On the identification of neural responses

The most common form of measuring electrical responses of nerve cells is the recording of a given cell's spike train profile to the parameters of a given input signal. In this paper we consider the conditions under which it is possible to relate such response measures to (a) the properties of the cell's underlying activity characteristics, (b) the neural network, and (c) the input signal.This work was partially supported by the Natural Sciences and Engineering Research Council of Canada under Grant A-4345 to M.N.O. and under Grant A-4395 to T.M.C. through the University of Alberta     

An inverse problem in neural processing

Any neural network aimed at the coding sensory events must contain computational properties which generally allow the organism to reconstruct the input signals with some degree of accuracy-else the association between stimulus and response would, at best, be uncertain. In this paper we investigate the problem of reconstructing external input signals to neural networks when the activity profiles of only some of its member cells are known. The evolution and activities of such cells are defined by an earlier formulation of one of us (Ouztöreli 1979) and, here, we restrict our application to local curcuits within the vertebrate retina. Solutions to this inverse coding problem are presented for specific network equations and examplified with 1, 3, and 5 neuron cases.This work was partially supported by the Natural Sciences and Engineering Research Council of Canada under Grant A-4345 to M.N.O. and grant A-4395 to T.M.C. through the University of Alberta     

Activity analysis of neural networks

In a series of articles (Leung et al., 1973, 1974; Ogztöreli, 1972, 1975, 1978, 1979; Stein et al., 1974) we have investigated some of the physiologically significant properties of a general neural model. In these papers the nature of the oscillations occuring in the model has been briefly analyzed by omitting the effects of the discrete time-lags in the interaction of neurons, although these time-lags were incoporated in the general model. In the present work we investigate the effects of the time-lags on the oscillations which are intrinsic to the neural model, depending on the structural parameters such as external inputs, interaction coefficients, self-inhibition, self-excitation and selfadaptation coefficients. The numerical solution of the neural model, the computation of the steady-state solutions and the natural modes of the oscillations around the steady-state solutions are described.This work was partly supported by the Natural Sciences and Engineering Research Council of Canada under Grant NRC-A-4345 through the University of Alberta     

Response analysis of vertebrate retina

In a recent work (Ouztöreli, 1980) a mathematical model for studying the neural activities in a vertebrate retina has been investigated, where the basic network contains five interconnected neurons: a receptor cell, a bipolar cell, a horizontal cell, an amacrine cell, and a retinal ganglion cell. More recently, in (Ouztöreli and O'Mara, 1980) the basic network has been extended to a larger network containing twelve neurons. In both of these works, the performances of the basic and extended models were discussed under different structural and processing conditions with constant inputs by using the results of one of our earlier work (Ouztöreli, 1979). In the present paper we investigate by simulations the responses of the basic retinal network to piecewise constant and periodic inputs. The step and frequency responses of the extended retinal network will be discussed in a forthcoming paper.This work was partially supported by the Natural Sciences and Engineering Research Council of Canada under Grant A-4345 through the University of alberta     

Analysis of a model for antagonistic muscles

In the present work a linear model for a pair of antogonistic muscles is analysed. Each constituent muscle in this model is identical to ones considered previously (Stein and Ouztöreli, 1976). Analytical properties of the antagonistic muscles and dynamics of the system are described and some numerical results are discussed. The natural modes of the system are determined by a fourth order polynomial, which most commonly has one pair of conjugate complex roots and two negative real roots. The filtering of neural inputs through the active state properties of the muscle increases the order of the system to fifth order for these inputs.This work was partly supported by the National Scientific and Engineering Research Council of Canada Grant NRC-A4345 and by the Medical Research Council of Canada Grant MRC-MT-3307 through the University of Alberta     

Underlying neural computations for some visual phenomena

In this paper we examine how a large array of neurons, and their associated neural circuitry, may determine known receptive field profile types and some well-known visual phenomena including Mach bands, edge enhancement, and visual masking of one signal by another. The neural model has a spatio-temporal structure and is described by a nonlinear integropartial differential difference equation with an isotropic Gabor kernel — a Gaussian apertured cosine modulation. Several simulations are presented.This project was partially funded by Grants A4345 to M. N. Ouztöreli and A2568 to T. M. Caelli from the Natural Sciences and Engineering Research Council of Canada through the University of Alberta     

Immunohistochemical localization of human α2macroglobulin in connective tissue

Summary The localization of ₂ macroglobulin (₂M) has been examined by an indirect immunofluorescent technique in frozen sections of various human tissues. The results indicate that ₂M is present only in connective tissues and blood. The outer medulla of the kidney and the submucosa of the gut showed the strongest reaction. Epithelia or endothelial cells were unreactive. In liver, only the Kupffer cells were stained. These results were confirmed with the immunoperoxidase technique and by the study of tissue extracts in crossed immunoelectrophoresis (CIE). As a positive control a polyspecific antiserum prepared against whole human fibroblasts as well as anti-albumin were used. Our findings are interpreted in the light of the observations that ₂M is synthesized and selectively ingested by fibroblasts.Abbreviation used in this Paper ₂M ₂macroglobulin This work was supported by a Grant from the Belgian Cancer Fund (Algemene Spaar en Lijfrente Kas), by Grant 3.0043.79 (Fonds voor Geneeskundig Wetenschappelijk Onderzoek) and by Research Fund OT/VII/30 (K.U. Leuven)F. Van Leuven is a Post Doctoral Research Fellow of the American Cystic Fibrosis Foundation     

Self-organization of day cycle and hierarchical associative memory in “live” neural network

The live neural network model is proposed on the basis of live neuron model and optimal learning rule. By means of numerical simulation the initial stages of neural network self-organization have been shown: (1) the formation of two activity forms, which are identified with sleep and awaking, and (2) the self-organization of hierarchical associative memory when feeding a receptor excitation to the neural network. The energetic profit of self-organization is demonstrated. The formation of neural ensembles, playing the role of generalized neurons, is obtained.     

Bayesian neural networks

A neural network that uses the basic Hebbian learning rule and the Bayesian combination function is defined. Analogously to Hopfield's neural network, the convergence for the Bayesian neural network that asynchronously updates its neurons' states is proved. The performance of the Bayesian neural network in four medical domains is compared with various classification methods. The Bayesian neural network uses more sophisticated combination function than Hopfield's neural network and uses more economically the available information. The naive Bayesian classifier typically outperforms the basic Bayesian neural network since iterations in network make too many mistakes. By restricting the number of iterations and increasing the number of fixed points the network performs better than the naive Bayesian classifier. The Bayesian neural network is designed to learn very quickly and incrementally.     

A gene apparently determining the extent of sialylation of lysosomal α-mannosidase in mouse liver

Organ-specific electrophoretic heterogeneity of lysosomal -mannosidase has been observed within individual strains of inbred mice. Polymorphism between C57BL/6J and CBA/J for liver lysosomal -mannosidase is determined by a single genetic locus on chromosome 5 and appears to be the result of differences in sialylation of the lysosomal enzyme. Two different patterns of expression of development of the liver electrophoretic forms have been observed.Supported in part by Grant GM-19521 from the U.S. Public Health Service. One of the authors (M.D.) was supported in part from USPHS Grant TAO-CA05016.     

Activation of LA-N-2 Cell Phospholipase D by Amyloid Beta Protein (25–35)

Amyloid protein is the major protein component of neuritic plaques found in the brain of Alzheimer's disease. The activation of phospholipase D by amyloid beta protein (25–35), quisqualate and phorbol 12, 13-dibutyrate was investigated in LA-N-2 cells by measuring phosphatidylethanol formation. The activation of phospholipase D by quisqualate and AP (25–35) was calcium-independent. The AP (25–35) and quisqualate activation of phospholipase D appeared to be mediated through a pertussis toxin-sensitive GTP-binding protein. Phospholipase D activation by AP (25–35), quisqualate and phorbol dibutyrate was not blunted by the protein kinase C inhibitors, staurosporine, H-7 and RO-31-8220. However, it was abolished by overnight exposure to phorbol dibutyrate. This activation of phospholipase D was prevented by the tyrosine kinase inhibitor, genistein but not by tyrophostin A. Several excitatory amino acid antagonists were tested for their ability to prevent the phospholipase D activation by quisqualate and AP (25–35). Only NBQX was effective with an IC₅₀ of 75 M for AP (25–35) and quisqualate. Activation of phospholipase D by AP or quisqualate was absent in LA-N-2 cells previously desensitized by quisqualate or AP (25–35), but the activation by phorbol dibutyrate was unaltered. The responsiveness to AP and quisqualate in previously desensitized cells reappeared subsequent to a period of resensitization. The observations with the antagonist NBQX, and the desensitization and resensitization experiments, are consistent with a receptor occupancy mediated activation of phospholipase D by quisqualate and by AP (25–35).     

A model for neural representation of binocular disparity in striate cortex: distributed representation and veto mechanism

A model in striate cortex is proposed for a distributed neural representation of binocular disparity with a simple cell. In the model, disparity is represented by far, near and tuned inhibitory simple cells. However, the representation will be vetoed by model cells where disparity is excessively large. The veto mechanism consists of a neural network of the model cell which received output from simple cells and which interacts with neighbors. The mechanism is necessary, the model cell responds like a simple cell, and the network is physiologically plausible in the brain. Computer simulation on the neural network model with random dot stereography indicates reasonable performance.     

Langevin machine: a neural network based on stochastically justifiable sigmoidal function

In neural networks the activation process controls the output as a nonlinear function of the input; and, this output remains bounded between limits as decided by a logistic function known as the sigmoid (S-shaped). Presently, by applying the considerations of Maxwell-Boltzmann statistics, the Langevin function is shown as the appropriate and justifiable sigmoid (instead of the conventional hyperbolic tangent function) to depict the bipolar nonlinear logic-operation enunciated by the collective stochastical response of artificial neurons under activation. That is, the graded response of a large network of neurons such as Hopfield's can be stochastically justified via the proposed model. The model is consistent with the established link between the Hopfield model and the statistical mechanics. The possible outcomes and implications of using the Langevin function (in lieu of conventional hyperbolic tangent and/or exponential sigmoids) in determining nonlinear decision boundaries, in characterizing the neural networks by the Langevin machine versus the Boltzmann machine, in sharpening and annealing schedules and in the optimization of nonlinear detector performance are discussed.     

Derivative-free neural network for optimizing the scoring functions associated with dynamic programming of pairwise-profile alignment

Background

A profile-comparison method with position-specific scoring matrix (PSSM) is among the most accurate alignment methods. Currently, cosine similarity and correlation coefficients are used as scoring functions of dynamic programming to calculate similarity between PSSMs. However, it is unclear whether these functions are optimal for profile alignment methods. By definition, these functions cannot capture nonlinear relationships between profiles. Therefore, we attempted to discover a novel scoring function, which was more suitable for the profile-comparison method than existing functions, using neural networks.

Results

Although neural networks required derivative-of-cost functions, the problem being addressed in this study lacked them. Therefore, we implemented a novel derivative-free neural network by combining a conventional neural network with an evolutionary strategy optimization method used as a solver. Using this novel neural network system, we optimized the scoring function to align remote sequence pairs. Our results showed that the pairwise-profile aligner using the novel scoring function significantly improved both alignment sensitivity and precision relative to aligners using existing functions.

Conclusions

We developed and implemented a novel derivative-free neural network and aligner (Nepal) for optimizing sequence alignments. Nepal improved alignment quality by adapting to remote sequence alignments and increasing the expressiveness of similarity scores. Additionally, this novel scoring function can be realized using a simple matrix operation and easily incorporated into other aligners. Moreover our scoring function could potentially improve the performance of homology detection and/or multiple-sequence alignment of remote homologous sequences. The goal of the study was to provide a novel scoring function for profile alignment method and develop a novel learning system capable of addressing derivative-free problems. Our system is capable of optimizing the performance of other sophisticated methods and solving problems without derivative-of-cost functions, which do not always exist in practical problems. Our results demonstrated the usefulness of this optimization method for derivative-free problems.

     

Optimal task performance of antagonistic muscles

Movements against a variety of loads are relatively invariant in form. These movements are controlled in general by antagonistic groups of muscles. In this paper optimal control strategies are computed for coupling antagonistic muscles so as to minimize deviations from a desired trajectory. Simulations are presented for linear and nonlinear decision functions linking control of the two muscles for a variety of movements in a way that may be compared with experimental observations.This work was partly supported by the Natural Sciences and Engineering Research Council of Canada Grant NSERC OGP-4345 and by the Medical Research Council of Canada Grant MRC PG-47 through the University of Alberta