test data, etc.: 34151. No. of bytes in distributed program, including test data, etc.: 424217. Distribution format: tar.gz. Programming language: Tcl 8.5. Computer: x86 Clusters, BlueGene/Q, Workstations. Operating system: Linux, IBM Compute Node Kernel. Has the code been vectorised or parallelised?: Yes. MPI Parallelism. Classification: 3. External routines: NAMD 2.9 (http://www.ks.uiuc.edu/Research/namd/). Nature of problem: Replica Exchange Molecular Dynamics. Solution method: Each replica runs through multiple cycles of heating and cooling with exchanges between them being attempted. Running time: Typically 30 mins, up to an hour.

Visualization of HIV-1 interactions with penile and foreskin epithelia: clues for female-to-male HIV transmission

To gain insight into female-to-male HIV sexual transmission and how male circumcision protects against this mode of transmission, we visualized HIV-1 interactions with foreskin and penile tissues in ex vivo tissue culture and in vivo rhesus macaque models utilizing epifluorescent microscopy. 12 foreskin and 14 cadaveric penile specimens were cultured with R5-tropic photoactivatable (PA)-GFP HIV-1 for 4 or 24 hours. Tissue cryosections were immunofluorescently imaged for epithelial and immune cell markers. Images were analyzed for total virions, proportion of penetrators, depth of virion penetration, as well as immune cell counts and depths in the tissue. We visualized individual PA virions breaching penile epithelial surfaces in the explant and macaque model. Using kernel density estimated probabilities of localizing a virion or immune cell at certain tissue depths revealed that interactions between virions and cells were more likely to occur in the inner foreskin or glans penis (from local or cadaveric donors, respectively). Using statistical models to account for repeated measures and zero-inflated datasets, we found no difference in total virions visualized at 4 hours between inner and outer foreskins from local donors. At 24 hours, there were more virions in inner as compared to outer foreskin (0.0495 +/- 0.0154 and 0.0171 +/- 0.0038 virions/image, p = 0.001). In the cadaveric specimens, we observed more virions in inner foreskin (0.0507 +/- 0.0079 virions/image) than glans tissue (0.0167 +/- 0.0033 virions/image, p<0.001), but a greater proportion was seen penetrating uncircumcised glans tissue (0.0458 +/- 0.0188 vs. 0.0151 +/- 0.0100 virions/image, p = 0.099) and to significantly greater mean depths (29.162 +/- 3.908 vs. 12.466 +/- 2.985 μm). Our in vivo macaque model confirmed that virions can breach penile squamous epithelia in a living model. In summary, these results suggest that the inner foreskin and glans epithelia may be important sites for HIV transmission in uncircumcised men.

Time-in-area represents foraging activity in a wide-ranging pelagic forager

Successful Marine Spatial Planning depends upon the identification of areas with high importance for particular species, ecosystems or processes. For seabirds, advancements in biologging devices have enabled us to identify these areas through the detailed study of at-sea behaviour. However, in many cases, only positional data are available and the presence of local biological productivity and hence seabird foraging behaviour is inferred from these data alone, under the untested assumption that foraging activity is more likely to occur in areas where seabirds spend more time. We fitted GPS devices and accelerometers to northern gannets Morus bassanus and categorised the behaviour of individuals outside the breeding colony as plunge diving, surface foraging, floating and flying. We then used the locations of foraging events to test the efficiency of 2 approaches: time-in-area and kernel density (KD) analyses, which are widely employed to detect highly-used areas and interpret foraging behaviour from positional data. For KD analyses, the smoothing parameter (h) was calculated using the ad hoc method (KDad hoc), and KDh=9.1, where h = 9.1 km, to designate core foraging areas from location data. A high proportion of foraging events occurred in core foraging areas designated using KDad hoc, KDh=9.1, and time-in-area. Our findings demonstrate that foraging activity occurs in areas where seabirds spend more time, and that both KD analysis and the time-in-area approach are equally efficient methods for this type of analysis. However, the time-in-area approach is advantageous in its simplicity, and in its ability to provide the shapes commonly used in planning. Therefore, the time-in-area approach can be used as a simple way of using seabirds to identify ecologically important locations from both tracking and survey data.

Microstructure and texture evolution during tensile deformation of symmetric/asymmetric-rolled low carbon microalloyed steel

The deformation and fracture mechanisms of a low carbon microalloyed steel processed by asymmetric rolling (AsR) and symmetric rolling (SR) were compared by microstructural and texture evolutions during uniaxial tensile deformation. A realistic microstructure-based micromechanical modeling was involved as well. AsR provides more effective grain refinement and beneficial shear textures, leading to higher ductility and extraordinary strain hardening with improved yield and ultimate tensile stresses as well as promoting the occurrence of ductile fracture. This was verified and further explained by means of the different fracture modes during quasi-static uniaxial deformation, the preferred void nucleation sites and crack propagation behavior, and the change in the dislocation density based on the kernel average misorientation (KAM) distribution. The equivalent strain/stress partitioning during tensile deformation of AsR and SR specimens was modeled based on a two-dimensional (2D) representative volume element (RVE) approach. The trend of strain/stress partitioning in the ferrite matrix agrees well with the experimental results.

Predicting protein function using multiple kernels

High-throughput experimental techniques provide a wide variety of heterogeneous proteomic data sources. To exploit the information spread across multiple sources for protein function prediction, these data sources are transformed into kernels and then integrated into a composite kernel. Several methods first optimize the weights on these kernels to produce a composite kernel, and then train a classifier on the composite kernel. As such, these approaches result in an optimal composite kernel, but not necessarily in an optimal classifier. On the other hand, some approaches optimize the loss of binary classifiers and learn weights for the different kernels iteratively. For multi-class or multi-label data, these methods have to solve the problem of optimizing weights on these kernels for each of the labels, which are computationally expensive and ignore the correlation among labels. In this paper, we propose a method called Predicting Protein Function using Multiple K ernels (ProMK). ProMK iteratively optimizes the phases of learning optimal weights and reduces the empirical loss of multi-label classifier for each of the labels simultaneously. ProMK can integrate kernels selectively and downgrade the weights on noisy kernels. We investigate the performance of ProMK on several publicly available protein function prediction benchmarks and synthetic datasets. We show that the proposed approach performs better than previously proposed protein function prediction approaches that integrate multiple data sources and multi-label multiple kernel learning methods. The codes of our proposed method are available at https://sites.google.com/site/guoxian85/promk.

A lightweight virtualization solution for android devices

Mobile virtualization has emerged fairly recently and is considered a valuable way to mitigate security risks on Android devices. However, major challenges in mobile virtualization include runtime, hardware, resource overhead, and compatibility. In this paper, we propose a lightweight Android virtualization solution named Condroid, which is based on container technology. Condroid utilizes resource isolation based on namespaces feature and resource control based on cgroups feature. By leveraging them, Condroid can host multiple independent Android virtual machines on a single kernel to support mutilple Android containers. Furthermore, our implementation presents both a system service sharing mechanism to reduce memory utilization and a filesystem sharing mechanism to reduce storage usage. The evaluation results on Google Nexus 5 demonstrate that Condroid is feasible in terms of runtime, hardware resource overhead, and compatibility. Therefore, we find that Condroid has a higher performance than other virtualization solutions.

Fast one-class support vector machine for novelty detection

Novelty detection arises as an important learning task in several applications. Kernel-based approach to novelty detection has been widely used due to its theoretical rigor and elegance of geometric interpretation. However, computational complexity is a major obstacle in this approach. In this paper, leveraging on the cutting-plane framework with the well-known One-Class Support Vector Machine, we present a new solution that can scale up seamlessly with data. The first solution is exact and linear when viewed through the cutting-plane; the second employed a sampling strategy that remarkably has a constant computational complexity defined relatively to the probability of approximation accuracy. Several datasets are benchmarked to demonstrate the credibility of our framework.