Magnification factors for the SOM and GTM algorithms

Magnification factors specify the extent to which the area of a small patch of the latent (or `feature') space of a topographic mapping is magnified on projection to the data space, and are of considerable interest in both neuro-biological and data analysis contexts. Previous attempts to consider magnification factors for the self-organizing map (SOM) algorithm have been hindered because the mapping is only defined at discrete points (given by the reference vectors). In this paper we consider the batch version of SOM, for which a continuous mapping can be defined, as well as the Generative Topographic Mapping (GTM) algorithm of Bishop et al. (1997) which has been introduced as a probabilistic formulation of the SOM. We show how the techniques of differential geometry can be used to determine magnification factors as continuous functions of the latent space coordinates. The results are illustrated here using a problem involving the identification of crab species from morphological data.

Neural Networks: A Diagnostic Tool for Gastric Electrical Uncoupling?

Neural Networks have been successfully employed in different biomedical settings. They have been useful for feature extractions from images and biomedical data in a variety of diagnostic applications. In this paper, they are applied as a diagnostic tool for classifying different levels of gastric electrical uncoupling in controlled acute experiments on dogs. Data was collected from 16 dogs using six bipolar electrodes inserted into the serosa of the antral wall. Each dog underwent three recordings under different conditions: (1) basal state, (2) mild surgically-induced uncoupling, and (3) severe surgically-induced uncoupling. For each condition half-hour recordings were made. The neural network was implemented according to the Learning Vector Quantization model. This is a supervised learning model of the Kohonen Self-Organizing Maps. Majority of the recordings collected from the dogs were used for network training. Remaining recordings served as a testing tool to examine the validity of the training procedure. Approximately 90% of the dogs from the neural network training set were classified properly. However, only 31% of the dogs not included in the training process were accurately diagnosed. The poor neural-network based diagnosis of recordings that did not participate in the training process might have been caused by inappropriate representation of input data. Previous research has suggested characterizing signals according to certain features of the recorded data. This method, if employed, would reduce the noise and possibly improve the diagnostic abilities of the neural network.

Self-adaptive probabilistic roadmap generation for intelligent virtual agents

Agents inhabiting large scale environments are faced with the problem of generating maps by which they can navigate. One solution to this problem is to use probabilistic roadmaps which rely on selecting and connecting a set of points that describe the interconnectivity of free space. However, the time required to generate these maps can be prohibitive, and agents do not typically know the environment in advance. In this paper we show that the optimal combination of different point selection methods used to create the map is dependent on the environment, no point selection method dominates. This motivates a novel self-adaptive approach for an agent to combine several point selection methods. The success rate of our approach is comparable to the state of the art and the generation cost is substantially reduced. Self-adaptation therefore enables a more efficient use of the agent's resources. Results are presented for both a set of archetypal scenarios and large scale virtual environments based in Second Life, representing real locations in London.

Network construction and graph theoretical analysis of functional language networks in pediatric epilepsy

This dissertation introduces a new approach for assessing the effects of pediatric epilepsy on the language connectome. Two novel data-driven network construction approaches are presented. These methods rely on connecting different brain regions using either extent or intensity of language related activations as identified by independent component analysis of fMRI data. An auditory description decision task (ADDT) paradigm was used to activate the language network for 29 patients and 30 controls recruited from three major pediatric hospitals. Empirical evaluations illustrated that pediatric epilepsy can cause, or is associated with, a network efficiency reduction. Patients showed a propensity to inefficiently employ the whole brain network to perform the ADDT language task; on the contrary, controls seemed to efficiently use smaller segregated network components to achieve the same task. To explain the causes of the decreased efficiency, graph theoretical analysis was carried out. The analysis revealed no substantial global network feature differences between the patient and control groups. It also showed that for both subject groups the language network exhibited small-world characteristics; however, the patient's extent of activation network showed a tendency towards more random networks. It was also shown that the intensity of activation network displayed ipsilateral hub reorganization on the local level. The left hemispheric hubs displayed greater centrality values for patients, whereas the right hemispheric hubs displayed greater centrality values for controls. This hub hemispheric disparity was not correlated with a right atypical language laterality found in six patients. Finally it was shown that a multi-level unsupervised clustering scheme based on self-organizing maps, a type of artificial neural network, and k-means was able to fairly and blindly separate the subjects into their respective patient or control groups. The clustering was initiated using the local nodal centrality measurements only. Compared to the extent of activation network, the intensity of activation network clustering demonstrated better precision. This outcome supports the assertion that the local centrality differences presented by the intensity of activation network can be associated with focal epilepsy.^

The influence of interhemispheric connections on ongoing and evoked orientation preference maps and spiking activity in the cat primary visual cortex

In the primary visual cortex, neurons with similar physiological features are clustered together in columns extending through all six cortical layers. These columns form modular orientation preference maps. Long-range lateral fibers are associated to the structure of orientation maps since they do not connect columns randomly; they rather cluster in regular intervals and interconnect predominantly columns of neurons responding to similar stimulus features. Single orientation preference maps – the joint activation of domains preferring the same orientation - were observed to emerge spontaneously and it was speculated whether this structured ongoing activation could be caused by the underlying patchy lateral connectivity. Since long-range lateral connections share many features, i.e. clustering, orientation selectivity, with visual inter-hemispheric connections (VIC) through the corpus callosum we used the latter as a model for long-range lateral connectivity. In order to address the question of how the lateral connectivity contributes to spontaneously generated maps of one hemisphere we investigated how these maps react to the deactivation of VICs originating from the contralateral hemisphere. To this end, we performed experiments in eight adult cats. We recorded voltage-sensitive dye (VSD) imaging and electrophysiological spiking activity in one brain hemisphere while reversible deactivating the other hemisphere with a cooling technique. In order to compare ongoing activity with evoked activity patterns we first presented oriented gratings as visual stimuli. Gratings had 8 different orientations distributed equally between 0º and 180º. VSD imaged frames obtained during ongoing activity conditions were then compared to the averaged evoked single orientation maps in three different states: baseline, cooling and recovery. Kohonen self-organizing maps were also used as a means of analysis without prior assumption (like the averaged single condition maps) on ongoing activity. We also evaluated if cooling had a differential effect on evoked and ongoing spiking activity of single units. We found that deactivating VICs caused no spatial disruption on the structure of either evoked or ongoing activity maps. The frequency with which a cardinally preferring (0º or 90º) map would emerge, however, decreased significantly for ongoing but not for evoked activity. The same result was found by training self-organizing maps with recorded data as input. Spiking activity of cardinally preferring units also decreased significantly for ongoing when compared to evoked activity. Based on our results we came to the following conclusions: 1) VICs are not a determinant factor of ongoing map structure. Maps continued to be spontaneously generated with the same quality, probably by a combination of ongoing activity from local recurrent connections, thalamocortical loop and feedback connections. 2) VICs account for a cardinal bias in the temporal sequence of ongoing activity patterns, i.e. deactivating VIC decreases the probability of cardinal maps to emerge spontaneously. 3) Inter- and intrahemispheric long-range connections might serve as a grid preparing primary visual cortex for likely junctions in a larger visual environment encompassing the two hemifields.

The influence of interhemispheric connections on ongoing and evoked orientation preference maps and spiking activity in the cat primary visual cortex

In the primary visual cortex, neurons with similar physiological features are clustered together in columns extending through all six cortical layers. These columns form modular orientation preference maps. Long-range lateral fibers are associated to the structure of orientation maps since they do not connect columns randomly; they rather cluster in regular intervals and interconnect predominantly columns of neurons responding to similar stimulus features. Single orientation preference maps – the joint activation of domains preferring the same orientation - were observed to emerge spontaneously and it was speculated whether this structured ongoing activation could be caused by the underlying patchy lateral connectivity. Since long-range lateral connections share many features, i.e. clustering, orientation selectivity, with visual inter-hemispheric connections (VIC) through the corpus callosum we used the latter as a model for long-range lateral connectivity. In order to address the question of how the lateral connectivity contributes to spontaneously generated maps of one hemisphere we investigated how these maps react to the deactivation of VICs originating from the contralateral hemisphere. To this end, we performed experiments in eight adult cats. We recorded voltage-sensitive dye (VSD) imaging and electrophysiological spiking activity in one brain hemisphere while reversible deactivating the other hemisphere with a cooling technique. In order to compare ongoing activity with evoked activity patterns we first presented oriented gratings as visual stimuli. Gratings had 8 different orientations distributed equally between 0º and 180º. VSD imaged frames obtained during ongoing activity conditions were then compared to the averaged evoked single orientation maps in three different states: baseline, cooling and recovery. Kohonen self-organizing maps were also used as a means of analysis without prior assumption (like the averaged single condition maps) on ongoing activity. We also evaluated if cooling had a differential effect on evoked and ongoing spiking activity of single units. We found that deactivating VICs caused no spatial disruption on the structure of either evoked or ongoing activity maps. The frequency with which a cardinally preferring (0º or 90º) map would emerge, however, decreased significantly for ongoing but not for evoked activity. The same result was found by training self-organizing maps with recorded data as input. Spiking activity of cardinally preferring units also decreased significantly for ongoing when compared to evoked activity. Based on our results we came to the following conclusions: 1) VICs are not a determinant factor of ongoing map structure. Maps continued to be spontaneously generated with the same quality, probably by a combination of ongoing activity from local recurrent connections, thalamocortical loop and feedback connections. 2) VICs account for a cardinal bias in the temporal sequence of ongoing activity patterns, i.e. deactivating VIC decreases the probability of cardinal maps to emerge spontaneously. 3) Inter- and intrahemispheric long-range connections might serve as a grid preparing primary visual cortex for likely junctions in a larger visual environment encompassing the two hemifields.

Contributions to 3D Data Registration and Representation

Nowadays, new computers generation provides a high performance that enables to build computationally expensive computer vision applications applied to mobile robotics. Building a map of the environment is a common task of a robot and is an essential part to allow the robots to move through these environments. Traditionally, mobile robots used a combination of several sensors from different technologies. Lasers, sonars and contact sensors have been typically used in any mobile robotic architecture, however color cameras are an important sensor due to we want the robots to use the same information that humans to sense and move through the different environments. Color cameras are cheap and flexible but a lot of work need to be done to give robots enough visual understanding of the scenes. Computer vision algorithms are computational complex problems but nowadays robots have access to different and powerful architectures that can be used for mobile robotics purposes. The advent of low-cost RGB-D sensors like Microsoft Kinect which provide 3D colored point clouds at high frame rates made the computer vision even more relevant in the mobile robotics field. The combination of visual and 3D data allows the systems to use both computer vision and 3D processing and therefore to be aware of more details of the surrounding environment. The research described in this thesis was motivated by the need of scene mapping. Being aware of the surrounding environment is a key feature in many mobile robotics applications from simple robotic navigation to complex surveillance applications. In addition, the acquisition of a 3D model of the scenes is useful in many areas as video games scene modeling where well-known places are reconstructed and added to game systems or advertising where once you get the 3D model of one room the system can add furniture pieces using augmented reality techniques. In this thesis we perform an experimental study of the state-of-the-art registration methods to find which one fits better to our scene mapping purposes. Different methods are tested and analyzed on different scene distributions of visual and geometry appearance. In addition, this thesis proposes two methods for 3d data compression and representation of 3D maps. Our 3D representation proposal is based on the use of Growing Neural Gas (GNG) method. This Self-Organizing Maps (SOMs) has been successfully used for clustering, pattern recognition and topology representation of various kind of data. Until now, Self-Organizing Maps have been primarily computed offline and their application in 3D data has mainly focused on free noise models without considering time constraints. Self-organising neural models have the ability to provide a good representation of the input space. In particular, the Growing Neural Gas (GNG) is a suitable model because of its flexibility, rapid adaptation and excellent quality of representation. However, this type of learning is time consuming, specially for high-dimensional input data. Since real applications often work under time constraints, it is necessary to adapt the learning process in order to complete it in a predefined time. This thesis proposes a hardware implementation leveraging the computing power of modern GPUs which takes advantage of a new paradigm coined as General-Purpose Computing on Graphics Processing Units (GPGPU). Our proposed geometrical 3D compression method seeks to reduce the 3D information using plane detection as basic structure to compress the data. This is due to our target environments are man-made and therefore there are a lot of points that belong to a plane surface. Our proposed method is able to get good compression results in those man-made scenarios. The detected and compressed planes can be also used in other applications as surface reconstruction or plane-based registration algorithms. Finally, we have also demonstrated the goodness of the GPU technologies getting a high performance implementation of a CAD/CAM common technique called Virtual Digitizing.

Network Construction and Graph Theoretical Analysis of Functional Language Networks in Pediatric Epilepsy

This dissertation introduces a new approach for assessing the effects of pediatric epilepsy on the language connectome. Two novel data-driven network construction approaches are presented. These methods rely on connecting different brain regions using either extent or intensity of language related activations as identified by independent component analysis of fMRI data. An auditory description decision task (ADDT) paradigm was used to activate the language network for 29 patients and 30 controls recruited from three major pediatric hospitals. Empirical evaluations illustrated that pediatric epilepsy can cause, or is associated with, a network efficiency reduction. Patients showed a propensity to inefficiently employ the whole brain network to perform the ADDT language task; on the contrary, controls seemed to efficiently use smaller segregated network components to achieve the same task. To explain the causes of the decreased efficiency, graph theoretical analysis was carried out. The analysis revealed no substantial global network feature differences between the patient and control groups. It also showed that for both subject groups the language network exhibited small-world characteristics; however, the patient’s extent of activation network showed a tendency towards more random networks. It was also shown that the intensity of activation network displayed ipsilateral hub reorganization on the local level. The left hemispheric hubs displayed greater centrality values for patients, whereas the right hemispheric hubs displayed greater centrality values for controls. This hub hemispheric disparity was not correlated with a right atypical language laterality found in six patients. Finally it was shown that a multi-level unsupervised clustering scheme based on self-organizing maps, a type of artificial neural network, and k-means was able to fairly and blindly separate the subjects into their respective patient or control groups. The clustering was initiated using the local nodal centrality measurements only. Compared to the extent of activation network, the intensity of activation network clustering demonstrated better precision. This outcome supports the assertion that the local centrality differences presented by the intensity of activation network can be associated with focal epilepsy.

An intelligent data mining technique for product quality improvement

Advances in data mining have provided techniques for automatically discovering underlying knowledge and extracting useful information from large volumes of data. Data mining offers tools for quick discovery of relationships, patterns and knowledge in large complex databases. Application of data mining to manufacturing is relatively limited mainly because of complexity of manufacturing data. Growing self organizing map (GSOM) algorithm has been proven to be an efficient algorithm to analyze unsupervised DNA data. However, it produced unsatisfactory clustering when used on some large manufacturing data. In this paper a data mining methodology has been proposed using a GSOM tool which was developed using a modified GSOM algorithm. The proposed method is used to generate clusters for good and faulty products from a manufacturing dataset. The clustering quality (CQ) measure proposed in the paper is used to evaluate the performance of the cluster maps. The paper also proposed an automatic identification of variables to find the most probable causative factor(s) that discriminate between good and faulty product by quickly examining the historical manufacturing data. The proposed method offers the manufacturers to smoothen the production flow and improve the quality of the products. Simulation results on small and large manufacturing data show the effectiveness of the proposed method.

Neural network-based detection of virtual environment anomalies

The increasingly widespread use of large-scale 3D virtual environments has translated into an increasing effort required from designers, developers and testers. While considerable research has been conducted into assisting the design of virtual world content and mechanics, to date, only limited contributions have been made regarding the automatic testing of the underpinning graphics software and hardware. In the work presented in this paper, two novel neural network-based approaches are presented to predict the correct visualization of 3D content. Multilayer perceptrons and self-organizing maps are trained to learn the normal geometric and color appearance of objects from validated frames and then used to detect novel or anomalous renderings in new images. Our approach is general, for the appearance of the object is learned rather than explicitly represented. Experiments were conducted on a game engine to determine the applicability and effectiveness of our algorithms. The results show that the neural network technology can be effectively used to address the problem of automatic and reliable visual testing of 3D virtual environments.

Classification of magnetic resonance brain images using wavelets as input to support vector machine and neural network

In this paper. we propose a novel method using wavelets as input to neural network self-organizing maps and support vector machine for classification of magnetic resonance (MR) images of the human brain. The proposed method classifies MR brain images as either normal or abnormal. We have tested the proposed approach using a dataset of 52 MR brain images. Good classification percentage of more than 94% was achieved using the neural network self-organizing maps (SOM) and 98% front support vector machine. We observed that the classification rate is high for a Support vector machine classifier compared to self-organizing map-based approach.

Multiple contour extraction from graylevel images using an artificial neural network

For active contour modeling (ACM), we propose a novel self-organizing map (SOM)-based approach, called the batch-SOM (BSOM), that attempts to integrate the advantages of SOM- and snake-based ACMs in order to extract the desired contours from images. We employ feature points, in the form of ail edge-map (as obtained from a standard edge-detection operation), to guide the contour (as in the case of SOM-based ACMs) along with the gradient and intensity variations in a local region to ensure that the contour does not "leak" into the object boundary in case of faulty feature points (weak or broken edges). In contrast with the snake-based ACMs, however, we do not use an explicit energy functional (based on gradient or intensity) for controlling the contour movement. We extend the BSOM to handle extraction of contours of multiple objects, by splitting a single contour into as many subcontours as the objects in the image. The BSOM and its extended version are tested on synthetic binary and gray-level images with both single and multiple objects. We also demonstrate the efficacy of the BSOM on images of objects having both convex and nonconvex boundaries. The results demonstrate the superiority of the BSOM over others. Finally, we analyze the limitations of the BSOM.

Modeling habits as self-sustaining patterns of sensorimotor behavior

In the recent history of psychology and cognitive neuroscience, the notion of habit has been reduced to a stimulus-triggered response probability correlation. In this paper we use a computational model to present an alternative theoretical view (with some philosophical implications), where habits are seen as self-maintaining patterns of behavior that share properties in common with self-maintaining biological processes, and that inhabit a complex ecological context, including the presence and influence of other habits. Far from mechanical automatisms, this organismic and self-organizing concept of habit can overcome the dominating atomistic and statistical conceptions, and the high temporal resolution effects of situatedness, embodiment and sensorimotor loops emerge as playing a more central, subtle and complex role in the organization of behavior. The model is based on a novel "iterant deformable sensorimotor medium (IDSM)," designed such that trajectories taken through sensorimotor-space increase the likelihood that in the future, similar trajectories will be taken. We couple the IDSM to sensors and motors of a simulated robot, and show that under certain conditions, the IDSM conditions, the IDSM forms self-maintaining patterns of activity that operate across the IDSM, the robot's body, and the environment. We present various environments and the resulting habits that form in them. The model acts as an abstraction of habits at a much needed sensorimotor "meso-scale" between microscopic neuron-based models and macroscopic descriptions of behavior. Finally, we discuss how this model and extensions of it can help us understand aspects of behavioral self-organization, historicity and autonomy that remain out of the scope of contemporary representationalist frameworks.

Guided self-organization in a dynamic embodied system based on attractor selection mechanism

Guided self-organization can be regarded as a paradigm proposed to understand how to guide a self-organizing system towards desirable behaviors, while maintaining its non-deterministic dynamics with emergent features. It is, however, not a trivial problem to guide the self-organizing behavior of physically embodied systems like robots, as the behavioral dynamics are results of interactions among their controller, mechanical dynamics of the body, and the environment. This paper presents a guided self-organization approach for dynamic robots based on a coupling between the system mechanical dynamics with an internal control structure known as the attractor selection mechanism. The mechanism enables the robot to gracefully shift between random and deterministic behaviors, represented by a number of attractors, depending on internally generated stochastic perturbation and sensory input. The robot used in this paper is a simulated curved beam hopping robot: a system with a variety of mechanical dynamics which depends on its actuation frequencies. Despite the simplicity of the approach, it will be shown how the approach regulates the probability of the robot to reach a goal through the interplay among the sensory input, the level of inherent stochastic perturbation, i.e., noise, and the mechanical dynamics. © 2014 by the authors; licensee MDPI, Basel, Switzerland.

Grid Cell Hexagonal Patterns Formed by Fast Self-Organized Learning within Entorhinal Cortex

Grid cells in the dorsal segment of the medial entorhinal cortex (dMEC) show remarkable hexagonal activity patterns, at multiple spatial scales, during spatial navigation. How these hexagonal patterns arise has excited intense interest. It has previously been shown how a selforganizing map can convert firing patterns across entorhinal grid cells into hippocampal place cells that are capable of representing much larger spatial scales. Can grid cell firing fields also arise during navigation through learning within a self-organizing map? A neural model is proposed that converts path integration signals into hexagonal grid cell patterns of multiple scales. This GRID model creates only grid cell patterns with the observed hexagonal structure, predicts how these hexagonal patterns can be learned from experience, and can process biologically plausible neural input and output signals during navigation. These results support a unified computational framework for explaining how entorhinal-hippocampal interactions support spatial navigation.