Publications

Ultrasonic neurostimulation is a potentially potent noninvasive therapy, whose mechanism has yet to be elucidated. We designed a system capable of applying ultrasound with minimal reflections to neuronal cultures. Synaptic transmission was pharmacologically controlled, eliminating network effects, enabling examination of single-cell processes. Short single pulses of low-intensity ultrasound were applied, and time-locked responses were examined using calcium imaging.
Low-pressure (0.35 MPa) ultrasound directly stimulated ∼20% of pharmacologically disconnected neurons, regardless of membrane poration. Stimulation was resistant to the blockade of several purinergic receptor and mechanosensitive ion channel types. Stimulation was blocked, however, by suppression of action potentials. Surprisingly, even extremely short (4 μs) pulses were effective, stimulating ∼8% of the neurons. Lower-pressure pulses (0.35 MPa) were less effective than higher-pressure ones (0.65 MPa). Attrition effects dominated, with no indication of compromised viability.
Our results detract from theories implicating cavitation, heating, non-transient membrane pores >1.5 nm, pre-synaptic release, or gradual effects. They implicate a post-synaptic mechanism upstream of the action potential, and narrow down the list of possible targets involved.

Automatic speech recognition (ASR) and natural language processing (NLP) are expected to benefit from an effective, simple, and reliable method to automatically parse conversational speech. The ability to parse conversational speech depends crucially on the ability to identify boundaries between prosodic phrases. This is done naturally by the human ear, yet has proved surprisingly difficult to achieve reliably and simply in an automatic manner. Efforts to date have focused on detecting phrase boundaries using a variety of linguistic and acoustic cues. We propose a method which does not require model training and utilizes two prosodic cues that are based on ASR output. Boundaries are identified using discontinuities in speech rate (pre-boundary lengthening and phrase-initial acceleration) and silent pauses. The resulting phrases preserve syntactic validity, exhibit pitch reset, and compare well with manual tagging of prosodic boundaries. Collectively, our findings support the notion of prosodic phrases that represent coherent patterns across textual and acoustic parameters.

Studies have shown that Methylphenidate (MPH) affects cognitive performance on the neuropsychological tests and clinical symptoms of individuals diagnosed with attention deficit/hyperactivity disorder (ADHD). This study investigated the acute effects of MPH on neuropsychological tests to explore the interaction between MPH and test-retest effects. Twenty youths with ADHD were tested before and after MPH intake in a double-blind placebo-controlled crossover design and compared to twenty matched controls. Participants were tested on a range of standardized tasks including sustained attention to response, N-Back, and Word/Color Stroop. Identical tasks were administered twice each testing day, before and 1 hour after MPH/Placebo administration. Healthy controls were tested similarly with no intervention. Decreases in response time (RT) variability across tasks and in commission errors were found in ADHD after MPH. Conversely, a significant increase in RT variability and increase in omission errors were observed after the placebo. In the control group, RT variability and omission errors increased whereas commission errors decreased, suggesting fatigue and practice effects, respectively. Test-retest reliability was higher in controls than ADHD. It is suggested that cognitive tests are sensitive objective measures for the assessment of responses to MPH in ADHD but are also affected by repetition and fatigue.

Methylphenidate (MPH) is the leading drug for treatment of attention deficit/hyperactivity disorder (ADHD), yet its underlying neuronal mechanisms are still unclear. Here, we use a dynamical brain networks approach to explore the effects of cognitive effort and MPH on ADHD subjects. Electroencephalography data were recorded from 19 ADHD subjects and 18 controls during a Go/No-Go Task. ADHD subjects completed the task twice a day over 2 days. The second session was administered post-ingestion of placebo/MPH (alternately). Controls performed two tasks in 1 day. The data were divided into 300 ms windows from -300 pre-stimulus until 1200 ms post-stimulus. Brain networks were constructed per subject and window, from which network metrics were extracted and compared across the experimental conditions. We identified an immediate shift of global connectivity and of network segregation after the stimulus for both groups, followed by a gradual return to baseline. Decreased global connectivity was found to be 400-700 ms post-stimulus in ADHD compared with controls, and it was normalized post-MPH. An increase of the networks' segregation occurred post-placebo at 100-400 and 400-700 ms post-stimulus, yet it was inhibited post-MPH. These global alterations resulted mainly from changes in task-relevant frontal and parietal regions. The networks of medicated ADHD subjects and controls exhibited a more significant and lasting change, relative to baseline, compared with those of nonmedicated ADHD. These results suggest impaired network flexibility in ADHD, corrected by MPH.

Background: Down syndrome remains the main genetic cause of intellectual disability, with an incidence rate of about 1 in 700 live births. The Ts65Dn mouse strain, with an extra murine chromosome that includes genes from chromosomes 10, 16 and 17 of the mouse and the Tc1 strain with an extra human chromosome 21, are currently accepted as informative and well-studied models for Down Syndrome. Using whole cell patch clamp we recently showed changes in several types of transmembrane currents in hippocampal neuronal cultures of Ts65Dn and Tc1 embryos. The associated genetic changes responsible for these changes in physiology were yet to be studied.Methods: We used qPCR to measure RNA expression level of a few of the channel genes that we suspect are implicated in the previously reported changes of measured currents, and performed statistical analysis using Matlab procedures for the standard t-test and ANOVA and for calculating correlations between the RNA expression levels of several channel genes.Results: We present differential gene expression levels measured using qPCR of the potassium channel regulators KCNE1 and KCNE2 in both Ts65Dn and Tc1 embryos and pups compared to controls. In Tc1, the human genes KCNJ6 and KCNJ15 are expressed in addition to a statistically insignificant increase of expression in the mouse genes KCNJ6 and KCNJ15. All channel genes that we have measured with large replication, have the same up-regulation or down-regulation in both mouse models, indicating that the transcription mechanism acts similarly in these two mouse models. The large dataset furthermore allows us to observe correlations between different channel genes. We find that, despite the significant changes in expression levels, channels that are known to interact have a high and significant correlation in expression both in controls and in the Down syndrome mouse model.Conclusions: We suggest the differential expression of KCNE1 and KCNE2 as a possible cause for our previously reported changes in potassium currents. We report a KCNJ6 and KCNJ15 overexpression, which plays a role in the increased input conductance and the reduced cell excitability that we previously reported in the Tc1 mouse model. The large and significant positive (KCNQ2-KCNQ3, KCNE1-KCNE2, KCNQ3-KCNE1, KCNQ2-KCNE1, KCNQ2-KCNE2, KCNQ3-KCNE2) and negative correlations (KCNE1-KCNJ15, KCNE2-KCNJ15) that we find between channel genes indicate that these genes probably work in a cooperative or in a mutually exclusive manner.

A fundamental issue in the dynamics of complex systems is the resilience of the network in response to targeted attacks. This paper explores the local dynamics of the network attack process by investigating the order of removal of the nodes that have maximal degree, and shows that this dynamic network response can be predicted from the graph's initial connectivity. We demonstrate numerically that the maximal degree M(tau) of the network at time step t decays exponentially with tau via a topology-dependent exponent. Moreover, the order in which sites are removed can be approximated by considering the network's ``hierarchy'' function h, which measures for each node V-i how many of its initial nearest neighbors have lower degree versus those that have a higher one. Finally, we show that the exponents we identified for the attack dynamics are related to the exponential behavior of spreading activation dynamics. The results suggest that the function h, which has both local and global properties, is a novel nodal measurement for network dynamics and structure.

Background: Down syndrome incidence in humans increases dramatically with maternal age. This is mainly the result of increased meiotic errors, but factors such as differences in abortion rate may play a role as well. Since the meiotic error rate increases almost exponentially after a certain age, its contribution to the overall incidence aneuploidy may mask the contribution of other processes. Results: To focus on such selection mechanisms we investigated transmission in trisomic females, using data from mouse models and from Down syndrome humans. In trisomic females the a-priori probability for trisomy is independent of meiotic errors and thus approximately constant in the early embryo. Despite this, the rate of transmission of the extra chromosome decreases with age in females of the Ts65Dn and, as we show, for the Tc1 mouse models for Down syndrome. Evaluating progeny of 73 Tc1 births and 112 Ts65Dn births from females aged 130 days to 250 days old showed that both models exhibit a 3-fold reduction of the probability to transmit the trisomy with increased maternal ageing. This is concurrent with a 2-fold reduction of litter size with maternal ageing. Furthermore, analysis of previously reported 30 births in Down syndrome women shows a similar tendency with an almost three fold reduction in the probability to have a Down syndrome child between a 20 and 30 years old Down syndrome woman. Conclusions: In the two types of mice models for Down syndrome that were used for this study, and in human Down syndrome, older females have significantly lower probability to transmit the trisomy to the offspring. Our findings, taken together with previous reports of decreased supportive environment of the older uterus, add support to the notion that an older uterus negatively selects the less fit trisomic embryos.

Down syndrome (DS) mouse models exhibit cognitive deficits, and are used for studying the neuronal basis of DS pathology. To understand the differences in the physiology of DS model neurons, we used dissociated neuronal cultures from the hippocampi of Ts65Dn and Tc1 DS mice. Imaging of [Ca²⁺]i and whole cell patch clamp recordings were used to analyze network activity and single neuron properties, respectively. We found a decrease of ~30% in both fast (A-type) and slow (delayed rectifier) outward potassium currents. Depolarization of Ts65Dn and Tc1 cells produced fewer spikes than diploid cells. Their network bursts were smaller and slower than diploids, displaying a 40% reduction in δf/f₀ of the calcium signals, and a 30% reduction in propagation velocity. Additionally, Ts65Dn and Tc1 neurons exhibited changes in the action potential shape compared to diploid neurons, with an increase in the amplitude of the action potential, a lower threshold for spiking, and a sharp decrease of about 65% in the after-hyperpolarization amplitude.Numerical simulations reproduced the DS measured phenotype by variations in the conductance of the delayed rectifier and A-type, but necessitated also changes in inward rectifying and M-type potassium channels and in the hyperpolarization-activated cyclic nucleotide-gated (HCN) channels. We therefore conducted whole cell patch clamp measurements of M-type potassium currents, which showed a ~90% decrease in Ts65Dn neurons, while HCN measurements displayed an increase of ~65% in Ts65Dn cells. Quantitative real-time PCR analysis indicates overexpression of 40% of KCNJ15, an inward rectifying potassium channel, contributing to the increased inhibition. We thus find that changes in several types of potassium channels dominate the observed DS model phenotype.

Background: Down Syndrome (DS) is the most common chromosomal abnormality in humans, with many phenotypic characteristics-mental retardation is probably the most prominent one. Embryonic hippocampal neuronal cultures from two mouse models (TC1 and Ts65Dn) for DS were compared to control littermate cultures to detect possible cellular phenotypes of DS. Network behaviour was assessed using calcium imaging, and whole cell patch clamp technique was used to measure differences in single cell properties. Results: Calcium imaging, measured with fluo4, revealed smaller amplitude, shorter duration network bursts, with no change in basal calcium level, which was measured using Fura-2. Baclofen, a GABA-B agonist, suppressed network bursts with lower concentration in DS networks than in control networks. In patch clamped neurons, the threshold for excitation, after-hyperpolarization (AHP) amplitude and input resistance were reduced, while spike height was increased in DS cells compared to controls. Following blockage of sodium currents we found that potassium currents (slow and fast) are reduced in DS cells. Network activity was different with electrophysiological tools as well: bursts were shorter, the total amount of time the network spent bursting was shorter, and network bursts were less synchronized. Most of the differences appeared in both mouse models. Conclusions: DS cells were different both in the basic cell properties and in the assembly into a network. The changes in spike properties, along with reduced potassium currents all point to changes in potassium channels. We speculate that this may be due to possible overexpression of two potassium channel regulators, KCNE1 and KCNE2, whose genes reside on chromosome 21. The altered network activity along with baclofen experiments may be caused by increased inhibition. Two types of inwardly rectifying potassium channels genes reside on chromosome 21: KCNJ6, KCNJ15, and may be the cause of this increased network inhibition.

Neurotrophins are known to promote synapse development as well as to regulate the efficacy of mature synapses. We have previously reported that in two-dimensional rat hippocampal cultures, brain-derived neurotrophic factor (BDNF) and neurotrophin-3 significantly increase the number of excitatory input connections. Here we measure the effect of these neurotrophic agents on propagating fronts that arise spontaneously in quasi-one-dimensional rat hippocampal cultures. We observe that chronic treatment with BDNF increased the velocity of the propagation front by about 30%. This change is attributed to an increase in the excitatory input connectivity. We analyze the experiment using the Feinerman-Golomb/Ermentrout-Jacobi/Moses-Osan model for the propagation of fronts in a one-dimensional neuronal network with synaptic delay and introduce the synaptic connection probability between adjacent neurons as a new parameter of the model. We conclude that BDNF increases the number of excitatory connections by favoring the probability to form connections between neurons, but without significantly modifying the range of the connections (connectivity footprint).

The neurotrophic factors brain-derived neurotrophic factor (BDNF) and neurotrophin-3 (NT-3) have been shown to promote excitatory and inhibitory synapse development. However, a quantitative analysis of their influence on connectivity has proven in general difficult to achieve. In this work we use a novel experimental approach based on percolation concepts that provides a quantification of the average number of connections per neuron. In combination with electrophysiological measurements, we characterize the changes in network connectivity induced by BDNF and NT-3 in rat hippocampal cultures. We show that, on the one hand, BDNF and NT-3 accelerate the maturation of connectivity in the network by about 17 h. On the other hand, BDNF and NT-3 increase the number of excitatory input connections by a factor of about two, but without modifying the number of inhibitory input connections. This scenario of a dominant effect on the excitation is supported by the analysis of spontaneous population bursts in cultures treated with either BDNF or NT-3, which show burst amplitudes that are insensitive to the blockade of inhibition. A leaky integrate-and-fire model reproduces the experimental results well.

Functional logical microcircuits are an essential building block of computation in the brain. However, single neuronal connections are unreliable, and it is unclear how neuronal ensembles can be constructed to achieve high response fidelity. Here, we show that reliable, mesoscale logical devices can be created in vitro by geometrical design of neural cultures. We control the connections and activity by assembling living neural networks on quasi-one-dimensional configurations. The linear geometry yields reliable transmission lines. Incorporating thin lines creates 'threshold' devices and logical 'AND gates'. Breaking the symmetry of transmission makes neuronal 'diodes'. All of these function with error rates well below that of a single connection. The von Neumann model of redundancy and error correction accounts well for all of the devices, giving a quantitative estimate for the reliability of a neuronal connection and of threshold devices. These neuronal devices may contribute to the implementation of computation in vitro and, ultimately, to its understanding in vivo.

Improvements in technique in conjunction with an evolution of the theoretical and conceptual approach to neuronal networks provide a new perspective on living neurons in culture. Organization and connectivity are being measured quantitatively along with other physical quantities such as information, and are being related to function. In this review we first discuss some of these advances, which enable elucidation of structural aspects. We then discuss two recent experimental models that yield some conceptual simplicity. A one-dimensional network enables precise quantitative comparison to analytic models, for example of propagation and information transport. A two-dimensional percolating network gives quantitative information on connectivity of cultured neurons. The physical quantities that emerge as essential characteristics of the network in vitro are propagation speeds, synaptic transmission, information creation and capacity. Potential application to neuronal devices is discussed.

Thoughts and ideas are multidimensional and often concurrent, yet they can be expressed surprisingly well sequentially by the translation into language. This reduction of dimensions occurs naturally but requires memory and necessitates the existence of correlations, e.g., in written text. However, correlations in word appearance decay quickly, while previous observations of long-range correlations using random walk approaches yield little insight on memory or on semantic context. Instead, we study combinations of words that a reader is exposed to within a "window of attention," spanning about 100 words. We define a vector space of such word combinations by looking at words that co-occur within the window of attention, and analyze its structure. Singular value decomposition of the co-occurrence matrix identifies a basis whose vectors correspond to specific topics, or "concepts" that are relevant to the text. As the reader follows a text, the "vector of attention" traces out a trajectory of directions in this "concept space." We find that memory of the direction is retained over long times, forming power-law correlations. The appearance of power laws hints at the existence of an underlying hierarchical network. Indeed, imposing a hierarchy similar to that defined by volumes, chapters, paragraphs, etc. succeeds in creating correlations in a surrogate random text that are identical to those of the original text. We conclude that hierarchical structures in text serve to create long-range correlations, and use the reader's memory in reenacting some of the multidimensionality of the thoughts being expressed.

Dissociated neurons were cultured on lines of various lengths covered with adhesive material to obtain an experimental model system of linear signal transmission. The neuronal connectivity in the linear culture is characterized, and it is demonstrated that local spiking activity is relayed by synaptic transmission along the line of neurons to develop into a large-scale population burst. Formally, this can be treated as a one-dimensional information channel. Directional propagation of both spontaneous and stimulated bursts along the line, imaged with the calcium indicator Fluo-4, revealed the existence of two different propagation velocities. Initially, a small number of neighboring neurons fire, leading to a slow, small and presumably asynchronous wave of activity. The signal then spontaneously develops to encompass much larger and further populations, and is characterized by fast propagation of high-amplitude activity, which is presumed to be synchronous. These results are well described by an existing theoretical framework for propagation based on an integrate-and-fire model.

We study the dynamic network of e-mail traffic and find that it develops self-organized coherent structures similar to those appearing in many nonlinear dynamic systems. Such structures are uncovered by a general information theoretic approach to dynamic networks based on the analysis of synchronization among trios of users. In the e-mail network, coherent structures arise from temporal correlations when users act in a synchronized manner. These temporally linked structures turn out to be functional, goal-oriented aggregates that must react in real time to changing objectives and challenges (e.g., committees at a university). In contrast, static structures turn out to be related to organizational units (e.g., departments).

A double pipette system for local, controlled drug infusion is presented. Two concentric pipettes can be manipulated separately and pressurized independently by a designated double holder. The inner pipette is loaded with the desirable solution (drug), and functions as a source, while the outer one is a sink. This gives a flow of the solution between the two pipettes that protrudes only a small distance into the surrounding fluid and does not diffuse away. Time resolution of the infusion is highly controllable, and oscillatory flow can be generated. Three implementations of the double pipette system are demonstrated. We show that local application of neurotransmitters in neuronal networks is an efficient way of stimulating activity in the network. We then present a wet micro lithography technique using topical application of proteins onto the substrate. Finally, we show that we can localize a given drug on a small targeted part of a cell.

Gradual disruption of the actin cytoskeleton induces a series of structural shape changes in cells leading to a transformation of cylindrical cell extensions into a periodic chain of 'pearls.' Quantitative measurements of the pearling instability give a square-root behavior for the wavelength as a function of drug concentration. We present a theory that explains these observations in terms of the interplay between rigidity of the submembranous actin shell and tension that is induced by boundary conditions set by adhesion points. The theory allows estimation of the rigidity and thickness of this supporting shell. The same theoretical considerations explain the shape of nonadherent edges in the general case of untreated cells.

A thin solid, burning against an oxidizing wind, develops a fingering instability. The effect is observed in a narrow gap geometry, where free convection is suppressed. Focusing on the developed nonlinear state, we find that two length scales coexist. The spacing between fingers is determined by the Péclet number, and the finger width is determined by heat losses. Dense fingers develop by tip splitting. A phenomenological model accurately predicts the fingers’ spacing, and is generally applicable to diffusion limited systems. We suggest that the effect is a new, accurately controllable, version of the thermal-diffusive instability.

We report an experimental and theoretical study of wetting front propagation in anisotropic paper. The height-height correlation function is measured in the presence of large scale biasing and is found to scale with an exponent x = 0.4. Theoretical analysis of this two-dimensional invasion predicts logarithmic scaling for isotropic paper. The observed nonzero roughness is due to correlated randomness and anisotropy in the paper.

Transmission light micrographs show that domains in spinodal decomposition are elongated into extremely long strings in steady states of a polymer solution under strong shear flow. The shear flow stabilizes the string against their intrinsic surface tension instabilities. The string diameter decreases with increase in the shear rate and ultimately becomes of the order of the interface thickness, resulting in shear-induced homogenization.

We present an experimental study of stationary convection in a binary mixture at positive values of a separation ratio. The interplay between the Rayleigh-Bénard and the Sorét mechanisms of instability and the corresponding boundary conditions gives us the possibility to observe a transition from large- to small-scale structures as well as a transition between patterns with different symmetries. We also investigate an influence of lateral boundaries and the cell geometry on the pattern selection.

We present an experimental evidence of the existence of a large-scale global circulation flow in nonlinear traveling wave states in convecting binary mixtures. We covered three of the known TW patterns: linear counter-propagating waves (CPW), spatially and temporally modulated TW and spatially modulated TW. We found that the linear CPW that occur in transient regime do not carry mass while both nonlinear states do, and the mass-transport onset coincides with the transition to nonlinear TW. Using a photochromic technique we were able to show that in the spatially modulated TW state two types of particle trajectories exist: closed one which corresponds to mass transport with TW, and open one which represents mass transfer on average backwards. Good agreement with a simple model for the particle trajectories in TW was found. In spatially and temporally modulated TW state irregular behavior in mass transport probably indicates an existence of chaotic trajectories.

We present an experimental observation of convective transport in oscillatory convection of a binary mixture. The results show the existence of mass transfer in the direction of propagation of nonlinear traveling waves (TW) as well as that of a backflow. Good agreement with a simple model for particle trajectories in TW was found. We found that the two previously observed oscillatory states differ in mass transfer. In linear counterpropagating waves no mass transfer is detected, and the mass-transport onset coincides with the transition to nonlinear TW.

Pattern observations and heat-transport measurements of convection in ethanol-water mixtures at positive values of the separation ratio are presented. Close to onset, the convective flow manifests itself in a stationary square pattern with negligible change in heat transport compared with the conductive state. Far from threshold, convection selects the usual roll structure with a strong change in heat transport. In a crossover region, the competition between square and roll patterns leads to oscillations.