Pulse-coupled neural network performance for real-time identification of vegetation during forced landing

Safety concerns in the operation of autonomous aerial systems require safe-landing protocols be followed during situations where the mission should be aborted due to mechanical or other failure. This article presents a pulse-coupled neural network (PCNN) to assist in the vegetation classification in a vision-based landing site detection system for an unmanned aircraft. We propose a heterogeneous computing architecture and an OpenCL implementation of a PCNN feature generator. Its performance is compared across OpenCL kernels designed for CPU, GPU, and FPGA platforms. This comparison examines the compute times required for network convergence under a variety of images to determine the plausibility for real-time feature detection.

Mesenchymal stem cells, neural lineage potential, heparan sulfate proteoglycans and the matrix

Along with the tri-lineage of bone, cartilage and fat, human mesenchymal stem cells (hMSCs) retain neural lineage potential. Multiple factors have been described that influence lineage fate of hMSCs including the extracellular microenvironment or niche. The niche includes the extracellular matrix (ECM) providing structural composition, as well as other associated proteins and growth factors, which collectively influence hMSC stemness and lineage specification. As such, lineage specific differentiation of MSCs is mediated through interactions including cell–cell and cell–matrix, as well as through specific signalling pathways triggering downstream events. Proteoglycans (PGs) are ubiquitous within this microenvironment and can be localised to the cell surface or embedded within the ECM. In addition, the heparan sulfate (HS) and chondroitin sulfate (CS) families of PGs interact directly with a number of growth factors, signalling pathways and ECM components including FGFs, Wnts and fibronectin. With evidence supporting a role for HSPGs and CSPGs in the specification of hMSCs down the osteogenic, chondrogenic and adipogenic lineages, along with the localisation of PGs in development and regeneration, it is conceivable that these important proteins may also play a role in the differentiation of hMSCs toward the neuronal lineage. Here we summarise the current literature and highlight the potential for HSPG directed neural lineage fate specification in hMSCs, which may provide a new model for brain damage repair.

Colonizing while migrating : how do individual enteric neural crest cells behave?

Background Directed cell migration is essential for normal development. In most of the migratory cell populations that have been analysed in detail to date, all of the cells migrate as a collective from one location to another. However, there are also migratory cell populations that must populate the areas through which they migrate, and thus some cells get left behind while others advance. Very little is known about how individual cells behave to achieve concomitant directional migration and population of the migratory route. We examined the behavior of enteric neural crest-derived cells (ENCCs), which must both advance caudally to reach the anal end and populate each gut region. Results The behaviour of individual ENCCs was examined using live imaging and mice in which ENCCs express a photoconvertible protein. We show that individual ENCCs exhibit very variable directionalities and speed; as the migratory wavefront of ENCCs advances caudally, each gut region is populated primarily by some ENCCs migrating non-directionally. After populating each region, ENCCs remain migratory for at least 24 hours. Endothelin receptor type B (EDNRB) signaling is known to be essential for the normal advance of the ENCC population. We now show that perturbation of EDNRB principally affects individual ENCC speed rather than directionality. The trajectories of solitary ENCCs, which occur transiently at the wavefront, were consistent with an unbiased random walk and so cell-cell contact is essential for directional migration. ENCCs migrate in close association with neurites. We showed that although ENCCs often use neurites as substrates, ENCCs lead the way, neurites are not required for chain formation and neurite growth is more directional than the migration of ENCCs as a whole. Conclusions Each gut region is initially populated by sub-populations of ENCCs migrating non-directionally, rather than stopping. This might provide a mechanism for ensuring a uniform density of ENCCs along the growing gut.

THE quest for alpha : can artificial neural networks help?

The application of artificial neural networks (ANN) in finance is relatively new area of research. We employed ANNs that used both fundamental and technical inputs to predict future prices of widely held Australian stocks and used these predicted prices for stock portfolio selection over a 10-year period (2001-2011). We found that the ANNs generally do well in predicting the direction of stock price movements. The stock portfolios selected by the ANNs with median accuracy are able to generate positive alpha over the 10-year period. More importantly, we found that a portfolio based on randomly selected network configuration had zero chance of resulting in a significantly negative alpha but a 27% chance of yielding a significantly positive alpha. This is in stark contrast to the findings of the research on mutual fund performance where active fund managers with negative alphas outnumber those with positive alphas.

Predicting freeway pavement construction cost using a back-propagation neural network : a case study in Henan, China

The objective of this research was to develop a model to estimate future freeway pavement construction costs in Henan Province, China. A comprehensive set of factors contributing to the cost of freeway pavement construction were included in the model formulation. These factors comprehensively reflect the characteristics of region and topography and altitude variation, the cost of labour, material, and equipment, and time-related variables such as index numbers of labour prices, material prices and equipment prices. An Artificial Neural Network model using the Back-Propagation learning algorithm was developed to estimate the cost of freeway pavement construction. A total of 88 valid freeway cases were obtained from freeway construction projects let by the Henan Transportation Department during the period 1994−2007. Data from a random selection of 81 freeway cases were used to train the Neural Network model and the remaining data were used to test the performance of the Neural Network model. The tested model was used to predict freeway pavement construction costs in 2010 based on predictions of input values. In addition, this paper provides a suggested correction for the prediction of the value for the future freeway pavement construction costs. Since the change in future freeway pavement construction cost is affected by many factors, the predictions obtained by the proposed method, and therefore the model, will need to be tested once actual data are obtained.

Rho GTPases mediate the mechanosensitive lineage commitment of neural stem cells

Adult neural stem cells (NSCs) play important roles in learning and memory and are negatively impacted by neurological disease. It is known that biochemical and genetic factors regulate self-renewal and differentiation, and it has recently been suggested that mechanical and solid-state cues, such as extracellular matrix (ECM) stiffness, can also regulate the functions of NSCs and other stem cell types. However, relatively little is known of the molecular mechanisms through which stem cells transduce mechanical inputs into fate decisions, the extent to which mechanical inputs instruct fate decisions versus select for or against lineage-committed blast populations, or the in vivo relevance of mechanotransductive signaling molecules in native stem cell niches. Here we demonstrate that ECM-derived mechanical signals act through Rho GTPases to activate the cellular contractility machinery in a key early window during differentiation to regulate NSC lineage commitment. Furthermore, culturing NSCs on increasingly stiff ECMs enhances RhoA and Cdc42 activation, increases NSC stiffness, and suppresses neurogenesis. Likewise, inhibiting RhoA and Cdc42 or downstream regulators of cellular contractility rescues NSCs from stiff matrix- and Rho GTPase-induced neurosuppression. Importantly, Rho GTPase expression and ECM stiffness do not alter proliferation or apoptosis rates indicating that an instructive rather than selective mechanism modulates lineage distributions. Finally, in the adult brain, RhoA activation in hippocampal progenitors suppresses neurogenesis, analogous to its effect in vitro. These results establish Rho GTPase-based mechanotransduction and cellular stiffness as biophysical regulators of NSC fate in vitro and RhoA as an important regulatory protein in the hippocampal stem cell niche.

Out of their minds: legal theory in neural networks

This paper examines the use of connectionism (neural networks) in modelling legal reasoning. I discuss how the implementations of neural networks have failed to account for legal theoretical perspectives on adjudication. I criticise the use of neural networks in law, not because connectionism is inherently unsuitable in law, but rather because it has been done so poorly to date. The paper reviews a number of legal theories which provide a grounding for the use of neural networks in law. It then examines some implementations undertaken in law and criticises their legal theoretical naïvete. It then presents a lessons from the implementations which researchers must bear in mind if they wish to build neural networks which are justified by legal theories.

Cancer : the to and fro of tumour spread

The spread of cells from the primary site of a solid tumour to distant sites remains the major cause of disease and death associated with these cancers. For tumour cells to spread, or metastasize, they must modify their 'anchored' state and detach from their neighbouring cells; migrate through tissues into the blood and lymph systems; survive in these circulation systems; and then leave the vessels at an appropriate site to form another tumour1. Many of these events are favoured by conversions between two cellular states — the epithelial and mesenchymal phenotypes. But the role of these transitions in cancer metastasis is controversial. Writing in Cancer Cell, Tsai et al.2 and Ocaña et al.3 help to clarify this issue...

Determining epithelial contribution to in vivo mesenchymal tumour expression signature using species-specific microarray profiling analysis of xenografts

Gene expression profiling using microarrays and xenograft transplants of human cancer cell lines are both popular tools to investigate human cancer. However, the undefined degree of cross hybridization between the mouse and human genomes hinders the use of microarrays to characterize gene expression of both the host and the cancer cell within the xenograft. Since an increasingly recognized aspect of cancer is the host response (or cancer-stroma interaction), we describe here a bioinformatic manipulation of the Affymetrix profiling that allows interrogation of the gene expression of both the mouse host and the human tumour. Evidence of microenvironmental regulation of epithelial mesenchymal transition of the tumour component in vivo is resolved against a background of mesenchymal gene expression. This tool could allow deeper insight to the mechanism of action of anti-cancer drugs, as typically novel drug efficacy is being tested in xenograft systems.

Staurosporine augments EGF-mediated EMT in PMC42-LA cells through actin depolymerisation, focal contact size reduction and Snail1 induction : a model for cross-modulation

Background A feature of epithelial to mesenchymal transition (EMT) relevant to tumour dissemination is the reorganization of actin cytoskeleton/focal contacts, influencing cellular ECM adherence and motility. This is coupled with the transcriptional repression of E-cadherin, often mediated by Snail1, Snail2 and Zeb1/δEF1. These genes, overexpressed in breast carcinomas, are known targets of growth factor-initiated pathways, however it is less clear how alterations in ECM attachment cross-modulate to regulate these pathways. EGF induces EMT in the breast cancer cell line PMC42-LA and the kinase inhibitor staurosporine (ST) induces EMT in embryonic neural epithelial cells, with F-actin de-bundling and disruption of cell-cell adhesion, via inhibition of aPKC. Methods PMC42-LA cells were treated for 72 h with 10 ng/ml EGF, 40 nM ST, or both, and assessed for expression of E-cadherin repressor genes (Snail1, Snail2, Zeb1/δEF1) and EMT-related genes by QRT-PCR, multiplex tandem PCR (MT-PCR) and immunofluorescence +/- cycloheximide. Actin and focal contacts (paxillin) were visualized by confocal microscopy. A public database of human breast cancers was assessed for expression of Snail1 and Snail2 in relation to outcome. Results When PMC42-LA were treated with EGF, Snail2 was the principal E-cadherin repressor induced. With ST or ST+EGF this shifted to Snail1, with more extreme EMT and Zeb1/δEF1 induction seen with ST+EGF. ST reduced stress fibres and focal contact size rapidly and independently of gene transcription. Gene expression analysis by MT-PCR indicated that ST repressed many genes which were induced by EGF (EGFR, CAV1, CTGF, CYR61, CD44, S100A4) and induced genes which alter the actin cytoskeleton (NLF1, NLF2, EPHB4). Examination of the public database of breast cancers revealed tumours exhibiting higher Snail1 expression have an increased risk of disease-recurrence. This was not seen for Snail2, and Zeb1/δEF1 showed a reverse correlation with lower expression values being predictive of increased risk. Conclusion ST in combination with EGF directed a greater EMT via actin depolymerisation and focal contact size reduction, resulting in a loosening of cell-ECM attachment along with Snail1-Zeb1/δEF1 induction. This appeared fundamentally different to the EGF-induced EMT, highlighting the multiple pathways which can regulate EMT. Our findings add support for a functional role for Snail1 in invasive breast cancer.

Dormant but migratory tumour cells in desmoplastic stroma of invasive ductal carcinomas

Mortality in breast cancer is linked to metastasis and recurrence yet there is no acceptable biological model for cancer relapse. We hypothesise that there might exist primary tumour cells capable of escaping surgery by migration and resisting radiotherapy and chemotherapy to cause cancer recurrence. We investigated this possibility in invasive ductal carcinoma (IDC) tissue and observed the presence of solitary primary tumour cells (SPCs) in the dense collagen stroma that encapsulates intratumoural cells (ICs). In IDC tissue sections, collagen was detected with either Masson's Trichrome or by second harmonics imaging. Cytokeratin-19 (CK-19) and vimentin (VIM) antibodies were, respectively, used to identify epithelial-derived tumour cells and to indicate epithelial to mesenchymal transition (EMT). Confocal/multiphoton microscopy showed that ICs from acini were mainly CK-19 +ve and were encapsulated by dense stromal collagen. Within the stroma, SPCs were detected by their staining for both CK-19 and VIM (confirming EMT). ICs and SPCs were subsequently isolated by laser capture microdissection followed by multiplex tandem-PCR studies. SPCs were found to be enriched for pro-migratory and anti-proliferative genes relative to ICs. In vitro experiments using collagen matrices at 20 mg/cm 3, similar in density to tumour matrices, demonstrated that SPC-like cells were highly migratory but dormant, phenotypes that recapitulated the genotypes of SPCs in clinical tissue. These data suggest that SPCs located at the breast cancer perimeter are invasive and dormant such that they may exceed surgical margins and resist local and adjuvant therapies. This study has important connotations for a role of SPCs in local recurrence.

Artificial neural networks to predict activity type and energy expenditure in youth

Previous studies have demonstrated that pattern recognition approaches to accelerometer data reduction are feasible and moderately accurate in classifying activity type in children. Whether pattern recognition techniques can be used to provide valid estimates of physical activity (PA) energy expenditure in youth remains unexplored in the research literature. Purpose: The objective of this study is to develop and test artificial neural networks (ANNs) to predict PA type and energy expenditure (PAEE) from processed accelerometer data collected in children and adolescents. Methods: One hundred participants between the ages of 5 and 15 yr completed 12 activity trials that were categorized into five PA types: sedentary, walking, running, light-intensity household activities or games, and moderate-to-vigorous intensity games or sports. During each trial, participants wore an ActiGraph GTIM on the right hip, and (V) Over dotO(2) was measured using the Oxycon Mobile (Viasys Healthcare, Yorba Linda, CA) portable metabolic system. ANNs to predict PA type and PAEE (METs) were developed using the following features: 10th, 25th, 50th, 75th, and 90th percentiles and the lag one autocorrelation. To determine the highest time resolution achievable, we extracted features from 10-, 15-, 20-, 30-, and 60-s windows. Accuracy was assessed by calculating the percentage of windows correctly classified and root mean square en-or (RMSE). Results: As window size increased from 10 to 60 s, accuracy for the PA-type ANN increased from 81.3% to 88.4%. RMSE for the MET prediction ANN decreased from 1.1 METs to 0.9 METs. At any given window size, RMSE values for the MET prediction ANN were 30-40% lower than the conventional regression-based approaches. Conclusions: ANNs can be used to predict both PA type and PAEE in children and adolescents using count data from a single waist mounted accelerometer.

Mechanisms of tumour invasion and metastasis : emerging targets for therapy

The progression of a tumour from one of benign and delimited growth to one that is invasive and metastatic is the major cause of poor clinical outcome in cancer patients. The invasion and metastasis of tumours is a highly complex and multistep process that requires a tumour cell to modulate its ability to adhere, degrade the surrounding extracellular matrix, migrate, proliferate at a secondary site and stimulate angiogenesis. Knowledge of the process has greatly increased and this has resulted in the identification of a number of molecules that are fundamental to the process. The involvement of these molecules has been shown to relate not only to the survival and proliferation of the tumour cell but, also to the processes of tumour cell adhesion, migration, and the tumour cells ability to degrade and escape the primary site as well as play a role in angiogenesis. These molecules may provide important therapeutic targets that represent the ability to target specific steps in the process of invasion and metastasis and provide additional therapies. The review focuses on representative key targets in each of these processes and summarises the state of play in each case.