Pattern computation in neural communication systems

Biological data suggests that activity patterns emerging in small- and large-scale neural systems may play an important role in performing the functions of the neural system, and in particular, neural computations. It is proposed in this paper that neural systems can be understood in terms of pattern computation and abstract communication systems theory. It is shown that analysing high-resolution surface EEG data, it is possible to determine abstract probabilistic rules that describe how emerging activity patterns follow earlier activity patterns. The results indicate the applicability of the proposed approach for understanding the working of complex neural systems.     

A canonical form for neural nets without circles

By “neural net” will be meant “neural net without circles.” Every neural net effects a transformation from inputs (i.e., firing patterns of the input neurons) to outputs (firing patterns of the output neurons). Two neural nets will be calledequivalent if they effect the same transformation from inputs to outputs. A canonical form is found for neural nets with respect to equivalence; i.e., a class of neural nets is defined, no two of which are equivalent, and which contains a neural net equivalent to any given neural net. This research was supported by the U.S. Air Force under Contract AF 49(638)-414 monitored by the Air Force Office of Scientific Research.     

Fly vision: neural mechanisms of motion computation

A wide range of novel approaches are being used to dissect the visual system of the fly, both the neural networks of motion detection and the performance of these networks under complex natural stimulus conditions.     

FPNA: interaction between FPGA and neural computation

Neural networks are usually considered as naturally parallel computing models. But the number of operators and the complex connection graph of standard neural models can not be directly handled by digital hardware devices. More particularly, several works show that programmable digital hardware is a real opportunity for flexible hardware implementations of neural networks. And yet many area and topology problems arise when standard neural models are implemented onto programmable circuits such as FPGAs, so that the fast FPGA technology improvements can not be fully exploited. Therefore neural network hardware implementations need to reconcile simple hardware topologies with complex neural architectures. The theoretical and practical framework developed, allows this combination thanks to some principles of configurable hardware that are applied to neural computation: Field Programmable Neural Arrays (FPNA) lead to powerful neural architectures that are easy to map onto FPGAs, thanks to a simplified topology and an original data exchange scheme. This paper shows how FPGAs have led to the definition of the FPNA computation paradigm. Then it shows how FPNAs contribute to current and future FPGA-based neural implementations by solving the general problems that are raised by the implementation of complex neural networks onto FPGAs.     

Escape behavior: linking neural computation to action

A new study uses a combination of physiological and optogenetic techniques to identify visual neurons in fruit flies that detect approaching objects, and whose activation is integral in escaping an oncoming threat.     

Frizzled7 mediates canonical Wnt signaling in neural crest induction

The neural crest is a multipotent cell population that migrates from the dorsal edge of the neural tube to various parts of the embryo where it differentiates into a remarkable variety of different cell types. Initial induction of neural crest is mediated by a combination of BMP, Wnt, FGF, Retinoic acid and Notch/Delta signaling. The two-signal model for neural crest induction suggests that BMP signaling induces the competence to become neural crest. The second signal involves Wnt acting through the canonical pathway and leads to expression of neural crest markers such as slug. Wnt signals from the neural plate, non-neural ectoderm and paraxial mesoderm have all been suggested to play a role in neural crest induction. We show that Xenopus frizzled7 (Xfz7) is expressed in the dorsal ectoderm including early neural crest progenitors and is a key mediator of the Wnt inductive signal. We demonstrate that Xfz7 expression is induced in response to a BMP antagonist, noggin, and that Xfz7 can induce neural crest specific genes in noggin-treated ectodermal explants (animal caps). Morpholino-mediated or dominant negative inhibition of Xfz7 inhibits Wnt induced Xslug expression in the animal cap assay and in the whole embryo leading to a loss of neural crest derived pigment cells. Full-length Xfz7 rescues the morpholino-induced phenotype, as does activated beta-catenin, suggesting that Xfz7 is signaling through the canonical pathway. We therefore demonstrate that Xfz7 is regulated by BMP antagonism and is required for neural crest induction by Wnt in the developing vertebrate embryo.     

Towards a general theory of neural computation based on prediction by single neurons

Although there has been tremendous progress in understanding the mechanics of the nervous system, there has not been a general theory of its computational function. Here I present a theory that relates the established biophysical properties of single generic neurons to principles of Bayesian probability theory, reinforcement learning and efficient coding. I suggest that this theory addresses the general computational problem facing the nervous system. Each neuron is proposed to mirror the function of the whole system in learning to predict aspects of the world related to future reward. According to the model, a typical neuron receives current information about the state of the world from a subset of its excitatory synaptic inputs, and prior information from its other inputs. Prior information would be contributed by synaptic inputs representing distinct regions of space, and by different types of non-synaptic, voltage-regulated channels representing distinct periods of the past. The neuron's membrane voltage is proposed to signal the difference between current and prior information ("prediction error" or "surprise"). A neuron would apply a Hebbian plasticity rule to select those excitatory inputs that are the most closely correlated with reward but are the least predictable, since unpredictable inputs provide the neuron with the most "new" information about future reward. To minimize the error in its predictions and to respond only when excitation is "new and surprising," the neuron selects amongst its prior information sources through an anti-Hebbian rule. The unique inputs of a mature neuron would therefore result from learning about spatial and temporal patterns in its local environment, and by extension, the external world. Thus the theory describes how the structure of the mature nervous system could reflect the structure of the external world, and how the complexity and intelligence of the system might develop from a population of undifferentiated neurons, each implementing similar learning algorithms.     

Statistical independence and neural computation in the leech ganglion

 In this report, the input/output relations in an isolated ganglion of the leech Hirudo medicinalis were studied by simultaneously using six or eight suction pipettes and two intracellular electrodes. Sensory input was mimicked by eliciting action potentials in mechanosensory neurons with intracellular electrodes. The integrated neural output was measured by recording extracellular voltage signals with pipettes sucking the roots and the connectives. A single evoked action potential activated electrical activity in at least a dozen different neurons, some of which were identified. This electrical activity was characterized by a high degree of temporal and spatial variability. The action potentials of coactivated neurons, i.e. activated by the same mechanosensory neuron, did not show any significant pairwise correlation. Indeed, the analysis of evoked action potentials indicates clear statistical independence among coactivated neurons, presumably originating from the independence of synaptic transmission at distinct synapses. This statistical independence may be used to increase reliability when neuronal activity is averaged or pooled. It is suggested that statistical independence among coactivated neurons may be a usual property of distributed processing of neuronal networks and a basic feature of neural computation. Received: 20 September 1999 / Accepted in revised form: 3 March 2000     

A neural net capable of competitive and cooperative computation

An algorithm of learning in multilayer threshold nets without feedbacks is proposed. The net is. built of threshold elements with binary inputs. During a learning process each input vector x is accompanied by a teacher's decision ({1,...,M}). The pairs (x[n], [n]) appear in successive steps independently according to some unknown stationary distribution p(x,). The problem of learning of a threshold net has been decomposed to a series of problems of learning of the threshold elements. The proposed learning algorithm of the threshold elements has a perceptron-like form. It was proven that a decision rule of the threshold net stabilizes after a finite number of steps. For definite classes {p(x, )} _* ^K of distributions p(x,), an optimal decision rule stabilizes after a finite number of steps. These classes {p(x, )} _* ^K also contain distributions describing learning processes with perturbations.     

Efficient computation via sparse coding in electrosensory neural networks

The electric sense combines spatial aspects of vision and touch with temporal features of audition. Its accessible neural architecture shares similarities with mammalian sensory systems and allows for recordings from successive brain areas to test hypotheses about neural coding. Further, electrosensory stimuli encountered during prey capture, navigation, and communication, can be readily synthesized in the laboratory. These features enable analyses of the neural circuitry that reveal general principles of encoding and decoding, such as segregation of information into separate streams and neural response sparsification. A systems level understanding arises via linkage between cellular differentiation and network architecture, revealed by in vitro and in vivo analyses, while computational modeling reveals how single cell dynamics and connectivity shape the sparsification process.     

Quantum-like model of behavioral response computation using neural oscillators

In this paper we propose the use of neural interference as the origin of quantum-like effects in the brain. We do so by using a neural oscillator model consistent with neurophysiological data. The model used was shown elsewhere to reproduce well the predictions of behavioral stimulus-response theory. The quantum-like effects are brought about by the spreading activation of incompatible oscillators, leading to an interference-like effect mediated by inhibitory and excitatory synapses.     

Ciliary margin transdifferentiation from neural retina is controlled by canonical Wnt signaling

The epithelial layers of the ciliary body (CB) and iris are non-neural structures that differentiate from the anterior region of the eyecup, the ciliary margin (CM). We show here that activation of the canonical Wnt signaling pathway is sufficient and necessary for the normal development of anterior eye structures. Pharmacological activation of beta-catenin signaling with lithium (Li(+)) treatment in retinal explants in vitro induced the ectopic expression of the CM markers Otx1 and Msx1. Cre-mediated stabilization of beta-catenin expression in the peripheral retina in vivo induced a cell autonomous upregulation of CM markers at the expense of neural retina (NR) markers and inhibited neurogenesis. Consistent with a cell autonomous conversion to peripheral eye fates, the proliferation index in the region of the retina that expressed stabilized beta-catenin was identical to the wild-type CM and there was an expansion of CB-like structures at later stages. Conversely, Cre-mediated inactivation of beta-catenin reduced CM marker expression as well as the size of the CM and CB/iris. Aberrant CB development in both mouse models was also associated with a reduction in the number of retinal stem cells in vitro. In summary, activation of canonical Wnt signaling is sufficient to promote the development of peripheral eyecup fates at the expense of the NR and is also required for the normal development of anterior eyecup structures.     

Normalization of a chromosomal contact map

ABSTRACT: BACKGROUND: Chromatin organization has been increasingly studied in relation with its important influence on DNA-related metabolic processes such as replication or regulation of gene expression. Since its original design ten years ago, capture of chromosome conformation (3C) has become an essential tool to investigate the overall conformation of chromosomes. It relies on the capture of long-range trans and cis interactions of chromosomal segments whose relative proportions in the final bank reflect their frequencies of interactions, hence their average spatial proximity. The recent coupling of 3C with deep sequencing approaches now allows the generation of high resolution genome-wide chromosomal contact maps. Different protocols have been used to generate such maps in various organisms. This includes mammals, drosophila and yeast. The massive amount of raw data generated by the genomic 3C has to be carefully processed to alleviate the various biases and byproducts generated by the experiments. Our study aims at proposing a simple normalization procedure to take into account these unwanted but inevitable events. RESULTS: Careful analysis of the raw data generated previously for budding yeast Saccharomyces cerevisiae led to the identification of three main biases affecting the final datasets, including an original bias resulting from the circularization of DNA molecules exhibiting specific lengths in accordance with laws from polymer physics. We then developed a simple normalization procedure to process the data and allow the generation of a normalized, highly contrasted, chromosomal contact map for S. cerevisiae. The same method was then extended to the first human genome contact map. Using the normalized data, we revisited the preferential interactions originally described between subsets of discrete chromosomal features. Notably, the detection of preferential interactions between tRNA in yeast and CTCF, PolII binding sites in human can vary with the normalization procedure used. CONCLUSIONS: We quantitatively reanalyzed the genomic 3C data obtained for S. cerevisiae, identified some of the biases inherent to the technique and proposed a simple normalization procedure to analyze them. Such an approach can be easily generalized for genomic 3C experiments in other organisms. More experiments and analysis will be necessary to reach optimal resolution and accuracies of the maps generated through these approaches. Working with cell population presenting highest levels of homogeneity will prove useful in this regards.     

Disparity tuning as simulated by a neural net

Previous research has suggested that the processing of binocular disparity in complex cells may be described with an energy formalism. The energy formalism allows for a representation of disparity by differences in the position or in the phase of monocular receptive subfields of binocular cells, or by combination of these two types. We studied the coding of disparities with an approach complementary to previous algorithmic investigations. Since realization of these representations is probably not genetically determined but learned during ontogeny, we used backpropagation networks to study which of these three possibilities were realized within neural nets. Three types of networks were trained with noise patterns in analogy to the three types of energy models. The networks learned the task and generalized to untrained correlated noise pattern input. Outputs were broadly tuned to spatial frequency and did not respond to anti-correlated noise patterns. Although the energy model was not explicitly implemented, we could analyze the outputs of the networks using predictions of the energy formalism. After learning was completed, the model neurons preferred position shifts over phase shifts in representing disparity. We discuss the general meaning of these findings and the correspondences and deviations between the energy model, V1 neurons, and our networks. Received: 6 August 1999 / Accepted in revised form: 26 January 2000     

Semantic commitments as a mode of non-programmable computation in the brain.

The natural language processor in the brain can cope with non-programmable computation. The average number of different lexical meanings per word serves as a quantitative figure in terms of which the extent of being non-programmable can be evaluated. The possible maximum average number of different lexical meanings per word that the brain of the subject reading the text can cope with while comprehending the context is found to be 3.3 with its standard deviation 0.15, beyond which the brain can no more succeed in comprehending the context. In contrast, the maximum average number of different lexical meanings per word that would make lexical disambiguation programmable is e = 2.718. Natural language processing in the brain is non-programmable in the sense that the manageable average number of different meanings per word is greater than e, but does not exceed roughly 3.3.     

Neucrin, a novel secreted antagonist of canonical Wnt signaling, plays roles in developing neural tissues in zebrafish

Wnt signaling plays crucial roles in neural development. We previously identified Neucrin, a neural-specific secreted antagonist of canonical Wnt/β-catenin signaling, in humans and mice. Neucrin has one cysteine-rich domain, in which the positions of 10 cysteine residues are similar to those in the second cysteine-rich domain of Dickkopfs, secreted Wnt antagonists. Here, we have identified zebrafish neucrin to understand its roles in vivo. Zebrafish Neucrin also has one cysteine-rich domain, which is significantly similar to that of mouse Neucrin. Zebrafish neucrin was also predominantly expressed in developing neural tissues. To examine roles of neucrin in neural development, we analyzed neucrin knockdown embryos. Neural development in zebrafish embryos was impaired by the knockdown of neucrin. The knockdown of neucrin caused increased expression of the Wnt/β-catenin target genes. In contrast, overexpression of neucrin reduced the expression of the Wnt/β-catenin target genes. The knockdown of neucrin affected specification of dorsal region in the midbrain and hindbrain. The knockdown of neucrin also suppressed neuronal differentiation and caused increased cell proliferation and apoptosis in developing neural tissues. Neucrin is a unique secreted Wnt antagonist that is predominantly expressed in developing neural tissues and plays roles in neural development in zebrafish.     

The enhancement of neural growth by amino-functionalization on carbon nanotubes as a neural electrode

This paper reports the success of amino-functionalization on multi-walled carbon nanotubes (MWCNTs) to promote neuronal cells growth on MWCNT electrode for extracellular recording, attributed to the formation of positive charge of NH(2) molecules on their surfaces. Besides, the surface of MWCNT electrode becomes hydrophilic after amino-functionalization (AF-MWCNTs) which can enhance electrical conductivity because of lower MWCNT/electrolyte interfacial impedance and higher interfacial capacitance. Durability tests show that electrical characteristics of the MWCNTs treated by 2 wt% 1,4-diaminobutane solution (2 wt%-AF-MWCNTs) can last for at least six months in air ambient. The neural recording of crayfish shows that 2 wt%-AF-MWCNTs can provide better capability on detecting action potentials of caudal photoreceptor (CPR) interneuron compared to suction glass pipette from the evidence of a higher S/N ratio (126 versus 23). The amino-functionalized MWCNT electrode is feasible for long-term recording application according to the results of biocompatibility tests. As the MWCNTs were directly synthesized on Si-based substrates by catalyst-assisted thermal chemical vapor deposition (CVD) at a low temperature (400 °C), these self-aligned MWCNT electrodes could be friendly implemented in integrated circuits fabrications.     

Molecular networks as a sub-neural factor of neural networks

We describe a new approach in the research of neural networks. This research is based on molecular networks in the neuron. If we use molecular networks as a sub-neuron factor of neural networks, it is a more realistic approach than today's concepts in this new computer technology field, because the artificial neural activity profile is similar to the profile of the action potential in the natural neuron. The molecular networks approach can be used in three technologies: neurocomputer, neurochip and molecular chip. This means that molecular networks open new fields of science and engineering called molecular-like machines and molecular machines.     

Nonlinear transient computation as a potential "kernel trick" in cortical processing

Evidence has been found for the presence of chaotic dynamics at all levels of the mammalian brain. This has led to some searching questions about the potential role that nonlinear dynamics may have in neural information processing. We propose that chaos equips the brain with the equivalent of a kernel trick for solving hard nonlinear problems. The approach presented, which is described as nonlinear transient computation, uses the dynamics of a well known chaotic attractor. The paper provides experimental results to show that this approach can be used to solve some challenging pattern recognition tasks. The paper also offers evidence to suggest that the efficacy of nonlinear transient computation for nonlinear pattern classification is dependent only on the generic properties of chaotic attractors and is not sensitive to the particular dynamics of specific sub-regions of chaotic phase space. If, as this work suggests, nonlinear transient computation is independent of the particulars of any given chaotic attractor, then it could be offered as a possible explanation of how the chaotic dynamics that have been observed in brain structures contribute to neural information processing tasks.