Marconi Soares Barbosa

Graph Rigidity, Cyclic Belief Propagation, and Point Pattern Matching

A recent paper (Caetano et al., 2006) proposed a provably optimal, polynomial time method for performing near-isometric point pattern matching by means of exact probabilistic inference in a chordal graphical model. Its fundamental result is that the chordal graph in question is shown to be globally rigid, implying that exact inference provides the same matching solution as exact inference in a complete graphical model. This implies that the algorithm is optimal when there is no noise in the point patterns. In this paper, we present a new graph which is also globally rigid but has an advantage over the graph, its maximal clique size is smaller, rendering inference significantly more efficient. However, this graph is not chordal and thus standard junction tree algorithms cannot be directly applied. Nevertheless, we show that loopy belief propagation in such a graph converges to the optimal solution. This allows us to retain the optimality guarantee in the noiseless case, while substantially reducing both memory requirements and processing time. Our experimental results show that the accuracy of the proposed solution is indistinguishable when there is noise in the point patterns.

Extended morphometric analysis of neuronal cells with Minkowski valuations

Abstract.Minkowski valuations provide a systematic framework for quantifying different aspects of morphology. In this paper we apply vector- and tensor-valued Minkowski valuations to neuronal cells from the cats retina in order to describe their morphological structure in a comprehensive way. We introduce the framework of Minkowski valuations, discuss their implementation for neuronal cells and show how they can be used to characterize cells of different morphological categories. We also provide a comparison to a Sholl analysis.

670nm Photobiomodulation as a Novel Protection against Retinopathy of Prematurity: Evidence from Oxygen Induced Retinopathy Models

Introduction To investigate the validity of using 670nm red light as a preventative treatment for Retinopathy of Prematurity in two animal models of oxygen-induced retinopathy (OIR). Materials and Methods During and post exposure to hyperoxia, C57BL/6J mice or Sprague-Dawley rats were exposed to 670nm light for 3 minutes a day (9J/cm2). Whole mounted retinas were investigated for evidence of vascular abnormalities, while sections of neural retina were used to quantify levels of cell death using the TUNEL technique. Organs were removed, weighed and independent histopathology examination performed. Results 670nm light reduced neovascularisation, vaso-obliteration and abnormal peripheral branching patterns of retinal vessels in OIR. The neural retina was also protected against OIR by 670nm light exposure. OIR-exposed animals had severe lung pathology, including haemorrhage and oedema, that was significantly reduced in 670nm+OIR light-exposed animals. There were no significance differences in the organ weights of animals in the 670nm light-exposed animals, and no adverse effects of exposure to 670nm light were detected. Discussion Low levels of exposure to 670nm light protects against OIR and lung damage associated with exposure to high levels of oxygen, and may prove to be a non-invasive and inexpensive preventative treatment for ROP and chronic lung disease associated with prematurity.

Self-referred approach to lacunarity.

This paper describes an approach to lacunarity which adopts the pattern under analysis as the reference for the sliding window procedure. The superiority of such a scheme with respect to more traditional methodologies, especially when dealing with finite-size objects, is established and illustrated through applications to diffusion limited aggregation pattern characterization. It is also shown that, given the enhanced accuracy and sensitivity of this scheme, the shape of the window becomes an important parameter, with advantage for circular windows.

Neuromorphometric characterization with shape functionals

This work presents a procedure to extract morphological information from neuronal cells based on the variation of shape functionals as the cell geometry undergoes a dilation through a wide interval of spatial scales. The targeted shapes are alpha and beta cat retinal ganglion cells, which are characterized by different ranges of dendritic field diameter. Image functionals are expected to act as descriptors of the shape, gathering relevant geometric and topological features of the complex cell form. We present a comparative study of classification performance of additive shape descriptors, namely, Minkowski functionals, and the nonadditive multiscale fractal. We found that the proposed measures perform efficiently the task of identifying the two main classes alpha and beta based solely on scale invariant information, while also providing intraclass morphological assessment.

ACTIVE PERCOLATION ANALYSIS OF PYRAMIDAL NEURONS OF SOMATOSENSORY CORTEX: A COMPARISON OF WILD TYPE AND p21H-RasVal12 TRANSGENIC MICE

This article describes the investigation of morphological variations among two sets of neuronal cells, namely a control group of wild type mouse cells and a group of cells of a transgenic line. Special attention is given to singular points in the neuronal structure, namely the branching points and extremities of the dendritic processes. The characterization of the spatial distribution of such points is obtained by using a recently reported morphological technique based on forced percolation and window-size compensation, which is particularly suited to the analysis of scattered points, presenting several coexisting densities. Different dispersions were identified in our statistical analysis, suggesting that the transgenic line of neurons is characterized by a more pronounced morphological variation. A classification scheme based on a canonical discriminant function was also considered in order to identify the morphological differences.

Characterizing neuromorphologic alterations with additive shape functionals

The complexity of a neuronal cell shape is known to be related to its function. Specifically, among other indicators, a decreased complexity in the dendritic trees of cortical pyramidal neurons has been associated with mental retardation. In this paper we develop a procedure to address the characterization of morphological changes induced in cultured neurons by over-expressing a gene involved in mental retardation. Measures associated with the multiscale connectivity, an additive image functional, are found to give a reasonable separation criterion between two categories of cells. One category consists of a control group and two transfected groups of neurons, and the other, a class of cat ganglionary cells. The reported framework also identified a trend towards lower complexity in one of the transfected groups. Such results establish the suggested measures as an effective descriptors of cell shape.

Locally countable properties and the perceptual salience of textures

The human ability to discriminate structured from uniformly random binary textures has been shown to exploit third- and higher-order pixel correlations. We examine this ability in an experiment using a large number of texture families that can only be distinguished on the basis of these higher-order correlations. This study investigates statistical models based on possible explanatory variables involving spatial interactions of up to four pixels. Some of these explanatory variables have been recently associated with natural images, and others are somewhat less intuitive and are used here for the first time, to our knowledge. Our models are constructed using intraclass and cross-class feature selection by means of lasso/elastic net optimization and extensive cross-validation. We focus on a special set of locally countable image measures that seem to parsimoniously capture the observed discrimination performance. Among the measures underpinning the best models, we highlight a concept that can only exist in nine-pixel or larger image patches, but nonetheless is calculable based on the multiplicity of specific four-pixel patches in a texture. We show that this single geometric concept provides significant clues to explain texture discrimination.

Morphological Hopfield Networks

This paper reports on the investigation of the effects of neuronal shape, at both individual cell and network level, on the behavior of neuronal systems. More specifically, two-dimensional biologically realistic neuronal networks are obtained that take explicity into account the position and morphology of neuronal cells, with the respective behavior for associative recall being simulated through a diluted version of Hopfields model. While a specific probability density function is used for the placement of the cell bodies, images of real neuronal cells (namely alpha and beta ganglion cells from the cat retina) are used to obtain biologically realistic models. Several morphological measures – including fractal dimension, the excluded volume, and integral geometry functionals–are estimated for the considered cells, and their values are correlated with the potential of the network for associative recall, which is quantified in terms of the overlap between distorted version of the trained patterns and their original version. Such an approach allows the quantitative and objective characterization of the relationship between neuronal shape and function, an important issue in neuroscience. The obtained results substantiate interesting relationships between the neural morphology and function as determined by the performance of the network.

On the potential of the excluded volume and autocorrelation as neuromorphometric descriptors

This work investigates at what degree two neuromorphometric measurements, namely the autocorrelation and the excluded volume of a neuronal cell can influence the characterization and classification of such type of cells. While the autocorrelation function presents good potential for quantifying the dendrite–dendrite connectivity of cells in mosaic tilings, the excluded volume, i.e., the amount of the surround space which is geometrically not accessible to an axon or dendrite, provides a complementary characterization of the cell connectivity. The potential of such approaches is illustrated with respect to real neuronal cells.