Events | Clinical Neurophysiology (CNPH)


Date	Speaker(s)	Affiliation	Subject
28 December	Christmas holidays
14 December	Nick Luiken	Master student Applied Mathematics	Inverse problems for neural field equations: challenges, limitation and solutions
30 November	Arnaud Sors	Grenoble University Hospital Centre & CEA Leti, Grenoble, France	Deep learning tools for intensive care EEG analysis
16 November	Sjoerd Verduyn Lunel	Applied Analysis, Utrecht University	Using dynamics to analyse time series
2 November	Wessel Woldman	Mathematics and Physical Sciences, University of Exeter, UK	A computational biomarker for idiopathic generalised epilepsy derived from resting state EEG
19 October	Lo Bour	Applied Analysis	New stimulation protocols for deep brain stimulation
5 October	Annika de Goede	Clinical Neurophysiology	Repeatability of paired pulse transcranial magnetic stimulation in healthy subjects
7 September	Richard van Wezel	Biomedical Signals and Systems	Visual motion processing in mice: behaviour and two-photon imaging
July & Augustus	Summer holidays
29 June	Rob van der Lubbe	Cognitive Psychology and Ergonomics	Two sides of the same coin: ERP and wavelet analyses of visual potentials evoked and induced by task-relevant faces
15 June	Barry Ruijter	Clinical Neurophysiology	A neural mean field model for EEG recovery in postanoxic encephalopathy
18 May	Koen Dijkstra	Applied Analysis	A Biophysical Model for Cytotoxic Edema
20 April	Martijn Beudel	University Medical Center Groningen	Towards Adaptive Deep Brain Stimulation in Movement Disorders
6 April	Michel van Putten and Marleen Tjepkema-Cloostermans	Clinical Neurophysiology and Medisch Spectrum Twente	Improved cerebral recovery index based on a random forest classifier for the prediction of outcome after cardiac arrest
9 March	Promotie Hanneke Berends-van Genderen	PhD defense	Fluoxetine and motor imagination to facilitate recovery after ischemic stroke
24 February	Wim van Drongelen	The University of Chicago	Multiscale Aspects of High Gamma Activity in Human Neocortex
27 January	Niels Erkamp	Master student Biomedical Engineering	Reversibility of functional connectivity loss after transient mild hypoxia of varying degree and duration
13 January	Alexander Kuck	Biomedical Engineering	Neuromodulation of the spinal circuits by tsDCS for the rehabilitation of spinal cord injury

Nick Luiken, Master student Applied Mathematics

Inverse problems for neural field equations: challenges, limitation and solutions

We address the challenges and limitations in solving inverse problems in neural field theory and present solutions. We study the inverse problem for the Amari equation. The inverse problem for the Amari equation is to determine the synaptic weight kernel given a certain neural field generated by the Amari equation. The inverse problem is ill-posed. An inverse problem is ill-posed when there is no solution, the solution is not unique, or the solution is unstable, or a combination of the three. We study the causes for the ill-posedness and present regularization methods to deal with the ill-posedness in an optimal manner. We present four examples where different challenges and solutions are addressed.

Keywords: neural fields, Amari equation, inverse problem, integro-differential equation, regularization

Arnaud Sors, Grenoble University Hospital Centre & CEA Leti, Grenoble, France

Deep learning tools for intensive care EEG analysis

EEG has been shown to be an interesting tool for a number of intensive care applications. However its potential remains partially untapped because 'human interpretation' is not always feasible long continuous records. This talk will present some of the ongoing approaches that we are developing as part of my PhD work for automated analysis of intensive care EEG using machine learning.

First I will present a simple proof of concept on the use of deep networks on raw EEG samples for a different application: supervised sleep scoring using convolutional networks. Then I will delve into the more challenging problem of how to extract meaningful representations out of multivariate time series using unsupervised approaches. I will introduce why it is interesting to be able to represent signals without the aid of supervision information, and how such representational models can serve a clinical application system. Different models and associated architectures for doing this in practice will be showcased and compared: variational autoencoders, unsupervised approaches using joint clustering and feature learning, and generative adversarial models.

Sjoerd Verduyn Lunel, Applied Analysis, Utrecht University

Using dynamics to analyse time series

We present a review of recent work to analyze time series in a robust manner using Wasserstein distances which are numerical costs of an optimal transportation problem. Given a time series, the long-term behavior of the dynamical system represented by the time series is reconstructed by Takens delay embedding method. This results in probability distributions over phase space and to each pair we then assign a numerical distance that quantifies the differences in their dynamical properties. From the totality of all these distances a low-dimensional representation in a Euclidean space is derived. This representation shows the functional relationships between the time series under study. For example, it allows to assess synchronization properties and also offers a new way of numerical bifurcation analysis. Several examples are given to illustrate our results.

Wessel Woldman, Mathematics and Physical Sciences, University of Exeter, UK

A computational biomarker for idiopathic generalised epilepsy derived from resting state EEG

Clinical diagnosis of generalised epilepsy commonly relies on the following: (1) case history (2) observation of transient abnormal activity during electroencephalography (EEG), which may not be present during clinical evaluation; and (3) if diagnostic uncertainty occurs, undertaking prolonged monitoring in an attempt to capture abnormalities, which is costly. I will discuss the discovery and validation of a biomarker based on computational analysis of a short segment of resting-state (interictal) EEG. Using leave-one-out classification on a dataset comprising 30 people with IGE and 38 normal controls, we find 100% specificity at 57% sensitivity, and 100% sensitivity at 66% specificity. We believe this biomarker could provide additional support to the diagnostic process.

Lo Bour, Applied Analysis

New stimulation protocols for deep brain stimulation

Steering DBS, adaptive DBS and coordinated reset neuromodulation the last fifteen years have been proposed as new techniques to improve the effect of DBS. Application of these techniques strongly depends on the design of the electronic devices. Why has the design of conventional pulse generators not changed up till now?

Steering DBS opens the possibility to steer the DBS current into a specific direction after implantation. A special high-density DBS lead and a multichannel programmable device is required. Adaptive DBS has the advantage of the adjustment of the stimulation amplitude depending on local field potential parameters. This requires a sensing possibility of the DBS equipment. Coordinated reset neuromodulation includes a different stimulation protocol with independent asynchronous stimulation on different contact points.

All three new techniques have been tested experimentally per-operatively or shortly after neurosurgery. However, data on chronic stimulation are lacking since appropriate implantable pulse generators are missing. But also there is still uncertainty about how to implement these new techniques. More studies are needed on the advantage of steering DBS, on which local field potential parameters are most suitable for feedback and on what is the best protocol to take optimum advantage of coordinated reset.

Very promising new techniques for DBS therapy are under development which have the potential of making DBS more efficient, leading to much smaller batteries and giving more possibilities on automatic adjustment of temporal and spatial stimulus parameters.

Annika de Goede, Clinical Neurophysiology

Repeatability of paired pulse transcranial magnetic stimulation in healthy subjects

Transcranial magnetic stimulation (TMS) is widely used to assess cortical excitability. To detect changes in excitability with longitudinal studies, it is important to validate the repeatability of excitability measures within a subject between different sessions. Repeatability studies on paired pulse TMS are limited and reported agreement ranges from poor to good. Therefore this study aims to evaluate the repeatability of paired pulse TMS in healthy subjects.

We studies thirty healthy subjects twice, approximately one week apart. Both motor cortices were stimulated at six different interstimulus intervals (ISIs) between 50 and 300 ms. Repeatability was assessed using the intraclass correlation coefficient (ICC).

We found a large variation in repeatability at the subject and ISI level, ranging from poor to good agreement. On a group level, we found a good repeatability for the averaged paired pulse TMS curves, which decreased when individual curves were correlated.

In conclusion, for a correct interpretation of the cortical excitability outcomes it is important to know the subject specific repeatability and to analyse each ISI individually.

Rob van der Lubbe, Cognitive Psychology and Ergonomics

Two sides of the same coin: ERP and wavelet analyses of visual potentials evoked and induced by task-relevant faces

New analyzing techniques of the electroencephalogram (EEG) such as wavelet analysis open the possibility to address questions that may seriously improve our understanding of the EEG, and clarify its relation with event related potentials (ERP). Three issues were addressed: 1) to what extent can early ERP components be described as transient evoked oscillations in specific frequency bands? 2) total EEG power (TP) after a stimulus consists of pre-stimulus baseline power (BP), evoked power (EP), and induced power (IP), but what are their respective contributions? 3) the Phase Reset model proposes that BP predicts EP, while the evoked model holds that BP is unrelated to EP. Which model is the most valid one? EEG results on NoGo trials for 123 individuals that took part in an experiment with emotional facial expressions were examined by computing ERPs, and by performing wavelet analyses on the raw EEG and on ERPs. After performing several multiple regression analyses we obtained the following answers. First, the P1, N1, and P2 components can by and large be described as transient oscillations in the α and θ bands. Secondly, it appears possible to estimate the separate contributions of EP, BP, and IP to TP, and importantly, the contribution of IP is mostly larger than of EP. Finally, no strong support was obtained for neither the Phase Reset nor the Evoked model. Recent models are discussed that may better explain the relation between raw EEG and ERPs.

Barry Ruijter, Clinical Neurophysiology

A neural mean field model for EEG recovery in postanoxic encephalopathy

Continuous early EEG contributes to outcome prediction in postanoxic encephalopathy. The observed EEG patterns change as a function of time and often appear in fixed sequences. We aim to increase the understanding of this evolution by means of a neural mean field model. The model comprises excitatory and inhibitory neurons, their local synaptic connections, and thalamic afferents. The effect of anoxia is modeled as an aggravated activity-dependent synaptic depression, resulting from inhibition of ATP-dependent processes. For prolonged anoxia, the network becomes hyperexcitable as a result of long term potentiation of excitatory synapses. The effect of sedative medication is modeled as an increased duration of inhibitory postsynaptic potentials. In healthy conditions, the model generates an alpha rhythm. Anoxia initially leads to an isoelectric EEG. After brief anoxia, the alpha rhythm readily reappears. After prolonged anoxia, the EEG remains isoelectric for a longer period of time, then evolves to a burst-suppression pattern, and finally shows periodic discharges. These patterns do never evolve to physiological rhythms, unless long term potentiation effects are reversible. Sedatives only lead to a transient suppression of periodic discharges. The model simulations agree well with real-world EEG observations in postanoxic encephalopathy. Our findings indicate that these EEG abnormalities possibly result from activity-dependent synaptic depression and long term potentiation of excitatory neurotransmission. Assuming the latter effect to be irreversible, the model also explains why periodic discharges in postanoxic encephalopathy are often refractory to treatment.

Koen Dijkstra, Applied Analysis

A Biophysical Model for Cytotoxic Edema

We present a dynamical biophysical model to explain cytotoxic edema in conditions of low energy supply, as observed in cerebral ischemia. Our model contains Hodgkin-Huxley type ion currents, a recently discovered voltage-gated chloride flux through the ion exchanger SLC26A11, KCl cotransporters and ATP-dependent pumps. Based on electroneutrality in bulk solution and electrostatic and diffusive forces, we derive an expression for the osmotic pressure in Gibbs- Donnan equilibrium. Under simulated conditions of low energy supply, the model predicts the reduction of ion gradients and amount of cell swelling, including realistic time courses of reaching equilibrium states. We theoretically substantiate experimental observations of chloride influx generating cytotoxic edema, while the entry of sodium alone does not. We further show that a tipping point exists, where cell volume rapidly increases as a function of reduced activity of the ATP-dependent Na+/K+ pump, and that a depolarized, Gibbs-Donnan-like equilibrium state remains stable when the pump strength is returned to physiological levels. Finally, we demonstrate that this pathological state disappears when voltage-gated sodium channels are temporarily blocked, suggesting a possible therapeutic strategy to reduce cerebral edema after cerebral ischemia.

Martijn Beudel, University Medical Center Groningen

Towards Adaptive Deep Brain Stimulation in Movement Disorders

Movement Disorders (MDs) are neurological disorders characterised by a paucity of movements, involuntary movements or a combination of both. The most well-known MDs are Parkinson’s disease (PD) and essential tremor (ET) in which respectively a paucity of movements and involuntary (shaking) movements dominate. Most MDs can initially by treated medically but in their more progressed states medication often fails. In these disease stages advanced therapies, like deep brain stimulation (DBS), can become the treatment of choice. With DBS small electrical pulses are applied to deep brain nuclei. This is established by means of a stereotactic neurosurgical procedure by which one or more electrodes are implanted and connected to an internal pulse generator (IPG), usually placed under the clavicle. Although the exact working mechanism of DBS is not established yet, DBS has been successfully applied in PD, ET and dystonia for over 20 years. Nevertheless, there are still limitations in terms of efficacy, side-effects and efficiency. One of the most important reasons for this is that DBS not only influences pathological but also physiological neural activity. To overcome this, DBS might work better were it only to stimulate where and when necessary. In the last two decades, the knowledge of the neurophysiological basis of MDs has increased dramatically. One of the most important discoveries was that the (medication induced reduction of) the power of beta (13-30 Hz) oscillations in the subthalamic nucleus (STN) correlates with the contralateral stiff-and slowness. In very recent studies, this beta power has been used as a biomarker for titrating DBS, so called adaptive DBS (aDBS). At present, potential biomarkers for applying aDBS are also being developed for other MDs. Next to this, many companies are currently developing implantable IPG’s with closed-loop properties. With these simultaneous developments, DBS might be applied in a more intelligent way in the nearby future. This would not only lead to more efficacy, less side-effects and more efficiency but would also make it possible to implant DBS systems in patients that could not tolerate conventional (continuous) DBS due to side-effects. Finally, by automising DBS programming by means of adaptive algorithms DBS can act in synergy with concurrent medication and less hospital visits for titrating the stimulation parameters would be necessary. These potential advantages would not result in an increased quality of life for the patient but could also lead to a substantial cost reduction.

Wim van Drongelen, The University of Chicago

Multiscale Aspects of High Gamma Activity in Human Neocortex

High gamma (HG, 80-150Hz) activity in macroscopic clinical records is considered a marker for critical brain regions involved in seizure initiation and is correlated with pathological multiunit firing during neocortical seizures (Weiss et al., Brain 2013). However, the effects of the seizure’s spatiotemporal dynamics on HG power generation are not well understood.

We studied HG generation and propagation, using a three-step, multi-scale signal analysis and modeling approach. First, we analyzed concurrent neuronal and microscopic network HG activity in neocortical slices from seven intractable epilepsy patients. We found HG activity in these small networks, especially when single neurons displayed paroxysmal depolarization shifts (PDSs). Second, we examined HG activity acquired with microelectrode arrays (MEAs) recorded during human seizures (n=8). We confirmed the presence of synchronized HG power across microelectrode records and HG activity at the macroscale, both specifically associated with the seizure’s core region, i.e. the area that showed multiunit spiking correlated with the seizure activity. Third, we used volume conduction based modeling to relate HG activity and network synchrony at different network scales. We show that local HG oscillations require high levels of synchrony to cross spatial scales and that this requirement is met at the microscopic scale, but not within macroscopic networks. Instead, we present evidence that HG power at the macroscale may result from harmonics of ongoing seizure activity.

We conclude that ictal HG power can be a marker for the seizure core, but generating mechanisms can differ across spatial scales.

Niels Erkamp, Master student Biomedical Engineering

Reversibility of functional connectivity loss after transient mild hypoxia of varying degree and duration

Stroke is very common in western countries, and can lead to cognitive impairment or death. The absence of oxygen in the infarct core quickly leads to cell death. Remaining but limited perfusion in the region around the core (the penumbra) causes energy depletion in neurons, and disrupts synaptic transmission and thus connectivity. Connectivity loss probably correlates with loss of cognition and restoration of connectivity seems crucial for possible treatment. However, in vivo, connectivity is difficult to assess due to limited access to the neurons and restricted experimental freedom.

We studied the effects of hypoxia on functional connectivity in cultured neuronal networks on a multi-electrode array under varying degrees and durations of hypoxia. Connectivity was estimated with conditional firing probability analysis. Several electrodes were stimulated throughout the experiment to obtain a measure for synaptic functioning.

The first six hours of hypoxia resulted in reduced activity and connectivity, followed by restoration of activity and functional connectivity to baseline values upon re-oxygenation. Under persisting hypoxic conditions, after 6 hours some recovery occurred as observed by increased activity and connectivity, until approximately 24 hours. Cultures that survived 24 hours of severe hypoxia, showed permanently decreased activity and reduced functional connectivity. If severe hypoxia was maintained even longer, both activity and connectivity further dropped, and were irreversibly lost. In cultures that survived 48 hours of hypoxia, mean connectivity strength tended to remain around baseline values, though several individual connections were lost in the process. We found no significant difference between the baseline connectivity strength of surviving and lost connections, suggesting that protective mechanisms do not prefer weaker or stronger connections. The formation of new connections during hypoxia occurred at comparable rates as in normoxic control experiments, and suggests that the mechanism for recovery through new connections are limited.

Alexander Kuck, Biomedical Engineering

Neuromodulation of the spinal circuits by tsDCS for the rehabilitation of spinal cord injury

Spinal Cord Injury (SCI) is a severe injury to the pathways of the central nervous system (CNS). Despite a heavy post-injury physical rehabilitation regime, SCI patients are often bound to a wheelchair or left with other impairments diminishing their quality of life. Trans-spinal direct current stimulation (tsDCS) is a promising new technique for the treatment of SCI. During tsDCS a small direct current is applied to the spinal cord via two or more stimulation electrodes, placed on the back of the subject. The technique thereby aims to alter the response of the neural pathways in the spinal cord, which is hypothesized to have a positive effect on the recovery of the damaged spinal cord neurons. In previous studies, it has been shown that tsDCS is able to induce a polarity-dependent modulation of reflex and motor unit behavior as well as altering ascending proprioceptive information and associative plasticity effects on a corticospinal level. Current work in our laboratory focuses on further understanding and optimizing the use of tsDCS for the rehabilitation of SCI. We will therefore discuss the current developments in the field, the work done in our laboratory and potential future directions of the application of tsDCS to spinal cord injury rehabilitation.

Date

Speaker(s)

Affiliation

Subject