Experimental analysis and computational modeling of interburst intervals in spontaneous activity of cortical neuronal culture.

Rhythmic bursting is the most striking behavior of cultured cortical networks and may start in the second week after plating. In this study, we focus on the intervals between spontaneously occurring bursts, and compare experimentally recorded values with model simulations. In the models, we use standard neurons and synapses, with physiologically plausible parameters taken from literature. All networks had a random recurrent architecture with sparsely connected neurons. The number of neurons varied between 500 and 5,000. We find that network models with homogeneous synaptic strengths produce asynchronous spiking or stable regular bursts. The latter, however, are in a range not seen in recordings. By increasing the synaptic strength in a (randomly chosen) subset of neurons, our simulations show interburst intervals (IBIs) that agree better with in vitro experiments. In this regime, called weakly synchronized, the models produce irregular network bursts, which are initiated by neurons with relatively stronger synapses. In some noise-driven networks, a subthreshold, deterministic, input is applied to neurons with strong synapses, to mimic pacemaker network drive. We show that models with such "intrinsically active neurons" (pacemaker-driven models) tend to generate IBIs that are determined by the frequency of the fastest pacemaker and do not resemble experimental data. Alternatively, noise-driven models yield realistic IBIs. Generally, we found that large-scale noise-driven neuronal network models required synaptic strengths with a bimodal distribution to reproduce the experimentally observed IBI range. Our results imply that the results obtained from small network models cannot simply be extrapolated to models of more realistic size. Synaptic strengths in large-scale neuronal network simulations need readjustment to a bimodal distribution, whereas small networks do not require such changes.

Neural cell-cell and cell-substrate adhesion through N-cadherin, N-CAM and L1.

In this study neural (N)-cadherin, neural cell adhesion molecule (N-CAM) and L1 proteins and their antibody equivalents were covalently immobilized on a polyethylene-imine (PEI)-coated glass surface to form neuron-adhesive coatings. Impedance sensing and (supplementary) image analysis were used to monitor the effects of these CAMs. Immobilization of high concentrations of both N-cadherin protein and antibody led to good adhesion of neurons to the modified surface, better than surfaces treated with 30.0 and 100.0 µg ml(-1) N-CAM protein and antibody. L1 antibody and protein coating revealed no significant effect on neuronal cell-substrate adhesion. In a second series of combinatorial experiments, we used the same antibodies and proteins as medium-additives to inhibit cell-cell adhesion between neurons. Adhesion of neurons cultured on N-cadherin protein or antibody-modified surfaces was lowered by the addition of a soluble N-cadherin protein and antibody to the culturing medium, accelerating neuronal aggregation. The presence of a soluble N-CAM antibody or protein had no effect on the adhesion of neuronal cells on a N-cadherin protein-modified surface. On a N-cadherin antibody-coated surface, the addition of a soluble N-CAM protein led to cell death of neurons after 48 h, while a N-CAM antibody had no effect. In the presence of a soluble N-cadherin protein and antibody the aggregation of neurons was inhibited, both on N-CAM protein and N-CAM antibody-modified surfaces. Neurons cultured on immobilized antibodies were less affected by the addition of soluble CAM blockers than neurons cultured on immobilized proteins, indicating that antibody-protein bonds are more stable compared to protein-protein bonds.

Inhibition of neuronal cell-cell adhesion measured by the microscopic aggregation assay and impedance sensing.

Microscopic aggregation assay and impedance sensing (IS) were used to monitor a change in in vitro neuron-neuron adhesion in response to blocking of cell adhesion molecules. By blocking neuron-neuron adhesion, migration and aggregation of neuronal cells can be inhibited. This leads to better control of spatial arrangement of cells in culture. In the literature N-CAM, L1 and N-cadherin proteins are pointed out as main regulators of neuronal adhesion. In this study, these three main cell adhesion molecules were used to inhibit neuron-to-neuron adhesion and aggregation. Both soluble extracellular domains and antigen antibodies were added to these adhesion molecules. They were investigated for their blocking ability in neuronal cultures. First, in a 96 h aggregation assay on a low-adhesive substrate, the effect of inhibition of the three proteins on aggregation of cortical neurons was investigated optically. Both L1 antibody and L1 protein had no effect on the degree of aggregation. An N-cadherin antibody however was shown to be effective in aggregation inhibition at concentrations of 1 and 3 µg ml(-1). Up to 96 h no aggregation occurred. A similar effect was achieved by the N-cadherin protein, although less distinct. N-CAM blocking revealed no inhibition of aggregation. Second, results from IS corresponded to those of the aggregation assays. In these experiments neuron-neuron adhesion was also inhibited by blocking N-CAM L1 and N-cadherin. Cortical neurons were cultured in small wells containing circular 100 µm diameter gold electrodes, so small changes in cell-cell interactions in monolayers of neurons could be monitored by IS. Impedances of neuron-covered electrodes were significantly lower in the presence of the N-cadherin antibody and protein at concentrations of 1, 3 and 10 µg ml(-1), indicating a less profound binding between adjacent neurons. Results from the aggregation assays and impedance measurements demonstrate the applicability of blocking cell adhesion molecules for inhibition of cell-cell adhesion and aggregation.

Impedance sensing for monitoring neuronal coverage and comparison with microscopy.

We investigated the applicability of electric impedance sensing (IS) to monitor the coverage of adhered dissociated neuronal cells on glass substrates with embedded electrodes. IS is a sensitive method for the quantification of changes in cell morphology and cell mobility, making it suitable to study aggregation kinetics. Various sizes of electrodes were compared for the real-time recording of the impedance of adhering cells, at eight frequencies (range: 5 Hz-20 kHz). The real part of the impedance showed to be most sensitive at frequencies of 10 and 20 kHz for the two largest electrodes (7850 and 125,600 µm(2)). Compared to simultaneous microscopic evaluation of cell coverage and cell spreading, IS shows more detail.

Network bursts in cortical cultures are best simulated using pacemaker neurons and adaptive synapses.

One of the most specific and exhibited features in the electrical activity of dissociated cultured neural networks (NNs) is the phenomenon of synchronized bursts, whose profiles vary widely in shape, width and firing rate. On the way to understanding the organization and behavior of biological NNs, we reproduced those features with random connectivity network models with 5,000 neurons. While the common approach to induce bursting behavior in neuronal network models is noise injection, there is experimental evidence suggesting the existence of pacemaker-like neurons. In our simulations noise did evoke bursts, but with an unrealistically gentle rising slope. We show that a small subset of 'pacemaker' neurons can trigger bursts with a more realistic profile. We found that adding pacemaker-like neurons as well as adaptive synapses yield burst features (shape, width, and height of the main phase) in the same ranges as obtained experimentally. Finally, we demonstrate how changes in network connectivity, transmission delays, and excitatory fraction influence network burst features quantitatively.

Bifurcating microchannels as a scaffold to induce separation of regenerating neurites.

Many neural interfacing strategies, such as the sieve electrode and the cultured probe, rely on neurite growth to establish neural contact. But this growth is subject to natural fasciculation, compromising the effectiveness of these interfacing strategies by reducing potential selectivity. This in vitro study shows that the fasciculation mechanism can be manipulated by providing an appropriate microchannel scaffold to guide and influence growth, thereby achieving a high degree of selectivity. The microchannels employed have a bifurcation from a primary channel into two secondary channels. This bifurcating microstructure was able to support and promote fasciculated growth over 70% of the time for microchannels widths of 2.5, 5, 10 and 20 microm. Fasciculation is shown to be a strong force during ingrowth, with the initiation of neurite separation related to random spatial exploration. Narrower microchannels initiate separated growth better. Once separated growth starts fasciculation results in an even distribution of neurite growth across the bifurcation. The reduction from 20 microm to 10 microm wide channels also resulted in a 3-fold decrease in ingrowing neurites performing 180 degrees turns to exit the microchannel via the entrance. No neurite turning was observed for both the 5 and 2.5 microm wide channels.

Latency-related development of functional connections in cultured cortical networks.

To study plasticity, we cultured cortical networks on multielectrode arrays, enabling simultaneous recording from multiple neurons. We used conditional firing probabilities to describe functional network connections by their strength and latency. These are abstract representations of neuronal pathways and may arise from direct pathways between two neurons or from a common input. Functional connections based on direct pathways should reflect synaptic properties. Therefore, we searched for long-term potentiation (this mechanism occurs in vivo when presynaptic action potentials precede postsynaptic ones with interspike intervals up to approximately 20 ms) in vitro. To investigate if the strength of functional connections showed a similar latency-related development, we selected periods of monotonously increasing or decreasing strength. We observed increased incidence of short latencies (5-30 ms) during strengthening, whereas these rarely occurred during weakening. Furthermore, we saw an increased incidence of 40-65 ms latencies during weakening. Conversely, functional connections tended to strengthen in periods with short latency, whereas strengthening was significantly less than average during long latency. Our data suggest that functional connections contain information about synaptic connections, that conditional firing probability analysis is sensitive enough to detect it and that a substantial fraction of all functional connections is based on direct pathways.

Neural networks on chemically patterned electrode arrays: towards a cultured probe.

One type of future, improved neural interfaces is the 'cultured probe'. It is a hybrid type of neural information transducer or prosthesis, for stimulation and/or recording of neural activity. It would consist of a micro-electrode array (MEA) on a planar substrate, each electrode being covered and surrounded by a local circularly confined network ('island') of cultured neurons. The main purpose of the local networks is that they act as bio-friendly intermediates for collateral sprouts from the in vivo system, thus allowing for an effective and selective neuron electrode interface. As a secondary purpose, one may envisage future information processing applications of these intermediary networks. In this chapter, first, progress is shown on how substrates can be chemically modified to confine developing networks, cultured from dissociated rat cortex cells, to 'islands' surrounding an electrode site. Additional coating of neurophobic, polyimide coated substrate by tri-block-copolymer coating enhances neurophilic-neurophobic adhesion contrast. Secondly, results are given on neuronal activity in patterned, unconnected and connected, circular 'island' networks. For connected islands, the larger the island diameter (50, 100 or 150 microm), the more spontaneous activity is seen. Also, activity may show a very high degree of synchronization between two islands. For unconnected islands, activity may start at 22 days in vitro (DIV), which is two weeks later than in unpatterned networks.

Conditional firing probabilities in cultured neuronal networks: a stable underlying structure in widely varying spontaneous activity patterns.

To properly observe induced connectivity changes after training sessions, one needs a network model that describes individual relationships in sufficient detail to enable observation of induced changes and yet reveals some kind of stability in these relationships. We analyzed spontaneous firing activity in dissociated rat cortical networks cultured on multi-electrode arrays by means of the conditional firing probability. For all pairs (i, j) of the 60 electrodes, we calculated conditional firing probability (CFP(i,j)[tau]) as the probability of an action potential at electrode j at t = tau, given that one was detected at electrode i at t = 0. If a CFP(i,j)[tau] distribution clearly deviated from a flat one, electrodes i and j were considered to be related. For all related electrode pairs, a function was fitted to the CFP-curve to obtain parameters for 'strength' and 'delay' (i.e. maximum and latency of the maximum of the curve) of each relationship. In young cultures the set of identified relationships changed rather quickly. At 16 days in vitro (DIV) 50% of the set changed within 2 days. Beyond 25 DIV this set stabilized: during a week more than 50% of the set remained intact. Most individual relationships developed rather gradually. Moreover, beyond 25 DIV relational strength appeared quite stable, with coefficients of variation (100 x SD/mean) around 25% in periods of approximately 10 h. CFP analysis provides a robust method to describe the underlying probabilistic structure of highly varying spontaneous activity in cultured cortical networks. It may offer a suitable basis for plasticity studies, in the case of changes in the probabilistic structure. CFP analysis monitors all pairs of electrodes instead of just a selected one. Still, it is likely to describe the network in sufficient detail to detect subtle changes in individual relationships.

Longterm stability and developmental changes in spontaneous network burst firing patterns in dissociated rat cerebral cortex cell cultures on multielectrode arrays.

Spontaneous action potentials were recorded longitudinally for 4-7 weeks from dissociated rat occipital cortex cells cultured on planar multi-electrode plates, during their development from isolated neurons into synaptically connected neuronal networks. Activity typically consisted of generalized bursts lasting up to several seconds, separated by variable epochs of sporadic firing at some of the active sites. These network bursts displayed discharge patterns with age-dependent firing rate profiles, and durations significantly increasing in the 3rd week in vitro and decreasing after about 1 month in vitro, when they evolved into short events with prompt onsets. These findings indicate that after about a month in vitro these cultured neuronal networks have developed a degree of excitability that allows almost instantaneous triggering of generalized discharges. Individual neurons tend to fire in specific and persistent temporal relationships to one another within these network bursts, suggesting that network connectivity maintains a core topology during its development.

Long-term adhesion and survival of dissociated cortical neurons on miniaturised chemical patterns.

The influence of neuron-adhesive pattern geometry on long-term adhesion, survival and pattern compliance of cortical neuronal tissue was studied over a period of 15 days. The results are relevant for a successful, long-term integration of neuronal cells with electrodes from micro-electronic devices. Microwells (depth 0.5 microm), with diameters of 25, 50, 100 and 150 microm and spacing distances of 15, 30, 60 and 90 microm, were etched in a neuron-repellent fluorocarbon (FC) layer and coated with neuron-adhesive polyethylenimine (PEI). Results showed that adhesion, survival and compliance to the underlying patterns were geometry- and time-dependent. After 1 day, adhesion was inversely proportional to the diameter of the microwells, thus favouring the 25 microm microwells. However, adhesion was best on 50 microm microwells after 15 days. Survival of neurons was limited on 25 microm microwells (viability function V(D, T) was 0.08), as opposed to the better survival on 150 microm microwells (V(D, T) was 0.25) after 15 days. In summary, the study shows that the chemical patterns with microwells of 150 microm diameter (90 microm spacing gap) are most suitable for application on neuro-electronic devices owing to the better long-term survival and high pattern compliance of the neuronal cells.

Extracellular detection of active membrane currents in the neuron-electrode interface.

Although measurement of sealing resistance is an important tool in the assessment of the electrical contacts between cultured cells and substrate embedded microelectrodes, it does not offer information about the type of cell, i.e. neuron or non-neuronal cell. Also, rules for translation of a measured sealing resistance into parameters for successful stimulation, i.e. eliciting an action potential, are not available yet. Therefore, a method is proposed for the detection of active membrane currents, elicited by extracellular current stimulation. The method is based on the prediction of the linear part of the response to an applied stimulus current pulse using an impedance model of the neuron-electrode contact. Active membrane currents are detected in the nonlinear response, which is obtained by subtraction of the predicted linear response from the measured response. The required impedance model parameters are extracted from impedance spectroscopy or directly from the measured responses.

Experimental investigation on neural cell survival after dielectrophoretic trapping.

Negative dielectrophoretic forces can effectively be used to trap cortical rat neurons. The creation of dielectrophoretic forces requires electric fields of high non-uniformity. High electric field strengths, however, can cause excessive membrane potentials by which cells may unrecoverably be changed or it may lead to cell death. In a previous study it was found that cells trapped at 3 Vtt/14 MHz did not change morphologically as compared to cells that were not exposed to the electric field. This study investigates the viability of fetal cortical rat neurons after being trapped by negative dielectrophoretic forces at frequencies up to 1 MHz. A planar quadrupole micro-electrode structure was used for the creation of a non-uniform electric field. The sinusoidal input signal was varied in amplitude (3 and 5 Vtt) and frequency (10 kHz-1 MHz). The results presented in this paper show that the viability of dielectrophoretically trapped postnatal cortical rat cells was greatly frequency dependent. To preserve viability frequencies above 100 kHz (at 3 Vtt) or 1 MHz (5 Vtt) must be used.

Viability of dielectrophoretically trapped neural cortical cells in culture.

Negative dielectrophoretic trapping of neural cells is an efficient way to position neural cells on the electrode sites of planar micro-electrode arrays. The preservation of viability of the neural cells is essential for this approach. This study investigates the viability of postnatal cortical rat cells that were dielectrophoretically trapped. Morphological characteristics as well as the ratio of the number of outgrowing to the number of non-outgrowing cortical cells were used to compare the viability of trapped cells to that of non-exposed cells. The morphological characteristics include the area of the cell, representing adhesive properties, and the number and length of the processes, as a measure for functional recovery. The results presented in this paper show that the viable state of dielectrophoretically trapped postnatal cortical rat cells under the conditions used was similar to that of non-exposed cells.

Dielectrophoretic trapping of dissociated fetal cortical rat neurons.

Recording and stimulating neuronal activity at multiple sites can be realized with planar microelectrode arrays. Efficient use of such arrays requires each site to be covered by at least one neuron. By application of dielectrophoresis (DEP), neurons can be trapped onto these sites. This study investigates negative dielectrophoretic trapping of fetal cortical rat neurons. A planar quadrupole microelectrode structure was used for the creation of a nonuniform electric field. The field was varied in amplitude (1, 3, and 5 V) and frequency (10 kHz-50 MHz). Experimental results were compared with a theoretical model to investigate the yield (the number of neurons trapped in the center of the electrode structure) with respect to time, amplitude and frequency of the field. The yield was a function of time(1/3) according to theory. However, unlike the model predicted, an amplitude-dependent frequency behavior was present and unexpected peaks occurred in the DEP-spectra above 1 MHz. Gain/phase measurements showed a rather unpredictable behavior of the electrode plate above 1 MHz, and temperature measurement showed that heating of the medium influenced the trapping effect, especially for larger amplitudes and higher frequencies.

Adhesion and patterning of cortical neurons on polyethylenimine- and fluorocarbon-coated surfaces.

Adhesion and patterning of cortical neurons was investigated on isolated islands of neuron-adhesive polyethylenimine (PEI) surrounded by a neuron-repellent fluorocarbon (FC) layer. In addition, the development of fasciculated neurites between the PEI-coated areas was studied over a time period of fifteen days. The patterns consisted of PEI-coated wells (diameter 150 microns, depth 0.5 micron) which were etched in a coating of fluorocarbon (FC) on top of polyimide (PI) coated glass. The separation distance between the PEI-coated wells were varied between 10 and 90 microns. This paper shows that chemical patterns of PEI and FC result in highly compliant patterns of adhering cortical neurons after one day in vitro. Interconnecting neurite fascicles between PEI-coated wells were especially present on patterns with a separation distance of 10 microns after eight days in vitro. A significant lower number of interconnecting neurite fascicles was observed on 20 microns separated patterns. Effective isolation of neurons into PEI-coated wells was achieved on patterns with a separation distance of 80 microns as no interconnecting neurite fascicles were observed.

Endoneural selective stimulating using wire-microelectrode arrays.

In acute experiments eight 5- to 24-wire-microelectrode arrays were inserted into the common peroneal nerve of the rat, to investigate whether the electrodes could selectively stimulate motor units of the extensor digitorum longus (EDL) muscle. Twitch-force-recruitment curves were measured from the EDL for each array electrode. The curves were plotted on a double-logarithmic scale and parameterized by the low-force slope (which represents the power p in the power-law relationship of force F versus stimulus current I, or F approximately I(p)) and the threshold current. The slopes and threshold currents measured with array electrodes did not differ significantly from those obtained with randomly inserted single wire-microelectrodes. This indicates that, although involving a more invasive insertion procedure, electrode arrays provide neural contacts with low-force recruitment properties similar to those of single wires. Array results revealed partial blocking of neural conduction, similar to that reported with microneurographic insertion with single needles. The efficiency of the array was defined as the fraction of array electrodes selectively contacting a motor unit and evoking the corresponding threshold force. Efficiency thus expresses the practical value of the used electrode array in terms of the total number of distinct threshold forces that can be stimulated by selecting the appropriate electrodes. The eight arrays were capable of evoking threshold forces selectively with an average efficiency of 0.81 (or 81%).

Extracellular potentials from active myelinated fibers inside insulated and noninsulated peripheral nerve.

A model is presented that calculates the single-fiber extracellular field and action potential (ap) of an active myelinated nerve fiber placed centrally or eccentrically inside a nerve with a cylindrical geometry, representing essentially a one-fascicle nerve. This one-fascicle nerve has the dimensions and conductivities of the rat peroneal nerve branch. The results show a wide variety of wave shapes to be measured, depending on the position of the intraneural electrode with respect to the fiber axis and to the nodes of Ranvier and depending on the presence of an isolating cuff around the nerve. Action potential shapes may range from the "classical" quasi-biphasic one, to more triphasic, or even more complicated in the case of a short insulating cuff being present around the nerve. In the latter case, when measured bipolarly, ap-wave shapes become almost monophasic.