43 publications classées par:
type de publication
: Revue avec comité de lecture
Articles Cabessa J. & Villa A. E. P. (2012). The expressive power of analog recurrent neural networks on infinite input streams. Theoretical Computer Science, 436, 23-34. [doi] [web of science] [abstract]
We consider analog recurrent neural networks working on infinite input streams, provide a complete topological characterization of their expressive power, and compare it to the expressive power of classical infinite word reading abstract machines. More precisely, we consider analog recurrent neural networks as language recognizers over the Cantor space, and prove that the classes of image-languages recognized by deterministic and non-deterministic analog networks correspond precisely to the respective classes of image-sets and image-sets of the Cantor space. Furthermore, we show that the result can be generalized to more expressive analog networks equipped with any kind of Borel accepting condition. Therefore, in the deterministic case, the expressive power of analog neural nets turns out to be comparable to the expressive power of any kind of Büchi abstract machine, whereas in the non-deterministic case, analog recurrent networks turn out to be strictly more expressive than any other kind of Büchi or Muller abstract machine, including the main cases of classical automata, image-counter automata, image-counter automata, pushdown automata, and Turing machines.
Cabessa J. & Villa A. E. P. (2010). A hierarchical classification of first-order recurrent neural networks. Lecture Notes in Computer Science, 6031, 142-153. [doi] [web of science] [abstract]
We provide a refined hierarchical classification of first-order recurrent neural networks made up of McCulloch and Pitts cells. The classification is achieved by first proving the equivalence between the expressive powers of such neural networks and Muller automata, and then translating the Wadge classification theory from the automata-theoretic to the neural network context. The obtained hierarchical classification of neural networks consists of a decidable pre-well ordering of width 2 and height omega(omega), and a decidability procedure of this hierarchy is provided. Notably, tins classification is shown to be intimately related to the attractive properties of the networks, and hence provides a new refined measurement of the computational power of these networks in terms of their attractive behaviours.
Gollo L. L., Mirasso C. & Villa A. E. P. (2010). Dynamic control for synchronization of separated cortical areas through thalamic relay. NeuroImage, 52(3), 947-955. [doi] [web of science] [abstract]
Binding of features and information which are processed at different cortical areas is generally supposed to be achieved by synchrony despite the non-negligible delays between these areas. In this work we study the dynamics and synchronization properties of a simplified model of the thalamocortical circuit where different cortical areas are interconnected with a certain delay, that is longer than the internal time scale of the neurons. Using this simple model we find that the thalamus could serve as a central subcortical area that is able to generate zero-lag synchrony between distant cortical areas by means of dynamical relaying (Vicente et al., 2008). Our results show that the model circuit is able to generate fast oscillations in frequency ranges of the beta and gamma bands triggered by an external input to the thalamus formed by independent Poisson trains. We propose a control mechanism to turn "On" and "Off" the synchronization between cortical areas as a function of the relative rate of the external input fed into dorsal and ventral thalamic neuronal populations. The current results emphasize the hypothesis that the thalamus could control the dynamics of the thalamocortical functional networks enabling two separated cortical areas to be either synchronized (at zero-lag) or unsynchronized. This control may happen at a fast time scale, in agreement with experimental data, and without any need of plasticity or adaptation mechanisms which typically require longer time scales.
Iglesias J. & Villa A. E. P. (2010). Recurrent spatiotemporal firing patterns in large spiking neural networks with ontogenetic and epigenetic processes. Journal of Physiology - Paris, 104(3-4), 137-146. [doi] [web of science] [abstract]
Neural development and differentiation are characterized by an overproduction of cells and a transient exuberant number of connections followed by cell death and selective synaptic pruning. We simulated large spiking neural networks (10,000 units at its maximum size) with and without an ontogenetic process corresponding to a brief initial phase of apoptosis driven by an excessive firing rate mimicking cell death due to glutamatergic neurotoxicity and glutamate-triggered apoptosis. This phase was followed by the onset of spike timing dependent synaptic plasticity (STDP), driven by spatiotemporal patterns of stimulation. Despite the reduction in cell counts the apoptosis tended to increase the excitatory/inhibitory ratio because the inhibitory cells were affected at first. Recurrent spatiotemporal firing patterns emerged in both developmental condition but they differed in dynamics. They were less numerous but repeated more often after apoptosis. The results suggest that initial cell death may be necessary for the emergence of stable cell assemblies, able to sustain and process temporal information, from the initially randomly connected networks.
Villa A. E. P. & Tetko I. V. (2010). Cross-frequency coupling in mesiotemporal EEG recordings of epileptic patients. Journal of Physiology - Paris, 104(3-4), 197-202. [doi] [web of science] [abstract]
Semi-invasive foramen ovate (Fov) electrodes were used to record electrical activity in the vicinity of the inferior mesial temporal region of epileptic patients, in addition to standard scalp EEG. Third order cumulant analysis was used to measure the phase-coupled frequencies corresponding to non-linear coupling of spectral frequency components, somewhat analogous to frequencies of resonance. On the basis of the distribution of these frequencies, an index of resonance (IR) is defined as the ratio between the number of peaks in the gamma-band (40-55 Hz) vs. the number of peaks in the beta-band (15-30 Hz). The epileptogenic focus was located in the hemisphere with lower resonant frequencies because these frequencies were characteristic of a spread of the seizure over a broader area. In the case of Fov electrodes IR could differentiate a group of patients affected by a tumor compared to patients with mesial temporal sclerosis. The novel index IR appears as an interesting parameter to evaluate the level of interareal functional connectivity in Fov recordings in epileptic patients, but its usage is likely to be extended in electrophysiological studies.
Perrig S., Dutoit P., Espa-Cervena K., Shaposhnyk V., Pelletier L., Berger F. & Villa A. E. P. (2009). Changes in quadratic phase coupling of EEG signals during wake and sleep in two chronic insomnia patients, before and after cognitive behavioral therapy. Frontiers in Artificial Intelligence and Applications, 204, 217-228. [doi] [web of science] [abstract]
Quantitative EEG studies of primary insomnia (PI) suggest that increased high frequency and reduction in slow frequency EEG activity could be associated with cortical "hyperarousal" and sleep homeostasis dysregulation. This preliminary study is the first to apply higher order EEG analysis in chronic PI patients. We analyzed phase coupling in two patients against two control subjects. We defined an index of resonant frequency (IRF) and show that both patients were characterized by high IRF values that suggest an increase in local cortical information processing before treatment. We show that cognitive behavioral therapy for insomnia (CBT-T) is able to reverse EEG phase coupling towards control values in as little as eight sessions. After treatment the patients were characterized by lower index values, thus suggesting recovery of information processing over wide-spread cortical areas.
Shaposhnyk V., Dutoit P., Contreras-Lámus V., Perrig S. & Villa A. E. P. (2009). A framework for simulation and analysis of dynamically organized distributed neural networks. Lecture Notes in Computer Science, 5768, 277-286. [doi] [web of science] [abstract]
We present a framework for modelling and analyzing emerging neural activity from multiple interconnected modules, where each module is formed by a neural network. The neural network simulator operates a 2D lattice tissue of leaky integrate-and-fire neurons with genetic, ontogenetic and epigenetic features. The;lava. Agent DEvelopment (JADE) environment allows the implementation of an efficient automata-like virtually unbound and platform-independent system of agents exchanging hierarchically organized messages. This framework allowed us to develop linker agents capable to handle dynamic configurations characterized by the entrance and exit of additional modules at any time following simple rewiring rules. The development of a virtual electrode allows the recording of a "neural" generated signal, called electrochipogram (EChG), characterized by dynamics close to biological local field potentials and electroencephalograms (EEG). These signals can be used to compute Evoked Potentials by complex sensory inputs and comparisons with neurophysiological signals of similar kind.
Asai Y., Guha A. & Villa A. E. P. (2008). Deterministic neural dynamics transmitted through neural networks. Neural Networks, 21(6), 799-809. [doi] [web of science] [abstract]
Precise spatiotemporal sequences of neuronal discharges (i.e., intervals between epochs repeating more often than expected by chance), have been observed in a large set of experimental electrophysiological recordings. Sensitivity to temporal information, by itself, does not demonstrate that dynamics embedded in spike trains can be transmitted through a neural network. This study analyzes how synaptic transmission through three archetypical types of neurons (regular-spiking, thalamo-cortical and resonator), simulated by a simple spiking model, can affect the transmission of precise timings generated nonlinear deterministic system (i.e., the Zaslavskii mapping in the present study). The results show cells with subthreshold oscillations (resonators) are very sensitive to stochastic inputs, and are not a good candidate for transmitting temporally coded information. Thalamo-cortical neurons may transmit very well temporal patterns in the absence of background activity, but jitter accumulates along the synaptic chain. Conversely, we observed that cortical regular-spiking neurons can propagate filtered temporal information in a reliable way through the network, and with high temporal accuracy. We discuss the results in the general framework of neural dynamics and brain theories.
Asai Y. & Villa A. E. P. (2008). Effect of the background activity on the reconstruction of spike train by spike pattern detection. Lecture Notes in Computer Science, 5164, 607-616. [doi] [web of science] [abstract]
Deterministic nonlinearity has been observed in experimental electrophysiological recordings performed in several areas of the brain. However, little is known about the ability to transmit a complex temporally organized activity through different types of spiking neurons. This study investigates the response of a spiking neuron model representing five archetypical types to input spike trains including deterministic information generated by a chaotic attractor. The comparison between input. and output spike trains is carried out by the pattern grouping algorithm (PGA) as a function of the intensity of the background activity for each neuronal type. The results show that the thalamo-cortical, regular spiking and intrinsically busting model neurons can be good candidate in transmitting temporal information with different characteristics in a spatially organized neural network.
Asai Y. & Villa A. E. P. (2008). Reconstruction of underlying nonlinear deterministic dynamics embedded in noisy spike trains. Journal of Biological Physics, 34(3-4), 325-340. [doi] [web of science] [abstract]
An experimentally recorded time series formed by the exact times of occurrence of the neuronal spikes (spike train) is likely to be affected by observational noise that provokes events mistakenly confused with neuronal discharges, as well as missed detection of genuine neuronal discharges. The points of the spike train may also suffer a slight jitter in time due to stochastic processes in synaptic transmission and to delays in the detecting devices. This study presents a procedure aimed at filtering the embedded noise (denoising the spike trains) the spike trains based on the hypothesis that recurrent temporal patterns of spikes are likely to represent the robust expression of a dynamic process associated with the information carried by the spike train. The rationale of this approach is tested on simulated spike trains generated by several nonlinear deterministic dynamical systems with embedded observational noise. The application of the pattern grouping algorithm (PGA) to the noisy time series allows us to extract a set of points that form the reconstructed time series. Three new indices are defined for assessment of the performance of the denoising procedure. The results show that this procedure may indeed retrieve the most relevant temporal features of the original dynamics. Moreover, we observe that additional spurious events affect the performance to a larger extent than the missing of original points. Thus, a strict criterion for the detection of spikes under experimental conditions, thus reducing the number of spurious spikes, may raise the possibility to apply PGA to detect endogenous deterministic dynamics in the spike train otherwise masked by the observational noise.
Chibirova O., Iglesias J., Shaposhnyk V. & Villa A. E. P. (2008). Dynamics of firing patterns in evolvable hierarchically organized neural networks. Lecture Notes in Computer Science, 5216, 296-307. [doi] [web of science] [abstract]
A scalable hardware platform made of custom reconfigurable devices endowed with bio-inspired ontogenetic and epigenetic features is configured to run an artificial neural network with developmental and evolvable capabilities. The hardware architect tire allows internet-work communication and this study analyzes the simulated activity of two hierarchically organized spiking neural networks. The main features were, an initial developmental phase characterized by cell death (apoptosis driven by excessive firing, rate), followed by spike timing dependent synaptic plasticity in presence of background noise. The emergence of precise firing sequences formed by recurrent patterns of spike intervals above chance levels suggested the build-up of a connectivity, out of initially randomly connected networks, able to sustain temporal information processing. The relative frequency of precise firing sequences was higher in the, downstream network and their dynamics suggested the emergence of an unsupervised hierarchical activity-driven connectivity.
Iglesias J., Garcá-Ojalvo J. & Villa A. E. P. (2008). Effect of feedback strength in coupled spiking neural networks. Lecture Notes in Computer Science, 5164, 646-654. [doi] [web of science] [abstract]
We simulated the coupling of two large spiking neural networks works (10(4) units each) composed by 80% of excitatory units and 20% of inhibitory units, randomly connected by projections featuring spike-timing dependent plasticity, locality preference and synaptic pruning. Only the first network received a complex spatiotemporal stimulus and projected on the second network, in a setup akin to coupled semiconductor lasers. In a series of simulations, the strength of the feedback from the second network to the first was modified to evaluate the effect of the bidirectional coupling on the firing dynamics of the two networks. We observed that, unexpectedly. the number of neurons which activity is altered by the introduction of feedback increases in the second network more than in the first network, suggesting a qualitative change in the dynamics of the first network when feedback is increased.
Iglesias J. & Villa A. E. P. (2008). Emergence of preferred firing sequences in large spiking neural networks during simulated neuronal development. International Journal of Neural Systems, 18(4), 267-277. [doi] [web of science] [abstract]
Two main processes concurrently re. ne the nervous system over the course of development: cell death and selective synaptic pruning. We simulated large spiking neural networks (100 x 100 neurons "at birth") characterized by an early developmental phase with cell death due to excessive. ring rate, followed by the onset of spike timing dependent synaptic plasticity (STDP), driven by spatiotemporal patterns of stimulation. The cell death affected the inhibitory units more than the excitatory units during the early developmental phase. The network activity showed the appearance of recurrent spatiotemporal. ring patterns along the STDP phase, thus suggesting the emergence of cell assemblies from the initially randomly connected networks. Some of these patterns were detected throughout the simulation despite the activity-driven network modifications while others disappeared.
Aksenova T. I., Volkovych V. V. & Villa A. E. (2007). Detection of spectral instability in EEG recordings during the preictal period. Journal of Neural Engineering, 4(3), 173-178. [doi] [web of science] [abstract]
The study of EEG recordings during the interval prior to an epileptic seizure onset-the preictal period-is likely to detect changes in the ongoing brain activity consistent with seizure anticipation. A novel index of spectral instability (ISpI) based on multiple abrupt changes of EEG spectral features is presented here. Based on the analysis of control records, robust M-estimates are used to calculate the threshold and avoid false warnings. The results obtained with a small data set (three patients, ten preictal records per patient) have shown that the ISpI index provided a warning flag that anticipated the seizure onset by 13.1 (SD = 4.0) min on average.
Asai Y., Yokoi T. & Villa A. E. P. (2007). Deterministic nonlinear spike train filtered by spiking neuron model. Lecture Notes in Computer Science, 4668, 924-933. [doi] [web of science] [abstract]
Deterministic nonlinear dynamics has been observed in experimental electrophysiological recordings performed in several areas of the brain. However, little is known about the ability to transmit a complex temporally organized activity through different types of spiking neurons. This study investigates the response of a spiking neuron model representing three archetypical types (regular spiking, thalamocortical and resonator) to input spike trains composed of deterministic (chaotic) and stochastic processes with weak background activity. The comparison of the input and output spike trains allows to assess the transmission of information contained in the deterministic nonlinear dynamics. The pattern grouping algorithm (PGA) was applied to the output of the neuron to detect the dynamical attractor embedded in the original input spike train. The results show that the model of the thalamo-cortical neuron can be a better candidate than regular spiking and resonator type neurons in transmitting temporal information in a spatially organized neural network.
Farré-Castany M. A., Schwaller B., Gregory P., Barski J., Mariethoz C., Eriksson J. L. et al. (2007). Differences in locomotor behavior revealed in mice deficient for the calcium-binding proteins parvalbumin, calbindin D-28k or both. Behavioural Brain Research, 178(2), 250-261. [doi] [web of science] [abstract]
We investigated the role of the two calcium-binding proteins parvalbumin (PV) and calbindin D-28k (CB) in the locomotor activity and motorcoordination using null-mutant mice for PV (PV-/-), CB (CB-/-) or both proteins (PV-/-CB-/-). These proteins are expressed in distinct, mainly non-overlapping populations of neurons of the central and peripheral nervous system and PV additionally in fast-twitch muscles. In a test measuring repeated locomotor activity during 18-20 days, the analysis revealed a slightly increased activity in truce lacking either protein, while the lack of both decreased the number of beams crossed during active periods. An increase in the characteristic speed during the first 8 days could be attributed to PV-deficiency, while the elimination of CB in CB-/- and double-KO mice decreased the percentage of fast movements at all time points. In the latter, additionally a reduction of the fastest speed was observed. The alterations in locomotor activity (fast movements, fastest speed) strongly correlate with the impairment in locomotor coordination in mice deficient for CB evidenced in the runway assay and the rotarod assay. The graded locomotor phenotype (CB > PV) is qualitatively correlated with alterations in Purkinje cell firing reported previously in these mice. The presence or absence of either protein did not affect the spontaneous locomotor activity when animals were placed in a novel environment and tested only once for 30 min. In summary, the lack of these calcium-binding proteins yields characteristic, yet distinct phenotypes with respect to locomotor activity and coordination.
Iglesias J., Chibirova O. K. & Villa A. E. P. (2007). Nonlinear dynamics emerging in large scale neural networks with ontogenetic and epigenetic processes. Lecture Notes in Computer Science, 4668, 579-588. [doi] [web of science] [abstract]
We simulated a large scale spiking neural network characterized by an initial developmental phase featuring cell death driven by an excessive firing rate, followed by the onset of spike-timing-dependent synaptic plasticity (STDP), driven by spatiotemporal patterns of stimulation. The network activity stabilized such that recurrent preferred firing sequences appeared along the STDP phase. The analysis of the statistical properties of these patterns give hints to the hypothesis that a neural network may be characterized by a particular state of an underlying dynamical system that produces recurrent firing patterns.
Iglesias J. & Villa A. E. P. (2007). Effect of stimulus-driven pruning on the detection of spatiotemporal patterns of activity in large neural networks. BioSystems, 89, 287-293. [doi] [web of science] [abstract]
Adult patterns of neuronal connectivity develop from a transient embryonic template characterized by exuberant projections to both appropriate and inappropriate target regions in a process known as synaptic pruning. Trigger signals able to induce synaptic pruning could be related to dynamic functions that depend on the timing of action potentials. We stimulated locally connected random networks of spiking neurons and observed the effect of a spike-timing-dependent synaptic plasticity (STDP)-driven pruning process on the emergence of cell assemblies. The spike trains of the simulated excitatory neurons were recorded. We searched for spatiotemporal firing patterns as potential markers of the build-up of functionally organized recurrent activity associated with spatially organized connectivity.
Sirovich R., Sacerdote L. & Villa A. E. P. (2007). Effect of increasing inhibitory inputs on information processing within a small network of spiking neurons. Lecture Notes in Computer Science, 4507, 23-30. [doi] [web of science] [abstract]
In this paper the activity of a spiking neuron A that receives a background input from the network in which it is embedded and strong inputs from an excitatory unit E and an inhibitory unit I is studied. The membrane potential of the neuron A is described by a jump diffusion model. Several types of interspike interval distributions of the excitatory strong inputs are considered as Poissonian inhibitory inputs increase intensity. It is shown that, independently of the distribution of the excitatory inpu, they are more efficiently transmitted as inhibition increases to larger intensities.
Turova T. S. & Villa A. E. P. (2007). On a phase diagram for random neural networks with embedded spike timing dependent plasticity. BioSystems, 89, 280-286. [doi] [web of science] [abstract]
This paper presents an original mathematical framework based on graph theory which is a first attempt to investigate the dynamics of a model of neural networks with embedded spike timing dependent plasticity. The neurons correspond to integrate-and-fire units located at the vertices of a finite subset of 2D lattice. There are two types of vertices, corresponding to the inhibitory and the excitatory neurons. The edges are directed and labelled by the discrete values of the synaptic strength. We assume that there is an initial firing pattern corresponding to a subset of units that generate a spike. The number of activated externally vertices is a small fraction of the entire network. The model presented here describes how such pattern propagates throughout the network as a random walk on graph. Several results are compared with computational simulations and new data are presented for identifying critical parameters of the model.
Villa A. E. P., Asai Y. & Segundo J. P. (2007). Influence of the temporal distribution of electric pulses on transcallosal single unit responses. BioSystems, 89, 143-153. [doi] [web of science] [abstract]
We examined how differently timed stimuli to one auditory cortex affect the spike trains they drive in the controlateral homotopic field of anesthetized rats. Bipolar electrical stimulations consisted of trains of pulses (100 mu s, < 500 mu A) at rates of 25, 50 or 125 pulses/s and with different stimulus patterns (i.e., dispersions, sequences), called "pacemaker", "accelerando" or "decelerando". Trains lasted for 342 ms and were separated by 4 s. When trains were evaluated over times comparable to the stimulus duration changes characteristically involved an initial slowing followed by recovery and several discharges both stimulus- and neuron-dependent. When evaluated by cross-correlations between cortical cell pairs, the changes extended far beyond the stimulus end. Results suggest that interhemispheric projections, by way of their averages and patterns, play key, long duration roles in the spike-dependent properties of cortical synapses (e.g., potentiation, depression) and thus of cortical circuit operations.
Villa A. E. P. & Iglesias J. (2007). OpenAdap.net: Evolvable information processing environment. Lecture Notes in Artificial Intelligence, 4578, 227-236. [doi] [web of science] [abstract]
OpenAdap.net is an Open Source project aimed at breaking the barriers existing in the flow of information access and information processing. The infrastructure makes it possible the dissemination of resources like knowledge, tools or data, their exposure to evaluation in ways that might be unanticipated and hence support the evolution of communities of users around a specific domain. The architecture is designed by analogy with a virtual distributed operating system in which the dynamic resources are presented as files in a structured virtual file system featuring ownership and access permissions.
Asai Y., Yokoi T. & Villa A. E. P. (2006). Detection of a dynamical system attractor from spike train analysis. Lecture Notes in Computer Science, 4131, 623-631. [doi] [web of science] [abstract]
Dynamics of the activity of neuronal networks have been intensively studied from the view point of the nonlinear dynamical system. The neuronal activities are recorded as multivariate time series of the epochs of spike occurrences-the spike trains-which are often effected by intrinsic and measuring noise. The spike trains can be considered as a mixture of a realization of deterministic and stochastic processes. Within this framework we considered several simulated spike trains derived from the Zaslavskii map with additive noise. The ensemble of all preferred firing sequences detected by the pattern grouping algorithm (PGA) in the noisy spike trains form a new time series that retains the dynamics of the original mapping.
Eriksson J. L. & Villa A. E. P. (2006). Learning of auditory equivalence classes for vowels by rats. Behavioural Processes, 73(3), 348-359. [doi] [web of science] [abstract]
Four male Long-Evans rats were trained to discriminate between synthetic vowel sounds using a GO/NOGO response choice task. The vowels were characterized by an increase in fundamental frequency correlated with an upward shift in formant frequencies. In an initial phase we trained the subjects to discriminate between two vowel categories using two exemplars from each category. In a subsequent phase the ability of the rats to generalize the discrimination between the two categories was tested. To test whether rats might exploit the fact that attributes of training stimuli covaried, we used non-standard stimuli with a reversed relation between fundamental frequency and formants. The overall results demonstrate that rats are able to generalize the discrimination to new instances of the same vowels. We present evidence that the performance of the subjects depended on the relation between fundamental and formant frequencies that they had previously been exposed to. Simple simulation results with artificial neural networks could reproduce most of the behavioral results and support the hypothesis that equivalence classes for vowels are associated with an experience-driven process based on general properties of peripheral auditory coding mixed with elementary learning mechanisms. These results suggest that rats use spectral and temporal cues similarly to humans despite differences in basic auditory capabilities.
Iglesias J. & Villa A. E. P. (2006). Neuronal cell death and synaptic pruning driven by spike-timing dependent plasticity. Lecture Notes in Computer Science, 4131, 953-962. [doi] [web of science] [abstract]
The embryonic nervous system is refined over the course of development as a result of two main processes: apoptosis (programmed cell death) and selective axon pruning. We simulated a large scale spiking neural network characterized by an initial apoptotic phase, driven by an excessive firing rate, followed by the onset of spike-timing-dependent plastiticity (STDP), driven by spatiotemporal patterns of stimulation. In the apoptotic phase the cell death affected the inhibitory more than the excitatory units. The network activity stabilized such that recurrent preferred firing sequences appeared along the STDP phase, thus suggesting the emergence of cell assemblies from large randomly connected networks.
Moreno J. M., Iglesias J., Eriksson J. L. & Villa A. E. P. (2006). Physical mapping of spiking neural networks models on a bio-inspired scalable architecture. Lecture Notes in Computer Science, 4131, 936-943. [doi] [web of science] [abstract]
The paper deals with the physical implementation of biologically plausible spiking neural network models onto a hardware architecture with bio-inspired capabilities. After presenting the model, the work will illustrate the major steps taken in order to provide a compact and efficient digital hardware implementation of the model. Special emphasis will be given to the scalability features of the architecture, that will permit the implementation of large-scale networks. The paper will conclude with details about the physical mapping of the model, as well as with experimental results obtained when applying dynamic input stimuli to the implemented network.
Moreno J. M., Thoma Y., Sanchez E., Eriksson J., Iglesias J. & Villa A. (2006). The POEtic electronic tissue and its role in the emulation of large-scale biologically inspired spiking neural networks models. Complexus, 3, 32-47. [doi] [abstract]
One of the major obstacles found when trying to construct artefacts derived from principles observed in living beings is the lack of actual dynamic hardware with autonomous capabilities. Even if programmable devices offer the possibility of modifying the functionality implemented in the device, they rely on external hardware and software elements to provide its physical configuration. In this paper we present a new family of electronic devices, called POEtic, whose architecture has been derived from the basic properties that can be extracted from the three major organization principles present in living beings: phylogenesis, ontogenesis and epigenesis. We will demonstrate that the capabilities present in these new programmable devices make them an ideal candidate for the real-time emulation of large-scale biologically inspired spiking neural network models.
Sacerdote L., Villa A. E. P. & Zucca C. (2006). On the classification of experimental data modeled via a stochastic leaky integrate and fire model through boundary values. Bulletin of Mathematical Biology, 68(6), 1257-1274. [doi] [abstract]
We present a computational algorithm aimed to classify single unit spike trains on the basis of observed interspikes intervals (ISI). The neuronal activity is modeled with a stochastic leaky integrate and fire model and the inverse first passage time method is extended to the Ornstein-Uhlenbeck (ISI) process. Differences between spike trains are detected in terms of the boundary shape. The proposed classification method is applied to the analysis of multiple single units recorded simultaneously in the thalamus and in the cerebral cortex of unanesthetized rats during spontaneous activity. We show the existence of at least three different firing patterns that could not be classified using the usual statistical indices.
Iglesias J., Eriksson J., Grize F., Tomassini M. & Villa A. (2005). Dynamics of pruning in simulated large-scale spiking neural networks. Biosystems, 79, 11 - 20.
Actes de conférence (partie) A. E. P. Villa, J. Iglesias & S. Ghernaouti-Hélie (2010). OpenAdap.net: a Community-Based Sharing System. In Gerald Eichler, Peter G. Kropf, Ulrike Lechner, Phayung Meesad & Herwig Unger (Eds.), Lecture Notes in Informatics, 10th International Conference on Innovative Internet Community Services (I2CS), Jubilee Edition 2010, June 3-5, 2010, Bangkok, Thailand, 165 (pp. 321-328). GI. [url]
Brousse O., Guillot J., Gil T., Grize F., Sassatelli G., Moreno J. M. et al. (2009). JubiTool: Unified design flow for the Perplexus SIMD hardware accelerator. CEC '09. IEEE Congress on Evolutionary Computation, 2009 (pp. 2070-2075). [doi]
Sanchez E., Perez-Uribe A., Upegui A., Thoma Y., Moreno J. M., Villa A. et al. (2007). PERPLEXUS: Pervasive computing framework for modeling complex virtually-unbounded systems. Proceedings of the Second NASA/ESA Conference on Adaptive Hardware and Systems (pp. 587-591). [doi]
Villa A., Iglesias J. & Ghernaouti-Hélie S. (2007, Sep). OpenAdap.net - A socialware for knowledge sharing. Proceedings of International Conference on Internet Technologies and Applications (ITA), Wrexham, North Wales, UK. [url]
Villa A., Iglesias J. & Ghernaouti-Hélie S. (2007, Oct). Knowledge Sharing by means of OpenAdap.net. Proceedings of 2nd Workshop IST_Africa - Supporting Research Engagement with Africa.
Villa A., Iglesias J. & Ghernaouti-Hélie S. (2007, Juin). OpenAdapt.net: a dynamic middleware for knowledge production and distribution. Proceedings of International conference - Towards a Knowledge Society: Is Knowledge a Public Good? Dynamics of Knowledge Production and Distribution (ESSHRA). Bern. [url]
Eriksson J. L. & Villa A. E. P. (2006). Artificial Neural Networks simulation of learning of auditory equivalence classes for vowels. Proceedings of International Joint Conference on Neural Networks, 2006. (pp. 526-533). [doi]
Villa A.E.P., Iglesias J. & Ghernaouti-Hélie S. (2006, Juin). OpenAdap.Net: a Community-Based Shared System. Proceedings of 6th International Workshop Innovative Internet Community Systems. Neuchâtel, Switzerland.
Villa A.E.P., Iglesias J. & Ghernaouti-Hélie S. (2006, Oct). OpenAdap.net: a collaborative sharing environment. Proceedings of E-Challenges 2006. Barcelona, Spain. [url]
Villa A.E.P., Iglesias J. & Ghernaouti-Hélie S. (2006, Mai). OpenAdap.Net: a technical perspective. Proceedings Workshop IST_Africa - Supporting Research Engagement with Africa (IST-Africa 2006), Pretoria, South Africa.
Iglesias J., Eriksson J., Pardo B., Tomassini M. & Villa A. (2005). Stimulus-driven unsupervised synaptic pruning in large neural networks. Lecture Notes in Computer Science, Proceedings of BV & Ai, 3704 (pp. 59 - 68). Springer, Berlin.
Iglesias J., Eriksson J., Pardo B., Tomassini M. & Villa A. (2005). Emergence of oriented cell assemblies associated with spike-timing-dependent plasticity. Lecture Notes in Computer Science, Proceedings of ICANNN 2005, 3696 (pp. 127 - 132). Springer, Berlin.
Thèses Mesrobian S. K., Villa A. E.P. (Dir.) (2015). Does working memory training affect decision making ? : a neuroeconomic study. Université de Lausanne, Faculté de biologie et médecine. [pdf] [abstract]
We all make decisions of varying levels of importance every day. Because making a decision implies that there are alternative choices to be considered, almost all decision involves some conflicts or dissatisfaction. Traditional economic models esteem that a person must weight the positive and negative outcomes of each option, and based on all these inferences, determines which option is the best for that particular situation. However, individuals rather act as irrational agents and tend to deviate from these rational choices. They somewhat evaluate the outcomes' subjective value, namely, when they face a risky choice leading to losses, people are inclined to have some preference for risk over certainty, while when facing a risky choice leading to gains, people often avoid to take risks and choose the most certain option. Yet, it is assumed that decision making is balanced between deliberative and emotional components. Distinct neural regions underpin these factors: the deliberative pathway that corresponds to executive functions, implies the activation of the prefrontal cortex, while the emotional pathway tends to activate the limbic system. These circuits appear to be altered in individuals with ADHD, and result, amongst others, in impaired decision making capacities. Their impulsive and inattentive behaviors are likely to be the cause of their irrational attitude towards risk taking. Still, a possible solution is to administrate these individuals a drug treatment, with the knowledge that it might have several side effects. However, an alternative treatment that relies on cognitive rehabilitation might be appropriate.¦This project was therefore aimed at investigate whether an intensive working memory training could have a spillover effect on decision making in adults with ADHD and in age-matched healthy controls. We designed a decision making task where the participants had to select an amount to gamble with the chance of 1/3 to win four times the chosen amount, while in the other cases they could loose their investment. Their performances were recorded using electroencephalography prior and after a one-month Dual N-Back training and the possible near and far transfer effects were investigated.¦Overall, we found that the performance during the gambling task was modulated by personality factors and by the importance of the symptoms at the pretest session. At posttest, we found that all individuals demonstrated an improvement on the Dual N-Back and on similar untrained dimensions. In addition, we discovered that not only the adults with ADHD showed a stable decrease of the symptomatology, as evaluated by the CAARS inventory, but this reduction was also detected in the control samples. In addition, Event-Related Potential (ERP) data are in favor of an change within prefrontal and parietal cortices.¦These results suggest that cognitive remediation can be effective in adults with ADHD, and in healthy controls. An important complement of this work would be the examination of the data in regard to the attentional networks, which could empower the fact that complex programs covering the remediation of several executive functions' dimensions is not required, a unique working memory training can be sufficient.¦--¦Nous prenons tous chaque jour des décisions ayant des niveaux d'importance variables. Toutes les décisions ont une composante conflictuelle et d'insatisfaction, car prendre une décision implique qu'il y ait des choix alternatifs à considérer. Les modèles économiques traditionnels estiment qu'une personne doit peser les conséquences positives et négatives de chaque option et en se basant sur ces inférences, détermine quelle option est la meilleure dans une situation particulière. Cependant, les individus peuvent dévier de ces choix rationnels. Ils évaluent plutôt les valeur subjective des résultats, c'est-à-dire que lorsqu'ils sont face à un choix risqué pouvant les mener à des pertes, les gens ont tendance à avoir des préférences pour le risque à la place de la certitude, tandis que lorsqu'ils sont face à un choix risqué pouvant les conduire à un gain, ils¦évitent de prendre des risques et choisissent l'option la plus su^re. De nos jours, il est considéré que la prise de décision est balancée entre des composantes délibératives et émotionnelles. Ces facteurs sont sous-tendus par des régions neurales distinctes: le chemin délibératif, correspondant aux fonctions exécutives, implique l'activation du cortex préfrontal, tandis que le chemin émotionnel active le système limbique. Ces circuits semblent être dysfonctionnels chez les individus ayant un TDAH, et résulte, entre autres, en des capacités de prise de décision altérées. Leurs comportements impulsifs et inattentifs sont probablement la cause de ces attitudes irrationnelles face au risque. Cependant, une solution possible est de leur administrer un traitement médicamenteux, en prenant en compte les potentiels effets secondaires. Un traitement alternatif se reposant sur une réhabilitation cognitive pourrait être appropriée.¦Le but de ce projet est donc de déterminer si un entrainement intensif de la mémoire de travail peut avoir un effet sur la prise de décision chez des adultes ayant un TDAH et chez des contrôles sains du même âge. Nous avons conçu une tâche de prise de décision dans laquelle les participants devaient sélectionner un montant à jouer en ayant une chance sur trois de gagner quatre fois le montant choisi, alors que dans l'autre cas, ils pouvaient perdre leur investissement. Leurs performances ont été enregistrées en utilisant l'électroencéphalographie avant et après un entrainement d'un mois au Dual N-Back, et nous avons étudié les possibles effets de transfert.¦Dans l'ensemble, nous avons trouvé au pré-test que les performances au cours du jeu d'argent étaient modulées par les facteurs de personnalité, et par le degré des sympt^omes. Au post-test, nous avons non seulement trouvé que les adultes ayant un TDAH montraient une diminutions stable des symptômes, qui étaient évalués par le questionnaire du CAARS, mais que cette réduction était également perçue dans l'échantillon des contrôles. Les rsultats expérimentaux mesurés à l'aide de l'éléctroencéphalographie suggèrent un changement dans les cortex préfrontaux et pariétaux.¦Ces résultats suggèrent que la remédiation cognitive est efficace chez les adultes ayant un TDAH, mais produit aussi un effet chez les contrôles sains. Un complément important de ce travail pourrait examiner les données sur l'attention, qui pourraient renforcer l'idée qu'il n'est pas nécessaire d'utiliser des programmes complexes englobant la remédiation de plusieurs dimensions des fonctions exécutives, un simple entraiment de la mémoire de travail devrait suffire.
Iglesias J., Tomassini M. (Dir.) (2005). Emergence of oriented circuits driven by synaptic pruning associated with spike-timing-dependent plasticity (STDP). Université de Lausanne, Faculté des sciences. [pdf] [abstract]
ABSTRACT:¦Massive synaptic pruning following over-growth is a general feature of mammalian brain maturation. Pruning starts near time of birth and is completed by time of sexual maturation. Trigger signals able to induce synaptic pruning could be related to dynamic functions that depend on the timing of action potentials. Spike-timing-dependent synaptic plasticity (STDP) is a change in the synaptic strength based on the ordering of pre- and postsynaptic spikes. The relation between synaptic efficacy and synaptic pruning suggests that the weak synapses may be modified and removed through competitive "learning" rules. This plasticity rule might produce the strengthening of the connections among neurons that belong to cell assemblies characterized by recurrent patterns of firing. Conversely, the connections that are not recurrently activated might decrease in efficiency and eventually be eliminated.¦The main goal of our study is to determine whether or not, and under which conditions, such cell assemblies may emerge out of a locally connected random network of integrate-and-fire units distributed on a 2D lattice receiving background noise and content-related input organized in both temporal and spatial dimensions. The originality of our study stands on the relatively large size of the network, 10,000 units, the duration of the experiment, 10E6 time units (one time unit corresponding to the duration of a spike), and the application of an original bio-inspired STDP modification rule compatible with hardware implementation.¦A first batch of experiments was performed to test that the randomly generated connectivity and the STDP-driven pruning did not show any spurious bias in absence of stimulation. Among other things, a scale factor was approximated to compensate for the network size on the ac¬tivity. Networks were then stimulated with the spatiotemporal patterns. The analysis of the connections remaining at the end of the simulations, as well as the analysis of the time series resulting from the interconnected units activity, suggest that feed-forward circuits emerge from the initially randomly connected networks by pruning.¦RESUME:¦L'élagage massif des synapses après une croissance excessive est une phase normale de la ma¬turation du cerveau des mammifères. L'élagage commence peu avant la naissance et est complété avant l'âge de la maturité sexuelle. Les facteurs déclenchants capables d'induire l'élagage des synapses pourraient être liés à des processus dynamiques qui dépendent de la temporalité rela¬tive des potentiels d'actions. La plasticité synaptique à modulation temporelle relative (STDP) correspond à un changement de la force synaptique basé sur l'ordre des décharges pré- et post- synaptiques. La relation entre l'efficacité synaptique et l'élagage des synapses suggère que les synapses les plus faibles pourraient être modifiées et retirées au moyen d'une règle "d'appren¬tissage" faisant intervenir une compétition. Cette règle de plasticité pourrait produire le ren¬forcement des connexions parmi les neurones qui appartiennent à une assemblée de cellules caractérisée par des motifs de décharge récurrents. A l'inverse, les connexions qui ne sont pas activées de façon récurrente pourraient voir leur efficacité diminuée et être finalement éliminées.¦Le but principal de notre travail est de déterminer s'il serait possible, et dans quelles conditions, que de telles assemblées de cellules émergent d'un réseau d'unités integrate-and¬-fire connectées aléatoirement et distribuées à la surface d'une grille bidimensionnelle recevant à la fois du bruit et des entrées organisées dans les dimensions temporelle et spatiale. L'originalité de notre étude tient dans la taille relativement grande du réseau, 10'000 unités, dans la durée des simulations, 1 million d'unités de temps (une unité de temps correspondant à une milliseconde), et dans l'utilisation d'une règle STDP originale compatible avec une implémentation matérielle.¦Une première série d'expériences a été effectuée pour tester que la connectivité produite aléatoirement et que l'élagage dirigé par STDP ne produisaient pas de biais en absence de stimu¬lation extérieure. Entre autres choses, un facteur d'échelle a pu être approximé pour compenser l'effet de la variation de la taille du réseau sur son activité. Les réseaux ont ensuite été stimulés avec des motifs spatiotemporels. L'analyse des connexions se maintenant à la fin des simulations, ainsi que l'analyse des séries temporelles résultantes de l'activité des neurones, suggèrent que des circuits feed-forward émergent par l'élagage des réseaux initialement connectés au hasard.