Model-Free Reconstruction of Excitatory Neuronal Connectivity from Calcium Imaging Signals

A systematic assessment of global neural network connectivity through direct electrophysiological assays has remained technically infeasible, even in simpler systems like dissociated neuronal cultures. We introduce an improved algorithmic approach based on Transfer Entropy to reconstruct structural connectivity from network activity monitored through calcium imaging. We focus in this study on the inference of excitatory synaptic links. Based on information theory, our method requires no prior assumptions on the statistics of neuronal firing and neuronal connections. The performance of our algorithm is benchmarked on surrogate time series of calcium fluorescence generated by the simulated dynamics of a network with known ground-truth topology. We find that the functional network topology revealed by Transfer Entropy depends qualitatively on the time-dependent dynamic state of the network (bursting or non-bursting). Thus by conditioning with respect to the global mean activity, we improve the performance of our method. This allows us to focus the analysis to specific dynamical regimes of the network in which the inferred functional connectivity is shaped by monosynaptic excitatory connections, rather than by collective synchrony. Our method can discriminate between actual causal influences between neurons and spurious non-causal correlations due to light scattering artifacts, which inherently affect the quality of fluorescence imaging. Compared to other reconstruction strategies such as cross-correlation or Granger Causality methods, our method based on improved Transfer Entropy is remarkably more accurate. In particular, it provides a good estimation of the excitatory network clustering coefficient, allowing for discrimination between weakly and strongly clustered topologies. Finally, we demonstrate the applicability of our method to analyses of real recordings of in vitro disinhibited cortical cultures where we suggest that excitatory connections are characterized by an elevated level of clustering compared to a random graph (although not extreme) and can be markedly non-local.

Transfer entropy reconstruction and labeling of neuronal connections from simulated calcium imaging

Neuronal dynamics are fundamentally constrained by the underlying structural network architecture, yet much of the details of this synaptic connectivity are still unknown even in neuronal cultures in vitro. Here we extend a previous approach based on information theory, the Generalized Transfer Entropy, to the reconstruction of connectivity of simulated neuronal networks of both excitatory and inhibitory neurons. We show that, due to the model-free nature of the developed measure, both kinds of connections can be reliably inferred if the average firing rate between synchronous burst events exceeds a small minimum frequency. Furthermore, we suggest, based on systematic simulations, that even lower spontaneous inter-burst rates could be raised to meet the requirements of our reconstruction algorithm by applying a weak spatially homogeneous stimulation to the entire network. By combining multiple recordings of the same in silico network before and after pharmacologically blocking inhibitory synaptic transmission, we show then how it becomes possible to infer with high confidence the excitatory or inhibitory nature of each individual neuron.

Lack of GDAP1 induces neuronal calcium and mitochondrial defects in a knockout mouse model of charcot-marie-tooth neuropathy.

Mutations in GDAP1, which encodes protein located in the mitochondrial outer membrane, cause axonal recessive (AR-CMT2), axonal dominant (CMT2K) and demyelinating recessive (CMT4A) forms of Charcot-Marie-Tooth (CMT) neuropathy. Loss of function recessive mutations in GDAP1 are associated with decreased mitochondrial fission activity, while dominant mutations result in impairment of mitochondrial fusion with increased production of reactive oxygen species and susceptibility to apoptotic stimuli. GDAP1 silencing in vitro reduces Ca2+ inflow through store-operated Ca2+ entry (SOCE) upon mobilization of endoplasmic reticulum (ER) Ca2+, likely in association with an abnormal distribution of the mitochondrial network. To investigate the functional consequences of lack of GDAP1 in vivo, we generated a Gdap1 knockout mouse. The affected animals presented abnormal motor behavior starting at the age of 3 months. Electrophysiological and biochemical studies confirmed the axonal nature of the neuropathy whereas histopathological studies over time showed progressive loss of motor neurons (MNs) in the anterior horn of the spinal cord and defects in neuromuscular junctions. Analyses of cultured embryonic MNs and adult dorsal root ganglia neurons from affected animals demonstrated large and defective mitochondria, changes in the ER cisternae, reduced acetylation of cytoskeletal α-tubulin and increased autophagy vesicles. Importantly, MNs showed reduced cytosolic calcium and SOCE response. The development and characterization of the GDAP1 neuropathy mice model thus revealed that some of the pathophysiological changes present in axonal recessive form of the GDAP1-related CMT might be the consequence of changes in the mitochondrial network biology and mitochondria-endoplasmic reticulum interaction leading to abnormalities in calcium homeostasis.

Synaptic depression and slow oscillatory activity in a biophysical network model of the cerebral cortex

Short-term synaptic depression (STD) is a form of synaptic plasticity that has a large impact on network computations. Experimental results suggest that STD is modulated by cortical activity, decreasing with activity in the network and increasing during silent states. Here, we explored different activity-modulation protocols in a biophysical network model for which the model displayed less STD when the network was active than when it was silent, in agreement with experimental results. Furthermore, we studied how trains of synaptic potentials had lesser decay during periods of activity (UP states) than during silent periods (DOWN states), providing new experimental predictions. We next tackled the inverse question of what is the impact of modifying STD parameters on the emergent activity of the network, a question difficult to answer experimentally. We found that synaptic depression of cortical connections had a critical role to determine the regime of rhythmic cortical activity. While low STD resulted in an emergent rhythmic activity with short UP states and long DOWN states, increasing STD resulted in longer and more frequent UP states interleaved with short silent periods. A still higher synaptic depression set the network into a non-oscillatory firing regime where DOWN states no longer occurred. The speed of propagation of UP states along the network was not found to be modulated by STD during the oscillatory regime; it remained relatively stable over a range of values of STD. Overall, we found that the mutual interactions between synaptic depression and ongoing network activity are critical to determine the mechanisms that modulate cortical emergent patterns.

Expérience comportementale et modélisation par réseau neuronal des différences entre les processus de catégorisation par règles logiques et par ressemblance familiale

Notre étude est bipartite. En premier lieu nous avons effectué une étude empirique des différences entre les processus de catégorisation explicite (verbalisable) et implicite (non-verbalisable). Nous avons examiné la difficulté et le temps nécessaire pour apprendre trois tâches de catégorisation dites par air de famille, par règle logique conjonctive et par règle logique disjonctive. Nous avons ensuite utilisé un réseau neuronal pour modéliser la catégorisation en lui faisant compléter les mêmes tâches. La comparaison entre les deux nous permet de juger de l’adéquation du modèle. Les données empiriques ont montré un effet de la typicité et de la familiarité en accord avec la documentation et nous trouvons que la tâche de catégorisation par règle disjonctive est la plus difficile alors que la tâche de catégorisation par air de famille est la plus facile. La modélisation par le réseau est une réussite partielle mais nous présentons des solutions afin qu’un réseau futur puisse modéliser le processus catégoriel humain efficacement

Fonction de la protéine Ceroid lipofuscinosis neuronal 5 (CLN5) dans le tri et le recyclage à l’endosome

Le tri et le transport efficace des hydrolases acides vers le lysosome jouent un rôle critique pour la fonction des cellules. Plus de 50 maladies humaines sont dues à des mutations des enzymes lysosomales, des protéines régulant des processus-clés du transport vers le lysosome ou des enzymes effectuant des modifications posttraductionnelles importantes pour la fonction du lysosome. L’objectif de cette thèse est d’identifier des protéines et des mécanismes permettant à la cellule de réguler le transport des enzymes vers le lysosome. Nous avons formulé l’hypothèse que des protéines mutées dans des maladies lysosomales et dont les fonctions étaient inconnues pouvaient jouer un rôle dans le transport vers le lysosome. Les céroïdes-lipofuscinoses neuronales forment une famille de maladies lysosomales rares mais sont aussi les maladies neurodégénératives infantiles les plus fréquentes. Plusieurs gènes impliqués dans les NCL encodent des protéines aux fonctions inconnues. Les travaux présentés dans cette thèse ont identifié la protéine « ceroid lipofuscinosis neuronal-5 » (CLN5) qui est localisée à l’endosome et au lysosome comme élément nécessaire au recrutement et à l’activation de rab7. Rab7 est une protéine Rab-clé qui contrôle le trafic à l’endosome tardif. Cette petite GTPase est impliquée dans le recrutement de retromer, un complexe protéique qui régule le trafic de l’endosome vers l’appareil de Golgi des récepteurs de tri lysosomal comme sortilin et le récepteur du mannose-6-phosphate. Dans les cellules où CLN5 est déplété, les récepteurs de tri lysosomal sont moins recyclés plus rapidement dégradés. En utilisant des expériences de photomarquage nous avons aussi pu démontrer que Rab7 est moins activées en l’absence de CLN5. Pour exécuter leur fonction les protéines rabs doivent être recrutée à la membrane et activées par l’échange d’une molécule de GDP pour une molécule de GTP. Le recrutement des Rabs à la membrane nécessite une modification posttraductionnelle lipidique pour être facilités. En utilisant un modèle de levures nous avons démontré que l’homologue de Rab7, Ypt7 est palmitoylée. Nous avons aussi démontré que la palmitoyltransférase Swif1 est nécessaire au recrutement de Ypt7 à la membrane. Nous avons aussi remarqué que les sous- unités de retromer chez la levure sont moins recrutées lorsque les palmitoyltransférases sont déplétées. Dans les cellules de mammifères nous avons démontré que Rab7 est également palmitoylé et que cette palmitoylation est possiblement effectuée par les palmitoyltransférases DHHC1 et DHHC8. La palmitoylation de Rab7 a lieu sur les cystéines en C-terminal qui sont nécessaires au recrutement membranaire et qui auparavant étaient uniquement décrites comme prénylées. En utilisant la méthode de « click chemistry » nous avons découvert que lorsque la prénylation de Rab7 est bloquée le niveau de palmitoylation augmente. Pour caractériser l’interaction entre CLN5 et Rab7 nous avons performé des expériences afin d’établir définitivement la topologie de cette protéine. Nous avons ainsi démontré que CLN5 est une protéine hautement glycosylée qui est initialement traduite en protéine transmembranaire et subséquemment clivée par un membre de la famille des peptidase de peptide signal (SPP). Cette protéine soluble peut alors possiblement interagir avec CLN3 qui est aussi palmitoylée pour recruter et activer Rab7. Nos études suggèrent pour la première fois que CLN5 pourrait être un recruteur et un activateur de Rab7 qui agirait avec la protéine CLN3 pour séquestrer Rab7 avec les autres récepteurs palmitoylés et permettre leur recyclage vers l’appareil de Golgi.

Architecture for neuronal cell control of a mobile robot

It is usually expected that the intelligent controlling mechanism of a robot is a computer system. Research is however now ongoing in which biological neural networks are being cultured and trained to act as the brain of an interactive real world robot - thereby either completely replacing or operating in a cooperative fashion with a computer system. Studying such neural systems can give a distinct insight into biological neural structures and therefore such research has immediate medical implications. In particular, the use of rodent primary dissociated cultured neuronal networks for the control of mobile `animals' (artificial animals, a contraction of animal and materials) is a novel approach to discovering the computational capabilities of networks of biological neurones. A dissociated culture of this nature requires appropriate embodiment in some form, to enable appropriate development in a controlled environment within which appropriate stimuli may be received via sensory data but ultimate influence over motor actions retained. The principal aims of the present research are to assess the computational and learning capacity of dissociated cultured neuronal networks with a view to advancing network level processing of artificial neural networks. This will be approached by the creation of an artificial hybrid system (animal) involving closed loop control of a mobile robot by a dissociated culture of rat neurons. This 'closed loop' interaction with the environment through both sensing and effecting will enable investigation of its learning capacity This paper details the components of the overall animat closed loop system and reports on the evaluation of the results from the experiments being carried out with regard to robot behaviour.

Emergence of a small-world functional network in cultured neurons

The functional networks of cultured neurons exhibit complex network properties similar to those found in vivo. Starting from random seeding, cultures undergo significant reorganization during the initial period in vitro, yet despite providing an ideal platform for observing developmental changes in neuronal connectivity, little is known about how a complex functional network evolves from isolated neurons. In the present study, evolution of functional connectivity was estimated from correlations of spontaneous activity. Network properties were quantified using complex measures from graph theory and used to compare cultures at different stages of development during the first 5 weeks in vitro. Networks obtained from young cultures (14 days in vitro) exhibited a random topology, which evolved to a small-world topology during maturation. The topology change was accompanied by an increased presence of highly connected areas (hubs) and network efficiency increased with age. The small-world topology balances integration of network areas with segregation of specialized processing units. The emergence of such network structure in cultured neurons, despite a lack of external input, points to complex intrinsic biological mechanisms. Moreover, the functional network of cultures at mature ages is efficient and highly suited to complex processing tasks.

Norepinephrine ignites local hot spots of neuronal excitation: how arousal amplifies selectivity in perception and memory

Arousal sometimes enhances and sometimes impairs perception and memory. In our Glutamate Amplifies Noradrenergic Effects (GANE) model, glutamate at active synapses interacts with norepinephrine released by the locus coeruleus to create local ‘hot spots’ of activity that enable the selective effects of arousal. This hot spot mechanism allows local cortical regions to self-regulate norepinephrine release based on current activation levels. In turn, hot spots bias global energetic delivery and functional network connectivity to enhance processing of high priority representations and impair processing of lower priority representations.

A multi-GPU algorithm for large-scale neuronal networks

Large-scale simulations of parts of the brain using detailed neuronal models to improve our understanding of brain functions are becoming a reality with the usage of supercomputers and large clusters. However, the high acquisition and maintenance cost of these computers, including the physical space, air conditioning, and electrical power, limits the number of simulations of this kind that scientists can perform. Modern commodity graphical cards, based on the CUDA platform, contain graphical processing units (GPUs) composed of hundreds of processors that can simultaneously execute thousands of threads and thus constitute a low-cost solution for many high-performance computing applications. In this work, we present a CUDA algorithm that enables the execution, on multiple GPUs, of simulations of large-scale networks composed of biologically realistic Hodgkin-Huxley neurons. The algorithm represents each neuron as a CUDA thread, which solves the set of coupled differential equations that model each neuron. Communication among neurons located in different GPUs is coordinated by the CPU. We obtained speedups of 40 for the simulation of 200k neurons that received random external input and speedups of 9 for a network with 200k neurons and 20M neuronal connections, in a single computer with two graphic boards with two GPUs each, when compared with a modern quad-core CPU. Copyright (C) 2010 John Wiley & Sons, Ltd.

Global slowing of network oscillations in mouse neocortex by diazepam

SCHEFFZUK, C. , KUKUSHKA, V. , VYSSOTSKI, A. L. , DRAGUHN, A. , TORT, A. B. L. , BRANKACK, J. . Global slowing of network oscillations in mouse neocortex by diazepam. Neuropharmacology , v. 65, p. 123-133, 2013.

Astrocytes and human cognition: Modeling information integration and modulation of neuronal activity

Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)

Endothelial and neuronal nitric oxide synthase inhibitors influences angiotensin II pressor effect in central nervous system

The present study investigated the central role of angiotensin II and nitric oxide on arterial blood pressure (MAP) in rats. Losartan and PD123349 AT 1 and AT 2 (selective no peptides antagonists angiotensin receptors), as well as FK 409 (a nitric oxide donor), N W-nitro-L-arginine methyl ester (L-NAME) a constituve nitric oxide synthase inhibitor endothelial (eNOSI) and 7-nitroindazol (7NI) a specific neuronal nitric oxide synthase inhibitor (nNOSI) were used. Holtzman strain, (Rattus norvergicus) weighting 200-250 g were anesthetized with zoletil 50 mg kg -1 (tiletamine chloridrate 125 mg and zolazepan chloridrate 125 mg) into quadriceps muscle anda stainless steel cannula was stereotaxically implanted into their Lateral Ventricle (LV). Controls were injected with a 0.5 μl volume of 0.15 M NaCl. Angiotensin II injected into LV increased MAP (19±3 vs. control 3±1 mm Hg), which is potentiated by prior injection of L-NAME in the same site 26±2 mm Hg. 7NI injected prior to ANG II into LV also potentiated the pressor effect of ANG II but with a higher intensity than L-NAME 32±3 mm Hg. FK 409 inhibited the pressor effect of ANG II (6±1 mm Hg). Losartan injected into LV before ANG II influences the pressor effect of ANG II (8±1 mm Hg). The PD 123319 decreased the pressor effects of ANG II (16±1 mm Hg). Losartan injected simultaneously with FK 409 blocked the pressor effect of ANG II (3±1 mm Hg). L-NAME produced an increase in the pressor effect of ANG II, may be due to local vasoconstriction and all at once by neuronal NOS inhibition but the main effect is of the 7-NIT an specific nNOS inhibitor. The AT 1 antagonist receptors improve basal nitric oxide (NO) production and release. These data suggest the involvement of constitutive and neuronal NOS in the control of arterial blood pressure induced by ANG II centrally, evolving AT 1 receptor-mediated vasoconstriction and AT 2 receptor-mediated vasodilatation. These results were confirmed by the experiment using FK 409. © 2006 Asian Network for Scientific Information.

Diverse levels of an inwardly rectifying potassium conductance generate heterogeneous neuronal behavior in a population of dorsal cochlear nucleus pyramidal neurons

Leao RM, Li S, Doiron B, Tzounopoulos T. Diverse levels of an inwardly rectifying potassium conductance generate heterogeneous neuronal behavior in a population of dorsal cochlear nucleus pyramidal neurons. J Neurophysiol 107: 3008-3019, 2012. First published February 29, 2012; doi:10.1152/jn.00660.2011.-Homeostatic mechanisms maintain homogeneous neuronal behavior among neurons that exhibit substantial variability in the expression levels of their ionic conductances. In contrast, the mechanisms, which generate heterogeneous neuronal behavior across a neuronal population, remain poorly understood. We addressed this problem in the dorsal cochlear nucleus, where principal neurons exist in two qualitatively distinct states: spontaneously active or not spontaneously active. Our studies reveal that distinct activity states are generated by the differential levels of a Ba2+-sensitive, inwardly rectifying potassium conductance (K-ir). Variability in K-ir maximal conductance causes variations in the resting membrane potential (RMP). Low K-ir conductance depolarizes RMP to voltages above the threshold for activating subthreshold-persistent sodium channels (Na-p). Once Na-p channels are activated, the RMP becomes unstable, and spontaneous firing is triggered. Our results provide a biophysical mechanism for generating neural heterogeneity, which may play a role in the encoding of sensory information.

Behavioral and EEG effects of GABAergic manipulation of the nigro-tectal pathway in the Wistar audiogenic rat (WAR) strain II: An EEG wavelet analysis and retrograde neuronal tracer approach

The role of the substantia nigra pars reticulata (SNPr) and superior colliculus (SC) network in rat strains susceptible to audiogenic seizures still remain underexplored in epileptology. In a previous study from our laboratory, the GABAergic drugs bicuculline (BIC) and muscimol (MUS) were microinjected into the deep layers of either the anterior SC (aSC) or the posterior SC (pSC) in animals of the Wistar audiogenic rat (WAR) strain submitted to acoustic stimulation, in which simultaneous electroencephalographic (EEG) recording of the aSC, pSC, SNPr and striatum was performed. Only MUS microinjected into the pSC blocked audiogenic seizures. In the present study, we expanded upon these previous results using the retrograde tracer Fluorogold (FG) microinjected into the aSC and pSC in conjunction with quantitative EEG analysis (wavelet transform), in the search for mechanisms associated with the susceptibility of this inbred strain to acoustic stimulation. Our hypothesis was that the WAR strain would have different connectivity between specific subareas of the superior colliculus and the SNPr when compared with resistant Wistar animals and that these connections would lead to altered behavior of this network during audiogenic seizures. Wavelet analysis showed that the only treatment with an anticonvulsant effect was MUS microinjected into the pSC region, and this treatment induced a sustained oscillation in the theta band only in the SNPr and in the pSC. These data suggest that in WAR animals, there are at least two subcortical loops and that the one involved in audiogenic seizure susceptibility appears to be the pSC-SNPr circuit. We also found that WARs presented an increase in the number of FG + projections from the posterior SNPr to both the aSC and pSC (primarily to the pSC), with both acting as proconvulsant nuclei when compared with Wistar rats. We concluded that these two different subcortical loops within the basal ganglia are probably a consequence of the WAR genetic background. (C) 2012 Elsevier Inc. All rights reserved.