Coemergence of regularity and complexity during neural network development

With the growing recognition that rhythmic and oscillatory patterns are widespread in the brain and play important roles in all aspects of the function of our nervous system, there has been a resurgence of interest in neuronal synchronized bursting activity. Here, we were interested in understanding the development of synchronized bursts as information-bearing neuronal activity patterns. For that, we have monitored the morphological organization and spontaneous activity of neuronal networks cultured on multielectrode-arrays during their self-executed evolvement from a mixture of dissociated cells into an active network. Complex collective network electrical activity evolved from sporadic firing patterns of the single neurons. On the system (network) level, the activity was marked by bursting events with interneuronal synchronization and nonarbitrary temporal ordering. We quantified these individual-to-collective activity transitions using newly-developed system level quantitative measures of time series regularity and complexity. We found that individual neuronal activity before synchronization was characterized by high regularity and low complexity. During neuronal wiring, there was a transient period of reorganization marked by low regularity, which then leads to coemergence of elevated regularity and functional (nonstochastic) complexity. We further investigated the morphology-activity interplay by modeling artificial neuronal networks with different topological organizations and connectivity schemes. The simulations support our experimental results by showing increased levels of complexity of neuronal activity patterns when neurons are wired up and organized in clusters (similar to mature real networks), as well as network-level activity regulation once collective activity forms.     

Bursting Reverberation as a Multiscale Neuronal Network Process Driven by Synaptic Depression-Facilitation

Neuronal networks can generate complex patterns of activity that depend on membrane properties of individual neurons as well as on functional synapses. To decipher the impact of synaptic properties and connectivity on neuronal network behavior, we investigate the responses of neuronal ensembles from small (5–30 cells in a restricted sphere) and large (acute hippocampal slice) networks to single electrical stimulation: in both cases, a single stimulus generated a synchronous long-lasting bursting activity. While an initial spike triggered a reverberating network activity that lasted 2–5 seconds for small networks, we found here that it lasted only up to 300 milliseconds in slices. To explain this phenomena present at different scales, we generalize the depression-facilitation model and extracted the network time constants. The model predicts that the reverberation time has a bell shaped relation with the synaptic density, revealing that the bursting time cannot exceed a maximum value. Furthermore, before reaching its maximum, the reverberation time increases sub-linearly with the synaptic density of the network. We conclude that synaptic dynamics and connectivity shape the mean burst duration, a property present at various scales of the networks. Thus bursting reverberation is a property of sufficiently connected neural networks, and can be generated by collective depression and facilitation of underlying functional synapses.     

Pharmacological Characterization of Cultivated Neuronal Networks: Relevance to Synaptogenesis and Synaptic Connectivity

Mental disorders, such as schizophrenia or Alzheimer’s disease, are associated with impaired synaptogenesis and/or synaptic communication. During development, neurons assemble into neuronal networks, the primary supracellular mediators of information processing. In addition to the orchestrated activation of genetic programs, spontaneous electrical activity and associated calcium signaling have been shown to be critically involved in the maturation of such neuronal networks. We established an in vitro model that recapitulates the maturation of neuronal networks, including spontaneous electrical activity. Upon plating, mouse primary hippocampal neurons grow neurites and interconnect via synapses to form a dish-wide neuronal network. Via live cell calcium imaging, we identified a limited period of time in which the spontaneous activity synchronizes across neurons, indicative of the formation of a functional network. After establishment of network activity, the neurons grow dendritic spines, the density of which was used as a morphological readout for neuronal maturity and connectivity. Hence, quantification of neurite outgrowth, synapse density, spontaneous neuronal activity, and dendritic spine density allowed to study neuronal network maturation from the day of plating until the presence of mature neuronal networks. Via acute pharmacological intervention, we show that synchronized network activity is mediated by the NMDA-R. The balance between kynurenic and quinolinic acid, both neuro-active intermediates in the tryptophan/kynurenine pathway, was shown to be decisive for the maintenance of network activity. Chronic modulation of the neurotrophic support influenced the network formation and revealed the extreme sensitivity of calcium imaging to detect subtle alterations in neuronal physiology. Given the reproducible cultivation in a 96-well setup in combination with fully automated analysis of the calcium recordings, this approach can be used to build a high-content screening assay usable for neurotoxicity screening, target identification/validation, or phenotypic drug screening.     

The phases of small networks of chemical reactors and neurons

We present an experimental study of the phase relationships observed in small reactor networks consisting of two and three continuous flow stirred tank reactors. In the three-reactor network one chemical oscillator is coupled to two other reactors in parallel in analogy to a small neural net. Each reactor contains an identical reaction mixture of the excitable Belousov-Zhabotinsky reaction which is characterized by its bifurcation diagram, where the electrical current is the bifurcation parameter. Coupling between the reactors is electrical via Pt-working electrodes and it can be either repulsive (inhibitory) or attractive (excitatory). An external electrical stimulus is applied to all three reactors in the form of an asymmetric electrical current pulse which sweeps across the bifurcation diagram. As a consequence, all three reactors oscillate with characteristic oscillation patterns or remain silent in analogy to the firing of neurons. The observed phase behavior depends on the type of coupling in a complex way. This situation is analogous to the in vivo measurements on single neurons (local neurons and projection neurons) performed by G. Laurent and co-workers on the olfactory system of the locust. We propose a simple neural network similar to the reactor network using the Hodgkin-Huxley model to simulate the action potentials of the coupled single neurons. Analogies between the reactor network and the neural network are discussed.     

Organization of neocortical neuronal networks

Multiple unit activity in deep layers of the frontal and motor cortices was recorded by chronically implanted semimicroelectrodes in waking cats with different levels of food motivation. From four to seven neuronal spike trains were selected from the recorded multiunit activity. Interactions between neighbouring neurons in the motor and frontal areas of the neocortex (within the local neuronal networks) and between the neurons of these areas (distributed neuronal networks) were estimated by means of statistical crosscorrelation analysis of spike trains within the range of delays from 0 to 100 ms. Neurons in the local networks were divided in two subgroups: the neurons with higher spike amplitudes with the dominance of divergent connections and neurons with lower spike amplitudes with the dominance of convergent connections. Strong monosynaptic connections (discharges with a delay of less than 2 ms) between the neurons with high- and low-amplitude spikes formed the background of the local networks. Connections between low-amplitude neurons in the frontal cortex and high-amplitude neurons in the motor cortex dominated in the distributed networks. A 24-hour food deprivation predominantly altered the late interneuronal crosscorrelations with time delays within the range of 2-100 ms in both local and distributed networks.     

Management of synchronized network activity by highly active neurons

Increasing evidence supports the idea that spontaneous brain activity may have an important functional role. Cultured neuronal networks provide a suitable model system to search for the mechanisms by which neuronal spontaneous activity is maintained and regulated. This activity is marked by synchronized bursting events (SBEs)--short time windows (hundreds of milliseconds) of rapid neuronal firing separated by long quiescent periods (seconds). However, there exists a special subset of rapidly firing neurons whose activity also persists between SBEs. It has been proposed that these highly active (HA) neurons play an important role in the management (i.e. establishment, maintenance and regulation) of the synchronized network activity. Here, we studied the dynamical properties and the functional role of HA neurons in homogeneous and engineered networks, during early network development, upon recovery from chemical inhibition and in response to electrical stimulations. We found that their sequences of inter-spike intervals (ISI) exhibit long time correlations and a unimodal distribution. During the network's development and under intense inhibition, the observed activity follows a transition period during which mostly HA neurons are active. Studying networks with engineered geometry, we found that HA neurons are precursors (the first to fire) of the spontaneous SBEs and are more responsive to electrical stimulations.     

Structure of spontaneous UP and DOWN transitions self-organizing in a cortical network model

Synaptic plasticity is considered to play a crucial role in the experience-dependent self-organization of local cortical networks. In the absence of sensory stimuli, cerebral cortex exhibits spontaneous membrane potential transitions between an UP and a DOWN state. To reveal how cortical networks develop spontaneous activity, or conversely, how spontaneous activity structures cortical networks, we analyze the self-organization of a recurrent network model of excitatory and inhibitory neurons, which is realistic enough to replicate UP–DOWN states, with spike-timing-dependent plasticity (STDP). The individual neurons in the self-organized network exhibit a variety of temporal patterns in the two-state transitions. In addition, the model develops a feed-forward network-like structure that produces a diverse repertoire of precise sequences of the UP state. Our model shows that the self-organized activity well resembles the spontaneous activity of cortical networks if STDP is accompanied by the pruning of weak synapses. These results suggest that the two-state membrane potential transitions play an active role in structuring local cortical circuits.     

Molecular and functional characterization of the electroneutral Na/HCO3 cotransporter NBCn1 in rat hippocampal neurons

We examined molecular and electrophysiological properties of the electroneutral sodium/bicarbonate cotransporter (NBCn1) that is present in rat hippocampal neurons. By PCR, a deletion variant (NBCn1-E) that lacks 123 amino acids in the cytoplasmic N-terminal domain was found in adult neurons. The previously characterized NBCn1-B, which does not have the deletion, was detected in embryonic neurons. In Xenopus oocytes, NBCn1-E raised the intracellular pH in the presence of HCO(3) without significantly affecting the membrane potential. Despite this electroneutral cotransport activity, the transporter mediated a steady-state current that positively shifted the resting potential by almost 30 mV. The mean reversal potential of the steady-state current was -21.2 mV, close to the resting potential of -21.4 mV. The reversal potential shifted 26 mV in response to a 10-fold increase of external Na(+) for concentrations above 10 mm. The current activity mediated by the transporter was unaffected by K(+), Mg(2+), Ca(2+), or Cl(-). Stable expression of NBCn1-E in human embryonic kidney cells also evoked an inward current that shifted the resting potentials more positive compared with the sham-transfected controls. In primary cultures of embryonic hippocampal neurons, the NBCn1 protein was localized in somatodendrites and synapses. NBCn1 protein was partially colocalized with the postsynaptic density protein PSD-95. Single-cell PCR showed that NBCn1 mRNA expression was present in both gamma-aminobutyric acid (GABA)ergic and non-GABAergic neurons. We propose that NBCn1 in hippocampal neurons may affect neuronal activity by regulating local pH as well as steady-state inward currents at synapses.     

Olfactory coding in the insect brain: molecular receptive ranges, spatial and temporal coding

Odor information is coded in the insect brain in a sequence of steps, ranging from the receptor cells, via the neural network in the antennal lobe, to higher order brain centers, among which the mushroom bodies and the lateral horn are the most prominent. Across all of these processing steps, coding logic is combinatorial, in the sense that information is represented as patterns of activity across a population of neurons, rather than in individual neurons. Because different neurons are located in different places, such a coding logic is often termed spatial, and can be visualized with optical imaging techniques. We employ in vivo calcium imaging in order to record odor‐evoked activity patterns in olfactory receptor neurons, different populations of local neurons in the antennal lobes, projection neurons linking antennal lobes to the mushroom bodies, and the intrinsic cells of the mushroom bodies themselves, the Kenyon cells. These studies confirm the combinatorial nature of coding at all of these stages. However, the transmission of odor‐evoked activity patterns from projection neuron dendrites via their axon terminals onto Kenyon cells is accompanied by a progressive sparsening of the population code. Activity patterns also show characteristic temporal properties. While a part of the temporal response properties reflect the physical sequence of odor filaments, another part is generated by local neuron networks. In honeybees, γ‐aminobutyric acid (GABA)‐ergic and histaminergic neurons both contribute inhibitory networks to the antennal lobe. Interestingly, temporal properties differ markedly in different brain areas. In particular, in the antennal lobe odor‐evoked activity develops over slow time courses, while responses in Kenyon cells are phasic and transient. The termination of an odor stimulus is reflected by a decrease in activity within most glomeruli of the antennal lobe and an off‐response in some glomeruli, while in the mushroom bodies about half of the odor‐activated Kenyon cells also exhibit off‐responses.     

Interfacing 3D Engineered Neuronal Cultures to Micro-Electrode Arrays: An Innovative In Vitro Experimental Model

Currently, large-scale networks derived from dissociated neurons growing and developing in vitro on extracellular micro-transducer devices are the gold-standard experimental model to study basic neurophysiological mechanisms involved in the formation and maintenance of neuronal cell assemblies. However, in vitro studies have been limited to the recording of the electrophysiological activity generated by bi-dimensional (2D) neural networks. Nonetheless, given the intricate relationship between structure and dynamics, a significant improvement is necessary to investigate the formation and the developing dynamics of three-dimensional (3D) networks. In this work, a novel experimental platform in which 3D hippocampal or cortical networks are coupled to planar Micro-Electrode Arrays (MEAs) is presented. 3D networks are realized by seeding neurons in a scaffold constituted of glass microbeads (30-40 µm in diameter) on which neurons are able to grow and form complex interconnected 3D assemblies. In this way, it is possible to design engineered 3D networks made up of 5-8 layers with an expected final cell density. The increasing complexity in the morphological organization of the 3D assembly induces an enhancement of the electrophysiological patterns displayed by this type of networks. Compared with the standard 2D networks, where highly stereotyped bursting activity emerges, the 3D structure alters the bursting activity in terms of duration and frequency, as well as it allows observation of more random spiking activity. In this sense, the developed 3D model more closely resembles in vivo neural networks.     

The interrelations between neurons of the amygdala and hypothalamus during conditioning with selection of food reinforcement quality in cats

Three cats were subjected to appetitive instrumental conditioning to light by the method of the "active choice" of the reinforcement quality. The short-delayed conditioned bar-pressings were reinforced by bread-meat mixture and the delayed response by meat. The animals differed in behavior strategy: two animals preferred bar-pressing with long delay (the so-called "self-control" group) and one animal preferred bar-pressing with short delay (the so-called "impulsive" group). The multiunit activity of the basolateral amygdala and nucleus lateralis of the hypothalamus was recorded through chronically implanted nichrome wire semimicroelecrodes. The interactions between the neighboring neurons in the lateral hypothalamus and basolateral amygdala (within the local neuronal network) and between the neurons of the basolateral amygdala and lateral hypothalamus (distributed neuronal networks in the direction amygdala--hypothalamus and vice versa) were evaluated by means of statistical crosscorrelation analysis of spike trains. The crosscorrelational interneuronal connections in the delay range of 0-100 ms were examined. It was shown that the number of crosscorrelations between the discharges on neurons both in the local networks of basolateral amygdala and distributed networks was significantly higher in "impulsive" cats. In both groups of animals, the percentage of crosscorrelations between neighbouring neurons in the local networks of the lateral hypothalamus was similar. We suggest that the local networks of the basolateral amygdala and amygdalar-hypothalamic distributed neuronal networks are involved in the system of brain structures which determine the individual features of animal behavior.     

Activity-dependent plasticity of electrical synapses: increasing evidence for its presence and functional roles in the mammalian brain

Gap junctions mediate electrical synaptic transmission between neurons. While the actions of neurotransmitter modulators on the conductance of gap junctions have been extensively documented, increasing evidence indicates they can also be influenced by the ongoing activity of neural networks, in most cases via local interactions with nearby glutamatergic synapses. We review here early evidence for the existence of activity-dependent regulatory mechanisms as well recent examples reported in mammalian brain. The ubiquitous distribution of both neuronal connexins and the molecules involved suggest this phenomenon is widespread and represents a property of electrical transmission in general.     

Seasonal encoding by the circadian pacemaker of the SCN

The circadian pacemaker of the suprachiasmatic nucleus (SCN) functions as a seasonal clock through its ability to encode day length [1-6]. To investigate the mechanism by which SCN neurons code for day length, we housed mice under long (LD 16:8) and short (LD 8:16) photoperiods. Electrophysiological recordings of multiunit activity (MUA) in the SCN of freely moving mice revealed broad activity profiles in long days and compressed activity profiles in short days. The patterns remained consistent after release of the mice in constant darkness. Recordings of MUA in acutely prepared hypothalamic slices showed similar differences between the SCN electrical activity patterns in vitro in long and short days. In vitro recordings of neuronal subpopulations revealed that the width of the MUA activity profiles was determined by the distribution of phases of contributing units within the SCN. The subpopulation patterns displayed a significantly broader distribution in long days than in short days. Long-term recordings of single-unit activity revealed short durations of elevated activity in both short and long days (3.48 and 3.85 hr, respectively). The data indicate that coding for day length involves plasticity within SCN neuronal networks in which the phase distribution of oscillating neurons carries information on the photoperiod's duration.     

Neuromodulation targets intrinsic cardiac neurons to attenuate neuronally mediated atrial arrhythmias

Our objective was to determine whether atrial fibrillation (AF) results from excessive activation of intrinsic cardiac neurons (ICNs) and, if so, whether select subpopulations of neurons therein represent therapeutic targets for suppression of this arrhythmogenic potential. Trains of five electrical stimuli (0.3-1.2 mA, 1 ms) were delivered during the atrial refractory period to mediastinal nerves (MSN) on the superior vena cava to evoke AF. Neuroanatomical studies were performed by injecting the neuronal tracer DiI into MSN sites that induced AF. Functional studies involved recording of neuronal activity in situ from the right atrial ganglionated plexus (RAGP) in response to MSN stimulation (MSNS) prior to and following neuromodulation involving either preemptive spinal cord stimulation (SCS; T(1)-T(3), 50 Hz, 200-ms duration) or ganglionic blockade (hexamethonium, 5 mg/kg). The tetramethylindocarbocyanine perchlorate (DiI) neuronal tracer labeled a subset (13.2%) of RAGP neurons, which also colocalized with cholinergic or adrenergic markers. A subset of DiI-labeled RAGP neurons were noncholinergic/nonadrenergic. MSNS evoked an ～4-fold increase in RAGP neuronal activity from baseline, which SCS reduced by 43%. Hexamethonium blocked MSNS-evoked increases in neuronal activity. MSNS evoked AF in 78% of right-sided MSN sites, which SCS reduced to 33% and hexamethonium reduced to 7%. MSNS-induced bradycardia was maintained with SCS but was mitigated by hexamethonium. We conclude that MSNS activates subpopulations of intrinsic cardiac neurons, thereby resulting in the formation of atrial arrhythmias leading to atrial fibrillation. Stabilization of ICN local circuit neurons by SCS or the local circuit and autonomic efferent neurons with hexamethonium reduces the arrhythmogenic potential.     

Development of coherent neuronal activity patterns in mammalian cortical networks: Common principles and local hetereogeneity

Many mammals are born in a very immature state and develop their rich repertoire of behavioral and cognitive functions postnatally. This development goes in parallel with changes in the anatomical and functional organization of cortical structures which are involved in most complex activities. The emerging spatiotemporal activity patterns in multi-neuronal cortical networks may indeed form a direct neuronal correlate of systemic functions like perception, sensorimotor integration, decision making or memory formation. During recent years, several studies – mostly in rodents – have shed light on the ontogenesis of such highly organized patterns of network activity. While each local network has its own peculiar properties, some general rules can be derived. We therefore review and compare data from the developing hippocampus, neocortex and – as an intermediate region – entorhinal cortex. All cortices seem to follow a characteristic sequence starting with uncorrelated activity in uncoupled single neurons where transient activity seems to have mostly trophic effects. In rodents, before and shortly after birth, cortical networks develop weakly coordinated multineuronal discharges which have been termed synchronous plateau assemblies (SPAs). While these patterns rely mostly on electrical coupling by gap junctions, the subsequent increase in number and maturation of chemical synapses leads to the generation of large-scale coherent discharges. These patterns have been termed giant depolarizing potentials (GDPs) for predominantly GABA-induced events or early network oscillations (ENOs) for mostly glutamatergic bursts, respectively. During the third to fourth postnatal week, cortical areas reach their final activity patterns with distinct network oscillations and highly specific neuronal discharge sequences which support adult behavior. While some of the mechanisms underlying maturation of network activity have been elucidated much work remains to be done in order to fully understand the rules governing transition from immature to mature patterns of network activity.     

Ontogenetic development of neural responses in the olfactory bulb of laboratory mice

Summary In laboratory mice (strain NMRI) the ontogenetic development of single unit activity in the olfactory bulb was investigated. From postnatal day 10 on, spontaneously active neurons were recorded with glass-microelectrodes, and their responses to olfactory stimuli were tested (butyric acid, geraniol, grass- and nest-odour).From day 10 to 13 only very few neurons were recordable (and most of these elements were too weak and were lost before being stimulated). At day 14 the number of recordable neurons increased rapidly, and by day 15 spontaneously active neurons reached adult level in terms of incidence and electric properties.There were 3 types of neurons: 1. respiration synchronous elements; 2. bursting neurons not correlated with respiration; 3. continuously, but randomly, firing elements (about 60% of all neurons). Reactions to odour stimuli (excitation, ca. 50%; inhibition, ca. 34%; complex reactions, ca. 12%; change in activity pattern, ca. 4%) occurred as soon as the cells were stable enough for testing. The reaction patterns showed no age specific differences; the duration of the responses varied from 100 ms to 100 s.In younger animals (P11–P14) the percentage of responses was slightly smaller (47%) than in the older ones (P30–P50; 64% response to olfactory stimulation). For some of the odours tested the proportion of responding cells differed depending on age (for instance grass odour evoked a response in 40% of the cells in young ones, but in 65% in adults).Abbreviations AP action potential - In interneuron - MTc mitral or tufted cell - P10 postnatal day 10     

Local interneurons and information processing in the olfactory glomeruli of the moth Manduca sexta

Intracellular recordings were made from the major neurites of local interneurons in the moth antennal lobe. Antennal nerve stimulation evoked 3 patterns of postsynaptic activity: (i) a short-latency compound excitatory postsynaptic potential that, based on electrical stimulation of the antennal nerve and stimulation of the antenna with odors, represents a monosynaptic input from olfactory afferent axons (71 out of 86 neurons), (ii) a delayed activation of firing in response to both electrical- and odor-driven input (11 neurons), and (iii) a delayed membrane hyperpolarization in response to antennal nerve input (4 neurons).Simultaneous intracellular recordings from a local interneuron with short-latency responses and a projection (output) neuron revealed unidirectional synaptic interactions between these two cell types. In 20% of the 30 pairs studied, spontaneous and current-induced spiking activity in a local interneuron correlated with hyperpolarization and suppression of firing in a projection neuron. No evidence for recurrent or feedback inhibition of projection neurons was found. Furthermore, suppression of firing in an inhibitory local interneuron led to an increase in firing in the normally quiescent projection neuron, suggesting that a disinhibitory pathway may mediate excitation in projection neurons. This is the first direct evidence of an inhibitory role for local interneurons in olfactory information processing in insects. Through different types of multisynaptic interactions with projection neurons, local interneurons help to generate and shape the output from olfactory glomeruli in the antennal lobe.Abbreviations AL antennal lobe - EPSP excitatory postsynaptic potential - GABA -aminobutyric acid - IPSP inhibitory postsynaptic potential - LN local interneuron - MGC macroglomerular complex - OB olfactory bulb - PN projection neuron - TES N-tris[hydroxymethyl]methyl-2-aminoethane-sulfonic acid     

Conditions increasing the adrenergic properties of dissociated chick superior cervical ganglion neurons grown in long-term culture

Neurons dissociated from the embryonic chick superior cervical ganglion (SCG) were separated from ganglionic nonneuronal cells using a density gradient formed with Percoll. The sympathetic neurons were then grown for 3-4 weeks in serum containing medium on a polyornithine substrate precoated with heart-conditioned medium. Both catecholamine (CA) and acetylcholine (ACh) are synthesized and accumulated by these neurons, but the amount of CA is higher and increases much more over time in culture than the amount of ACh. The cultures become therefore more adrenergic with time. We report here that the adrenergic properties of these cells can be enhanced. A 3-fold increase in CA synthesis, as expressed on a per neuron basis, is obtained by increasing neuron cell density 3- to 4-fold. ACh synthesis, however, is decreased at high neuronal density. Optimal CA production is obtained at densities of 120-150,000 neurons/cm2. This effect is due to direct cell contact since it cannot be transferred to low density cultures by medium conditioned by high density cultures. Nerve growth factor concentrations 5-10-fold higher than the amount necessary for optimal neuronal survival (1 microgram/ml 7S NGF) increases CA production but do not affect ACh synthesis. This effect is highest at low plating densities (20-30,000 neurons/cm2, 10-fold increase) and progressively decreases with increasing neuronal density. No increase is obtained in high density cultures where CA production is maximal. In addition, we made the novel observation that medium conditioned by chick liver cells in culture (LCM) increases CA production approximately 4-fold, whereas it does not increase ACh production by the SCG neurons. Work is in progress to biochemically characterize the active component(s) present in the LCM and to determine whether they favor the survival of a subpopulation of adrenergic neurons possible present in these ganglia. Alternatively, the adrenergic differentiation of neurons initially capable of synthesizing both CA and ACh could be selectively increased by LCM.