Peter Andras

Simulation of robustness against lesions of cortical networks

Structure entails function, and thus a structural description of the brain will help to understand its function and may provide insights into many properties of brain systems, from their robustness and recovery from damage to their dynamics and even their evolution. Advances in the analysis of complex networks provide useful new approaches to understanding structural and functional properties of brain networks. Structural properties of networks recently described allow their characterization as small‐world, random (exponential) and scale‐free. They complement the set of other properties that have been explored in the context of brain connectivity, such as topology, hodology, clustering and hierarchical organization. Here we apply new network analysis methods to cortical interareal connectivity networks for the cat and macaque brains. We compare these corticocortical fibre networks to benchmark rewired, small‐world, scale‐free and random networks using two analysis strategies, in which we measure the effects of the removal of nodes and connections on the structural properties of the cortical networks. The structural decay of the brain networks is in most respects similar to that of scale‐free networks. The results implicate highly connected hub‐nodes and bottleneck connections as a structural basis for some of the conditional robustness of brain systems. This informs the understanding of the development of connectivity of the brain networks.

Alternative sampling methods for estimating multivariate normal probabilities

We study the performance of alternative sampling methods for estimating multivariate normal probabilities through the GHK simulator. The sampling methods are randomized versions of some quasi-Monte Carlo samples (Halton, Niederreiter, Niederreiter-Xing sequences and lattice points) and some samples based on orthogonal arrays (Latin hypercube, orthogonal array and orthogonal array based Latin hypercube samples). In general, these samples turn out to have a better performance than Monte Carlo and antithetic Monte Carlo samples. Improvements over these are large for low-dimensional (4 and 10) cases and still significant for dimensions as large as 50.

On preserving statistical characteristics of accelerometry data using their empirical cumulative distribution

The majority of activity recognition systems in wearable computing rely on a set of statistical measures, such as means and moments, extracted from short frames of continuous sensor measurements to perform recognition. These features implicitly quantify the distribution of data observed in each frame. However, feature selection remains challenging and labour intensive, rendering a more generic method to quantify distributions in accelerometer data much desired. In this paper we present the ECDF representation, a novel approach to preserve characteristics of arbitrary distributions for feature extraction, which is particularly suitable for embedded applications. In extensive experiments on six publicly available datasets we demonstrate that it outperforms common approaches to feature extraction across a wide variety of tasks.

Kernel-Kohonen networks.

We investigate the combination of the Kohonen networks with the kernel methods in the context of classification. We use the idea of kernel functions to handle products of vectors of arbitrary dimension. We indicate how to build Kohonen networks with robust classification performance by transformation of the original data vectors into a possibly infinite dimensional space. The resulting Kohonen networks preserve a non-Euclidean neighborhood structure of the input space that fits the properties of the data. We show how to optimize the transformation of the data vectors in order to obtain higher classification performance. We compare the kernel-Kohonen networks with the regular Kohonen networks in the context of a classification task.

Environmental adversity and uncertainty favour cooperation

BackgroundA major cornerstone of evolutionary biology theory is the explanation of the emergence of cooperation in communities of selfish individuals. There is an unexplained tendency in the plant and animal world – with examples from alpine plants, worms, fish, mole-rats, monkeys and humans – for cooperation to flourish where the environment is more adverse (harsher) or more unpredictable.ResultsUsing mathematical arguments and computer simulations we show that in more adverse environments individuals perceive their resources to be more unpredictable, and that this unpredictability favours cooperation. First we show analytically that in a more adverse environment the individual experiences greater perceived uncertainty. Second we show through a simulation study that more perceived uncertainty implies higher level of cooperation in communities of selfish individuals.ConclusionThis study captures the essential features of the natural examples: the positive impact of resource adversity or uncertainty on cooperation. These newly discovered connections between environmental adversity, uncertainty and cooperation help to explain the emergence and evolution of cooperation in animal and human societies.

The Equivalence of Support Vector Machine and Regularization Neural Networks

We show in this brief paper the equivalence of the support vector machine and regularization neural networks. We prove both implication sides of the equivalence in a generally applicable way. The novelty lies in the effective construction of the regularization operator corresponding to a given support vector machine formulation. We give also a short introductory description of both neural network approximation frameworks.

Evaluating universities using simple scientometric research-output metrics: total citation counts per university for a retrospective seven-year rolling sample

We advocate a scientometric, top-down and institution-based research-assessment methodology that is based on total citations accumulated from all publications associated with a specific university during the survey period. The exercise could be done every year using a rolling seven-year retrospective sample and should be performed by at least two independent auditors. Identification of elite ‘revolutionary-science’ institutions could be accomplished using a metric derived from the distribution of science Nobel Prizes. Copyright , Beech Tree Publishing.

Research: metrics, quality, and management implications

Research evaluation is increasingly important in management decisions in universities. Research metrics provide an objective way to assess the research output of individuals, groups, departments and universities. Such metrics work well as quality assessment tools in the case of normal science research in mature sciences, and also in the case of early stage sciences containing a significant amount of research that is meant to be revolutionary. Revolutionary research in mature sciences and unfashionable revolutionary research in early stage sciences remain mostly invisible to research metrics in the short term. This kind of research may become measurable in the long term if it turns out to be successful and generates a large volume of follow-on research that becomes part of normal science. Pursuing revolutionary research is risky, and this risk is modulated by the availability of an appropriate research workforce and the funding environment. Hype and spin are part of the mechanisms of scientific public opinion, and dealing with these is important in the context of management decisions based on research metrics. Copyright , Beech Tree Publishing.

Environmental risk, cooperation, and communication complexity

The evolution of cooperation and communication in communities of individuals is a puzzling problem for a wide range of scientific disciplines, ranging from evolutionary theory to the theory and application of multi-agent systems. A key issue is to understand the factors that affect collaboration and communication evolution. To address this problem, here we choose the environmental risk as a compact descriptor of the environment in a model world of simple agents. We analyse the evolution of cooperation and communication as a function of the environmental risk. Our findings show that collaboration is more likely to rise to high levels within the agent society in a world characterised by high risk than in one characterised by low risk. With respect to the evolution of communication, we found that communities of agents with high levels of collaboration are more likely to use less complex communication than those which show lower levels of collaboration. Our results have important implications for understanding the evolution of both cooperation and communication, and the interrelationships between them.

Simultaneous measurement of membrane potential changes in multiple pattern generating neurons using voltage sensitive dye imaging

Optical imaging using voltage-sensitive dyes (VSDs) is a promising technique for the simultaneous activity recording of many individual neurons. While such simultaneous recordings are critical for the understanding of the integral functionality of neural systems, functional interpretations on a single neuron level are difficult without knowledge of the connectivity of the underlying circuit. Central pattern generating circuits, such as the pyloric and gastric mill circuits in the stomatogastric ganglion (STG) of crustaceans, allow such investigations due to their well-known connectivities and have already contributed much to our understanding of general neuronal mechanisms. Here we present for the first time simultaneous optical recordings of the pattern generating neurons in the STG of two crustacean species using bulk loading of the VSD di-4-ANEPPS. We demonstrate the recording of firing activities and synaptic interactions of the circuit neurons as well as inter-circuit interactions in their functional context, i.e. without artificial stimulation. Neurons could be uniquely identified using simple event-triggered averaging. We tested this technique in two different species of crustaceans (lobsters and crabs), since several crustacean species are used for studying motor pattern generation. The signal-to-noise ratio of the optical signal was high enough in both species to derive phase-relationship between the network neurons, as well as action potentials and excitatory and inhibitory postsynaptic potentials. We argue that imaging of neural networks with identifiable neurons with well-known connectivity, like in the STG, is crucial for the understanding of emergence of network functionality.