Corrigendum: Spectral Entropy Based Neuronal Network Synchronization Analysis Based on Microelectrode Array Measurements.

[This corrects the article DOI: 10.3389/fncom.2016.00112.].

Network-Wide Adaptive Burst Detection Depicts Neuronal Activity with Improved Accuracy.

Neuronal networks are often characterized by their spiking and bursting statistics. Previously, we introduced an adaptive burst analysis method which enhances the analysis power for neuronal networks with highly varying firing dynamics. The adaptation is based on single channels analyzing each element of a network separately. Such kind of analysis was adequate for the assessment of local behavior, where the analysis focuses on the neuronal activity in the vicinity of a single electrode. However, the assessment of the whole network may be hampered, if parts of the network are analyzed using different rules. Here, we test how using multiple channels and measurement time points affect adaptive burst detection. The main emphasis is, if network-wide adaptive burst detection can provide new insights into the assessment of network activity. Therefore, we propose a modification to the previously introduced inter-spike interval (ISI) histogram based cumulative moving average (CMA) algorithm to analyze multiple spike trains simultaneously. The network size can be freely defined, e.g., to include all the electrodes in a microelectrode array (MEA) recording. Additionally, the method can be applied on a series of measurements on the same network to pool the data for statistical analysis. Firstly, we apply both the original CMA-algorithm and our proposed network-wide CMA-algorithm on artificial spike trains to investigate how the modification changes the burst detection. Thereafter, we use the algorithms on MEA data of spontaneously active chemically manipulated in vitro rat cortical networks. Moreover, we compare the synchrony of the detected bursts introducing a new burst synchrony measure. Finally, we demonstrate how the bursting statistics can be used to classify networks by applying k-means clustering to the bursting statistics. The results show that the proposed network wide adaptive burst detection provides a method to unify the burst definition in the whole network and thus improves the assessment and classification of the neuronal activity, e.g., the effects of different pharmaceuticals. The results indicate that the novel method is adaptive enough to be usable on networks with different dynamics, and it is especially feasible when comparing the behavior of differently spiking networks, for example in developing networks.

Spectral Entropy Based Neuronal Network Synchronization Analysis Based on Microelectrode Array Measurements.

Synchrony and asynchrony are essential aspects of the functioning of interconnected neuronal cells and networks. New information on neuronal synchronization can be expected to aid in understanding these systems. Synchronization provides insight in the functional connectivity and the spatial distribution of the information processing in the networks. Synchronization is generally studied with time domain analysis of neuronal events, or using direct frequency spectrum analysis, e.g., in specific frequency bands. However, these methods have their pitfalls. Thus, we have previously proposed a method to analyze temporal changes in the complexity of the frequency of signals originating from different network regions. The method is based on the correlation of time varying spectral entropies (SEs). SE assesses the regularity, or complexity, of a time series by quantifying the uniformity of the frequency spectrum distribution. It has been previously employed, e.g., in electroencephalogram analysis. Here, we revisit our correlated spectral entropy method (CorSE), providing evidence of its justification, usability, and benefits. Here, CorSE is assessed with simulations and in vitro microelectrode array (MEA) data. CorSE is first demonstrated with a specifically tailored toy simulation to illustrate how it can identify synchronized populations. To provide a form of validation, the method was tested with simulated data from integrate-and-fire model based computational neuronal networks. To demonstrate the analysis of real data, CorSE was applied on in vitro MEA data measured from rat cortical cell cultures, and the results were compared with three known event based synchronization measures. Finally, we show the usability by tracking the development of networks in dissociated mouse cortical cell cultures. The results show that temporal correlations in frequency spectrum distributions reflect the network relations of neuronal populations. In the simulated data, CorSE unraveled the synchronizations. With the real in vitro MEA data, CorSE produced biologically plausible results. Since CorSE analyses continuous data, it is not affected by possibly poor spike or other event detection quality. We conclude that CorSE can reveal neuronal network synchronization based on in vitro MEA field potential measurements. CorSE is expected to be equally applicable also in the analysis of corresponding in vivo and ex vivo data analysis.

Analyzing the feasibility of time correlated spectral entropy for the assessment of neuronal synchrony.

In this paper, we study neuronal network analysis based on microelectrode measurements. We search for potential relations between time correlated changes in spectral distributions and synchrony for neuronal network activity. Spectral distribution is quantified by spectral entropy as a measure of uniformity/complexity and this measure is calculated as a function of time for the recorded neuronal signals, i.e., time variant spectral entropy. Time variant correlations in the spectral distributions between different parts of a neuronal network, i.e., of concurrent measurements via different microelectrodes, are calculated to express the relation with a single scalar. We demonstrate these relations with in vivo rat hippocampal recordings, and observe the time courses of the correlations between different regions of hippocampus in three sequential recordings. Additionally, we evaluate the results with a commonly employed causality analysis method to assess the possible correlated findings. Results show that time correlated spectral entropy reveals different levels of interrelations in neuronal networks, which can be interpreted as different levels of neuronal network synchrony.

Phase matters: responding to and learning about peripheral stimuli depends on hippocampal θ phase at stimulus onset.

Hippocampal θ (3-12 Hz) oscillations are implicated in learning and memory, but their functional role remains unclear. We studied the effect of the phase of local θ oscillation on hippocampal responses to a neutral conditioned stimulus (CS) and subsequent learning of classical trace eyeblink conditioning in adult rabbits. High-amplitude, regular hippocampal θ-band responses (that predict good learning) were elicited by the CS when it was timed to commence at the fissure θ trough (Trough group). Regardless, learning in this group was not enhanced compared with a yoked control group, possibly due to a ceiling effect. However, when the CS was consistently presented to the peak of θ (Peak group), hippocampal θ-band responding was less organized and learning was retarded. In well-trained animals, the hippocampal θ phase at CS onset no longer affected performance of the learned response, suggesting a time-limited role for hippocampal processing in learning. To our knowledge, this is the first study to demonstrate that timing a peripheral stimulus to a specific phase of the hippocampal θ cycle produces robust effects on the synchronization of neural responses and affects learning at the behavioral level. Our results support the notion that the phase of spontaneous hippocampal θ oscillation is a means of regulating the processing of information in the brain to a behaviorally relevant degree.

Quantification and automatized adaptive detection of in vivo and in vitro neuronal bursts based on signal complexity.

In this paper, we propose employing entropy values to quantify action potential bursts in electrophysiological measurements from the brain and neuronal cultures. Conventionally in the electrophysiological signal analysis, bursts are quantified by means of conventional measures such as their durations, and number of spikes in bursts. Here our main aim is to device metrics for burst quantification to provide for enhanced burst characterization. Entropy is a widely employed measure to quantify regularity/complexity of time series. Specifically, we investigate the applicability and differences of spectral entropy and sample entropy in the quantification of bursts in in vivo rat hippocampal measurements and in in vitro dissociated rat cortical cell culture measurement done with microelectrode arrays. For the task, an automatized and adaptive burst detection method is also utilized. Whereas the employed metrics are known from other applications, they are rarely employed in the assessment of burst in electrophysiological field potential measurements. Our results show that the proposed metrics are potential for the task at hand.

Quantitative examination of stimulus-response relations in cortical networks in vitro.

Variable responses of neuronal networks to repeated sensory or electrical stimuli reflect the interaction of the stimulus' response with ongoing activity in the brain and its modulation by adaptive mechanisms, such as cognitive context, network state, or cellular excitability and synaptic transmission capability. Here, we focus on reliability, length, delays, and variability of evoked responses with respect to their spatial distribution, interaction with spontaneous activity in the networks, and the contribution of GABAergic inhibition. We identified network-intrinsic principles that underlie the formation and modulation of spontaneous activity and stimulus-response relations with the use of state-dependent stimulation in generic neuronal networks in vitro. The duration of spontaneously recurring network-wide bursts of spikes was best predicted by the length of the preceding interval. Length, delay, and structure of responses to identical stimuli systematically depended on stimulus timing and distance to the stimulation site, which were described by a set of simple functions of spontaneous activity. Response length at proximal recording sites increased with the duration of prestimulus inactivity and was best described by a saturation function y(t) = A(1 - e(-αt)). Concomitantly, the delays of polysynaptic late responses at distant sites followed an exponential decay y(t) = Be(-ßt) + C. In addition, the speed of propagation was determined by the overall state of the network at the moment of stimulation. Disinhibition increased the number of spikes/network burst and interburst interval length at unchanged gross firing rate, whereas the response modulation by the duration of prestimulus inactivity was preserved. Our data suggest a process of network depression during bursts and subsequent recovery that limit evoked responses following distinct rules. We discuss short-term synaptic depression due to depletion of neurotransmitter vesicles as an underlying mechanism. The seemingly unreliable patterns of spontaneous activity and stimulus-response relations thus follow a predictable structure determined by the interdependencies of network structures and activity states.

Burst analysis tool for developing neuronal networks exhibiting highly varying action potential dynamics.

In this paper we propose a firing statistics based neuronal network burst detection algorithm for neuronal networks exhibiting highly variable action potential dynamics. Electrical activity of neuronal networks is generally analyzed by the occurrences of spikes and bursts both in time and space. Commonly accepted analysis tools employ burst detection algorithms based on predefined criteria. However, maturing neuronal networks, such as those originating from human embryonic stem cells (hESCs), exhibit highly variable network structure and time-varying dynamics. To explore the developing burst/spike activities of such networks, we propose a burst detection algorithm which utilizes the firing statistics based on interspike interval (ISI) histograms. Moreover, the algorithm calculates ISI thresholds for burst spikes as well as for pre-burst spikes and burst tails by evaluating the cumulative moving average (CMA) and skewness of the ISI histogram. Because of the adaptive nature of the proposed algorithm, its analysis power is not limited by the type of neuronal cell network at hand. We demonstrate the functionality of our algorithm with two different types of microelectrode array (MEA) data recorded from spontaneously active hESC-derived neuronal cell networks. The same data was also analyzed by two commonly employed burst detection algorithms and the differences in burst detection results are illustrated. The results demonstrate that our method is both adaptive to the firing statistics of the network and yields successful burst detection from the data. In conclusion, the proposed method is a potential tool for analyzing of hESC-derived neuronal cell networks and thus can be utilized in studies aiming to understand the development and functioning of human neuronal networks and as an analysis tool for in vitro drug screening and neurotoxicity assays.

Disrupting neural activity related to awake-state sharp wave-ripple complexes prevents hippocampal learning.

Oscillations in hippocampal local-field potentials (LFPs) reflect the crucial involvement of the hippocampus in memory trace formation: theta (4-8 Hz) oscillations and ripples (~200 Hz) occurring during sharp waves are thought to mediate encoding and consolidation, respectively. During sharp wave-ripple complexes (SPW-Rs), hippocampal cell firing closely follows the pattern that took place during the initial experience, most likely reflecting replay of that event. Disrupting hippocampal ripples using electrical stimulation either during training in awake animals or during sleep after training retards spatial learning. Here, adult rabbits were trained in trace eyeblink conditioning, a hippocampus-dependent associative learning task. A bright light was presented to the animals during the inter-trial interval (ITI), when awake, either during SPW-Rs or irrespective of their neural state. Learning was particularly poor when the light was presented following SPW-Rs. While the light did not disrupt the ripple itself, it elicited a theta-band oscillation, a state that does not usually coincide with SPW-Rs. Thus, it seems that consolidation depends on neuronal activity within and beyond the hippocampus taking place immediately after, but by no means limited to, hippocampal SPW-Rs.

Human embryonic stem cell-derived neuronal cells form spontaneously active neuronal networks in vitro.

The production of functional human embryonic stem cell (hESC)-derived neuronal cells is critical for the application of hESCs in treating neurodegenerative disorders. To study the potential functionality of hESC-derived neurons, we cultured and monitored the development of hESC-derived neuronal networks on microelectrode arrays. Immunocytochemical studies revealed that these networks were positive for the neuronal marker proteins beta-tubulin(III) and microtubule-associated protein 2 (MAP-2). The hESC-derived neuronal networks were spontaneously active and exhibited a multitude of electrical impulse firing patterns. Synchronous bursts of electrical activity similar to those reported for hippocampal neurons and rodent embryonic stem cell-derived neuronal networks were recorded from the differentiated cultures until up to 4 months. The dependence of the observed neuronal network activity on sodium ion channels was examined using tetrodotoxin (TTX). Antagonists for the glutamate receptors NMDA [D(-)-2-amino-5-phosphonopentanoic acid] and AMPA/kainate [6-cyano-7-nitroquinoxaline-2,3-dione], and for GABAA receptors [(-)-bicuculline methiodide] modulated the spontaneous electrical activity, indicating that pharmacologically susceptible neuronal networks with functional synapses had been generated. The findings indicate that hESC-derived neuronal cells can generate spontaneously active networks with synchronous communication in vitro, and are therefore suitable for use in developmental and drug screening studies, as well as for regenerative medicine.

Contribution of a single CA3 neuron to network synchrony.

Oscillations at theta (3-8 Hz) and gamma (30-80 Hz) frequencies co-occur during arousal, exploration, and rapid eye movement sleep and relate to information processing underlying learning and memory within neuronal networks. In hippocampus, gamma and theta frequency oscillations are associated with modification of synaptic weights, spatial learning, and short-term memory. These oscillations are referred to as network phenomena and, thereby, the role of single neuron oscillations in the generation of neuronal networks remains unclear. We report that an individual CA3 pyramidal cell can activate the CA1 neuronal network in vivo in rat hippocampus using electrical stimulations with simultaneous intracellular gamma and extracellular theta and slow (0.5-1 Hz) frequencies. These results suggest that an individual pyramidal cell can contribute to self-organization of a neuronal small-scale network.

Observing frequency content time evolution of independent hippocampal signals.

In this paper, we propose a method for observing frequency contents of independent hippocampal signals in time. The method is based on calculating independent component analysis (ICA) on electrophysiological multielectrode field potential measurements (MFPMs) in a running window. We have previously proposed a method for observing independently operating neural populations, i.e., functional populations (FUPOs) from MFPMs and outlined the concept, which is elaborated upon and extended in this paper, in order to facilitate analysis of functioning of the target brain area. In this paper, the proposed method is demonstrated with an example with three concurrent hippocampal measurements from an anesthetized rat brain. The proposed method can be applied in analysis of any recordings of neural networks in which contributions from a number of neural populations (NPs) are simultaneously recorded via a number of measurement points (MPs), as well in vivo as in vitro.

Independent component analysis of neural populations from multielectrode field potential measurements.

Independent component analysis (ICA) is proposed for analysis of neural population activity from multichannel electrophysiological field potential measurements. The proposed analysis method provides information on spatial extents of active neural populations, locations of the populations with respect to each other, population evolution, including merging and splitting of populations in time, and on time lag differences between the populations. In some cases, results of the proposed analysis may also be interpreted as independent information flows carried by neurons and neural populations. In this paper, a detailed description of the analysis method is given. The proposed analysis is demonstrated with an illustrative simulation, and with an exemplary analysis of an in vivo multichannel recording from rat hippocampus. The proposed method can be applied in analysis of any recordings of neural networks in which contributions from a number of neural populations or information flows are simultaneously recorded via a number of measurement points, as well in vivo as in vitro.