Posters and Patents


Presented by Anita Bhansali at SfN 2015. Here is a link to the poster AnitaBhansaliSfN2015.pdf

Failure of Synaptic Transmission may Contribute to Seizure Propagation

 Anita Bhansali1, 3, Wim van Drongelen1, 2, Andrew K. Tryba1 
1Department of Pediatrics, Univ of Chicago
2Committee on Computational Neuroscience, Univ of Chicago
3Section of Neurosurgery, Univ of Chicago

Epilepsy is one of the most prevalent neurological diseases, affecting about 1% of the world population, but the mechanisms underlying seizure initiation, propagation and termination are not well understood. For neurons embedded in an active network, it is generally assumed that there are straightforward linear or sigmoidal relationships between net synaptic input and neuronal firing. However, due to the extreme conditions during a seizure, such a straightforward input­output relationship may vanish. For example, a paroxysmal depolarization shift (PDS) can occur, in which a neuron undergoes high­amplitude suprathreshold synaptic depolarization while it fails to generate action potentials. The PDS burst, a cellular hallmark of epilepsy, represents a neuronal activity­driven depolarization block that affects transmission within the network and may contribute to counterintuitive emergent behavior during seizures. In order to investigate the effect of PDS bursts on synaptic transmission, we performed whole­cell patch­clamp recordings of synaptically coupled rat hippocampal neurons in dissociated culture. Depolarizing current injections were used to determine the minimal amount of current required to induce action potential firing or a PDS­like burst. Current clamp (n=8) and voltage clamp measurements (n=3) revealed that PDS depolarization block led to failure of synaptic transmission in both inhibitory (n=2) and excitatory (n=8) neurons. Importantly, the amount of current needed to evoke a PDS­like burst was ~40% less in inhibitory than excitatory neurons, which suggests that inhibitory neurons enter depolarization block more readily than excitatory neurons. At the cellular level, our studies demonstrate that PDS depolarization block can create a counterintuitive relationship, such that MORE excitatory input generates LESS synaptic output. In addition, PDS bursts can lead to failure of inhibitory synaptic transmission before excitatory output is affected, which may explain the collapse of inhibitory signaling that precedes seizure propagation, an observation that supports our group’s recent modeling work (Meijer et al., 2015).

Presented by Albert Wildeman at SfN 2015. Here is a link to the poster - WildemanSfn2015.pdf

Controlled Connectivity in Neuronal Cultures

Albert Wildeman1,2, Justin E. Jureller3, Janice Wang2, Jeremy D. Marks2, Wim van Drongelen1,2
1Committee on Computational Neuroscience, University of Chicago
2Dept. of Pediatrics, University of Chicago
3Inst. of Biophysical Dynamics NanoBiology Facility, University of Chicago

The difficulty in relating macroscopic, emergent phenomena to microscopic properties is a central problem in neuroscience, both in network physiology (e.g. memory, learning) and pathology (epilepsy, schizophrenia). Network function is determined by properties that emerge from the complex, nonlinear interactions between the nodes in the network and our aim is to develop a novel approach to study the effects of these interactions with high temporal precision, spatial resolution, and spatial range. Our methodology manipulates of neuronal connectivity to explore the effects of low-level network properties on emergent macroscopic behavior. We have designed a system in which cultures of dissociated neurons grown on a multi-electrode array (MEA) can be manipulated by the introduction of real-time feedback loops mimicking additional synaptic connectivity. This is achieved by illuminating a small area of the culture and thereby releasing caged neurotransmitter, in reaction to recorded activity elsewhere in the culture. To create the flexibility required to freely add connectivity anywhere within a culture, we incorporate a digital micromirror device (DMD), an array of many small, individually controlled mirrors. The DMD allows us to dynamically select sections of the culture to be illuminated. The 1024x768 DMD grants a 2 μm resolution when covering a square 1.6 mm MEA. We employ a custom high-power LED light source and have developed an optical setup to generate and guide light to the DMD and subsequently to the MEA. As natural synaptic connectivity occurs on a millisecond timescale, we implement the feedback loops, including spike detection on the MEA electrodes, artificial connectivity and the control of the DMD, on a digital signal processor (DSP), which is a dedicated processor embedded in the MEA data acquisition system and avoids the additional delay and jitter that would be introduced by having a PC control the feedback. Preliminary experiments have illustrated that we can alter the behavior of dissociated cultures of rat hippocampal neurons by feedback alone.Through the combination of MEA, DMD and DSP technologies, we are thus able to dynamically specify network interactions with micrometer and microsecond precision and create the possibility of manipulating connectivity on this scale in order to investigate its effects on emergent network properties.

Presented by Jyothsna Suresh at SfN 2015. Here is a link to the poster - SfN_Poster_2015_JoSuresh.pdf

Functional Synaptic Connectivity During Development of Hippocampal Neuronal Networks In-vitro

Jyothsna Suresh1, 2, Anita Bhansali1, 3, Jeremy Marks1, Janice Wang1, Andrew K. Tryba1, Wim van Drongelen1, 2.
1Department of Pediatrics, Univ of Chicago
2Committee on Computational Neuroscience, Univ of Chicago
3Section of Neurosurgery, Univ of Chicago

Dissociated rat hippocampal cell cultures develop unique network activity patterns at different maturation stages in-vitro. An important component of this network behavior is determined by synaptogenesis. In a previous study, we quantified anatomical synaptogenesis by counting synapses at four critical stages of maturation – 5, 8, 14 and 20 days in vitro (DIV). We are now interested in determining the development of functional synaptic connectivity at these stages. To accomplish this, we made whole cell patch clamp recordings of pairs of connected neurons to determine the strength of synaptic connectivity. We computed spike-triggered averages (STA) of postsynaptic activity triggered by action potentials in the presynaptic neuron. We compared the area under the curve (AUC) for STAs during spontaneous and evoked synaptic activity to quantify synaptic connectivity at different maturation stages. Evoked activity was achieved by current injections leading to bursts in the presynaptic neuron. Preliminary results show that at 5 DIV, there is no spontaneous spiking as evidenced by sub-threshold membrane potential values.  However current injections in the presynaptic neuron were able to trigger spikes in the postsynaptic neuron. By 8 DIV, there is significant amount of spontaneous spiking activity and the AUC of STA during evoked activity is larger compared to that at 5 DIV. However comparison of AUC during spontaneous and evoked activity within this stage shows no clear differences. At 14 DIV, in addition to spiking and bursting activity, there are also periods of sustained depolarization, characteristic of paroxysmal depolarization shifts (PDS). There is a clear increase in the AUC during spontaneous and evoked activity compared to 8DIV. Moreover, there is a relative increase in AUC for evoked activity when compared to spontaneous activity at this maturation stage. At 20 DIV, there is a further increase in the AUC of STAs during spontaneous and evoked activity compared to 14 DIV. The relationship between these findings and emergent network activity patterns, as well as histology of maturing in-vitro cell cultures will be discussed.

Presented by Jeremy Neuman at CNS 2015 @ Prague. Here is a link to the poster - CNS Poster JeremyNeuman.pdf

Damped traveling waves and localized responses in a Wilson-Cowan network

Jeremy Neuman1, Jack D Cowan2, Wim van Drongelen3
1 Dept. of Physics, University of Chicago, Chicago, IL 60637, USA
2 Dept. of Mathematics, University of Chicago, Chicago, IL 60637, USA
3 Dept. of Pediatrics, University of Chicago, Chicago, IL 60637, USA

Spontaneous and synaptically-driven neural activity exhibit a wide variety of dynamics. In the latter case, recent experiments using spike-triggered LFPs have been able to classify stimulated behavior into two distinct categories: 1) traveling waves with smooth attenuation when the input is weak; and, 2) localized responses when the impulse is strong. Unfortunately, our knowledge of the mechanisms behind these differences is lacking on both the cellular and network scales. This study, employing the spatiotemporal mean-field Wilson-Cowan equations, provides a model for the nature of these two modes at the population level. We detect damped traveling waves with exponential decay when the input is relatively small. When the stimulus increases, the activity stays localized as evidenced by the large slope in the peaks of the activity. 

Presented by Jyothsna Suresh at Society for Neuroscience Annual Meeting 2014. Here is a link to the poster - SfN_Poster_JoSuresh.pdf


Excitatory and inhibitory synaptogenesis during development of neuronal networks in vitro

Jyothsna Suresh1,2, Janice Wang2, Vytas Bindokas3, Jeremy D. Marks2, Wim van Drongelen1,2.
1Committee on Computational Neuroscience, University of Chicago, 2Dept. of Pediatrics, University of Chicago,3Dept. of Neurobiology, University of Chicago 

Evolution of network behavior is determined both by excitatory and inhibitory synaptogenesis. Attempts have been made to count synapses at different developmental stages in cell cultures, however, little is known about individual densities of excitatory and inhibitory synapses. To quantify synaptogenesis, we counted number of excitatory and inhibitory synapses at four critical stages during development: 5, 8, 14 and 20 days in vitro (DIV).  We used immunofluorescence to label: (1) dendrites (MAP-2), (2) pre- and (3) postsynaptic puncta of excitatory synapses (vGlut/PSD-95) and inhibitory synapses (VGAT/gephyrin), (4) nuclei in cell bodies (DAPI).  High resolution (50nm x 50nm pixel) images were acquired on Leica confocal microscope and deconvolved using Huygens software to remove distortions arising from a microscope’s point spread function. Subsequently, ImageJ software was employed to obtain synaptic counts from colocalization of dendrites, pre- and postsynaptic puncta. We developed a novel method to detect these colocalizations and to obtain a noise estimate associated with the detections. First, we created binary masks of (a) dendrites by tracing the outline structure, (b) pre- and (c) postsynaptic puncta, by extracting single pixel local-maxima of fluorescence intensity and expanding them by two pixels in all directions. Overlap detected from binary AND of the three masks was counted as a colocalized synapse on dendrites. Next, in order to estimate colocalizations that occur by chance, we repeated the detection procedure after destroying the spatial correlation between pre- and postsynaptic puncta by (a) randomizing images of local-maxima, (b) shifting the original pre- and postsynaptic masks relative to each other (cross-correlation). Both these methods produced similar results for colocalizations by chance, which was used as detector baseline noise. We thus estimated excitatory and inhibitory synaptic density as number of synaptic counts per unit dendritic area, normalized by total cell count. Our preliminary results show that the density of excitatory synapses increases rapidly from 5 DIV to 14 DIV and decreases during the last developmental stage, 20 DIV. In contrast, the density of inhibitory synapses grows steadily with age and approaches the density of excitatory synapses at the last stage. Given that the typical ratio of excitatory to inhibitory populations is 80:20 in hippocampal networks, these data suggest that excitatory synaptogenesis naturally predominates during the first three weeks in vitro, while inhibitory synaptogenesis increases steadily to eventually balance the excitation towards the fourth week of maturation.

Presented by Tahra Eissa at Society for Neuroscience Annual Meeting 2014. Here is a link to the poster - SfN_Poster_TahraEissa.pdf

The Relationship between Paraxysmal Depolarizations and High Frequency Oscillations in Focal Epilepsy

Tahra Eissa1, Andrew K. Tryba2, Faiza Ben-Mabrouk2, Sean Lew3, Charles Marcuccilli4, Catherine Schevon5, Wim van Drongelen1 4

1. Committee on Neurobiology, University of Chicago, 2. Depts. Physiology and 3. Neurosurgery, Medical College of Wisconsin 4. Dept. Pediatrics, University of Chicago 5. Dept. Neurology, Columbia University

In the human neocortex, aberrant rhythmic bursts of neural activity between 80 and 150 Hz - high frequency oscillations (HFOs) – are suggested to be a hallmark of focal seizures and may aid in localization of the focus. Clinically, this activity is typically detected via EEG electrodes, which involves recording from a relatively large (cm sized) neuronal network. However it is not clear whether such a large area is required for HFO generation or what mechanisms underlie its development. We hypothesize that HFO activity during seizures is (1) generated by microscopic networks, in association with paroxysmal depolarizing shifts (PDSs) of neurons, yet (2) remains detectible in macroscopic recordings because of: (a) volume conduction (a linear process obeying the rule of superposition) and (b) synchronization of many local networks during a seizure. We determined the presence of HFOs in single cell activity (intracellular recordings) and microscopic networks (extracellular recordings) from slice recordings of human neocortex during experimental seizures, characterized by series of PDSs. Using the multi-taper spectral estimate to reduce leakage across frequency bands (Thomson, 1982), we determined that power in the 80-150 Hz band was overtly present in single cell and sub-mm network recordings. This observation supports part (1) of our hypothesis that small networks are capable of generating HFOs. Part (2) of our hypothesis was examined using two approaches: processing of the human slice data and the use of multi-electrode array (MEA) recordings from epilepsy patients. Averaging intracellular and extracellular slice measurements was used to mimic volume conduction. Two types of averages were computed: one type triggered by PDS bursts (synchronous case) and the other used randomly selected epochs (asynchronous case). We detected significantly more HFO activity in the synchronous case as compared to the asynchronous case (p<0.02), supporting our hypothesis that volume conduction of synchronized neural activity can generate compound signals that include HFOs. Using recordings obtained from MEAs implanted in patients with epilepsy, we also detected HFOs during clinical seizures. Spectral estimation of both individual microelectrodes as well as averaged activity across the array (approximating clinical macroelectrode recordings) indicate there is an increase in high gamma rhythmic activity within the focal seizure core when seizures are fully developed (Weiss et al, 2013). Our data suggests a small-network source for clinically detected HFOs and may be used for improved localization of the seizure focus.

Presented at Gordon Research Conference 2012, Mechanisms of Epilepsy & Neuronal Synchronization. Here is a link  to the poster - Poster.pdf

A Novel Technique to Evaluate Network Properties

Wim van Drongelen, Albert Wildeman, Mihailo Radojicic, Jyothsna Suresh, Janice Wang, Jeremy Marks
Department of Pediatrics, The University of Chicago, Chicago IL, USA

The electrical signature of epileptiform activity is usually easily recognized in the EEG, in which each signal represents the compound activity of millions of neurons. There is little doubt that network connectivity plays a crucial role in generating and sustaining epileptiform activity. However, the precise relationship between network connectivity and global neuronal activity, as reflected in the EEG, is poorly understood. In part this is due to the lack of techniques that allow the study of neural activity at cellular and network levels simultaneously and experimental models that allow the study of networks while perturbing the connectivity properties.

Here we present a novel approach that allows the study of developing network activity while recording neuronal activities. Networks are cultured on multi-electrode arrays (MEAs) in which the electrodes are connect to an analog-to-digital converter and a digital-to-analog converter, both connected to an interface board containing a digital signal processor. Real time recording-stimulation loops are used to mimic network connections, thereby bringing connectivity properties under full experimental control.



Drongelen W van (2001) - Medical signal monitoring and display,
- US Patent US 6,224,549 B1.
- Canadian Patent CA 2335805
- European Patent EP 1089652

Drongelen W van (2003) – Electrode disconnect system and method for medical signal monitoring, US Patent, Patent #: 6560479.