• Home
• Preliminary Program
• Invited Speakers
• Speaker/Poster Information
• Conference Topics
• Programme Information
• Research Market/Beer & Wine Party
• Awards
• Company Market
• Important Dates
• Registration and Venue
• Authors Instructions
• Organizing Board
• Contact

13:00   Imaging EEG 13:00 15 mins NONLINEAR CONNECTIVITY IN SENSORIMOTOR CONTROL Yuan Yang, Teodoro Solis-Escalante, Jun Yao, Frans van der Helm, Alfred Schouten Abstract: Recent studies have indicated that the sensorimotor system is nonlinear. However, many previous studies investigated the sensorimotor system using the corticomuscular coherence (CMC), a method to measure the linear relation between brain and muscle signals. CMC can neither detect the potential nonlinear behaviour of the system nor distinguish the directionality, i.e. the effects of afferent and efferent pathways. To quantify the nonlinear connectivity in the sensorimotor control loop, we introduced a nonlinear phase coherence measure. This novel measure can access the high order nonlinear cross-frequency phase coupling, such as triple coupling (2f1-f2). Simulations showed that our method can effectively detected different order nonlinearity and assess the directional dependency between signals. Human data were obtained from a motor control study, where health subjects (n=11) maintained an isotonic wrist flexion, while receiving continuous position perturbations according to a periodic multisine signal. In this study, 126-channel electroencephalography (EEG) and two channel bipolar electromyography (EMG) were recorded. The nonlinear coherence was computed between the EEG and EMG to study nonlinear interactions in the afferent and efferent pathways of sensorimotor loops. We demonstrated that 1) our method can separately quantify nonlinear effects in the efferent (EEG → EMG) and afferent (EMG → EEG) pathways; and 2) a stronger cross-frequency coupling is detected in the afferent pathway in comparison with the efferent pathway in the corticospinal tract, with more significant values and higher magnitudes of nonlinear coherence. These results suggested our method is a useful tool for probing the nonlinear interactions in sensorimotor systems. 13:15 15 mins NETWORK ANALYSIS OF EEG RELATED FUNCTIONAL MRI CHANGES DUE TO MEDICATION WITHDRAWAL IN FOCAL EPILEPSY Kees Hermans, Pauly Ossenblok, Albert Colon, Petra van Houdt, Rudolf Verdaasdonk, Paul Boon, Jan de Munck Abstract: ABSTRACT EEG correlated functional MRI (EEG-fMRI) studies revealed substantial predictive value compared to invasive EEG and surgical outcome in patients with refractory focal epilepsy [1]. In this study we applied network analysis strategies to investigate EEG related fMRI changes due to withdrawal of anti-epileptic drugs (AEDs) to elaborate the clinical usefulness for epilepsy surgery candidates. EEG-fMRI was acquired for 10 patients with focal epilepsy who were candidates for epilepsy surgery. Patients were scanned before the start of standard pre-surgical video-EEG when still on the usual medication (condition A) and at the end of the video-EEG session, after anti-epileptic drug (AED) withdrawal (condition B). EEG-fMRI data were analyzed using both the general linear model (GLM) approach and independent component analysis (ICA). The EEG-fMRI correlation pattern obtained in condition B was used for spatial correlation with all the independent components to select the epileptic independent component (ICE). A similar procedure was used to select the well-known resting state components or networks (RSNs). Next, the difference in functional connectivity between conditions A and B was quantified using a GLM approach applied to the concatenated time series of the selected components. We were able to identify an ICE for both conditions A and B, even if the analysis of EEG-fMRI was inconclusive. Standard EEG-fMRI analysis was inconclusive in condition A for 7 out of the 10 patients studied. Spatially, these ICEs were similar for both conditions, before and after withdrawal of AEDs. Furthermore, significant increase in general functional connectivity of all selected resting state components was found for each patient after medication withdrawal. The results of this study suggest that sensitivity of EEG-fMRI as diagnostic tool is improved at the end of the video-EEG monitoring session. Our results indicate that the underlying reason is related to the increased excitability of the brain which is reflected in a general increased functional connectivity after withdrawal of AEDs. REFERENCES [1] P.J. van Houdt, P.P.W. Ossenblok, A.J. Colon, P.A.J.M. Boon, J.C. de Munck, “A framework to integrate EEG-correlated fMRI and intracerebral recordings” NeuroImage, Vol. 60, pp. 2042-53 (2012) 13:30 15 mins SEIZURE DETECTION IN ADULTS USING FEATURE BASELINE CORRECTION ON A NEONATAL EEG TRAINED CLASSIFIER Guy Bogaarts, Erik Gommer, Vivianne van Kranen-Mastenbroek, Danny Hilkman, Jos Reulen Abstract: Rationale Recently a feature baseline correction technique (FBC) has been introduced for neonatal seizure detection in the EEG using a support vector machine (SVM). It has been shown that a SVM classifier developed for neonatal seizure detection can be used without alteration for the detection of seizures in adults 1. We examined whether FBC could also improve epileptic seizure detection in adult patients using a SVM classifier trained on neonatal EEG. Furthermore a classifier trained on adult EEG is evaluated on neonatal patients. Methods The dataset consists of 57 neonatal and 21 adult standard EEG registrations (± 20 minutes). Using patient leave-one-out cross validation the performance of both neonatal and adult classifiers are tested. Performance is evaluated in terms of area under the ROC curve, percentage of seizures detected and amount of false detections per hour. Results FBC improves seizure detection performance on both adult and neonatal detection for both classifiers. The neonatal classifier outperforms the adult classifier on both patient groups. With FBC mean ROC curve areas are increased from 0.87 to 0.95 and 0.81 to 0.91 for the adult and neonatal patients respectively. Approximately 90% of the seizures in the adult patients and 80% in the neonatal patients are detected at a cost of ~1.5 false detections per hour. Conclusions Adult seizure detection performance significantly improves using FBC. Training with neonatal data results in higher performance compared to training with adult data. This might be due to the wider range of seizure morphologies in neonates. The lower amount of detected seizures also indicates that seizure detection is more difficult in neonates than in adults. We concluded that FBC is a useful alteration of the neonatal SVM classifier for detection of seizures in both adult and neonatal EEG. 1. Faul, S., A. Temko and W. Marnane Age-independent seizure detection. Conf. Proc. IEEE Eng. Med. Biol. Soc. 5:5332553, 2009. 13:45 15 mins OPTIMIZED CURRENT INJECTION SCHEME FOR TAILOR-MADE EEG FORWARD MODELS Juhani Dabek, Konstantina Kalogianni, Edwin Rotgans, Andreas Daffertshofer, Gert Kwakkel, Erwin van Wegen, Frans van der Helm, Jan de Munck Abstract: Electroencephalography (EEG) is a millisecond-resolution method for functional brain imaging with important clinical and research applications. Multichannel EEG can be used for source localisation, if additional information such as the geometry of the head and electrical conductivity of different tissues is available. The main limiting factor of its accuracy is the uncertainty in the conductivity parameters. Following Refs. [1] and [2], the possibility to determine conductivities in vivo is examined in this study, by systematically injecting small currents (around 1 mA) into pairs of EEG electrodes and fitting a model to the resulting potential distribution. Furthermore, the optimal measurement design is developed such that this technique can be applied routinely within an acceptable measurement time and an acceptable number of current injection pairs. In the conductivity model that is fit to the data, the geometry is assumed known and consists of three compartments: brain, skull and skin. It is assumed that σbrain = σskin, a standard model for EEG source localisation. Electrical current feeding simulations are performed wherein the optimal current injection pattern is sought, accounting for the following three noise sources: background EEG, imbalance in current feeding (injection–ground and ground–extraction), and deviation from the assumption that σbrain = σskin. For simplicity, a spherical head model is used, limiting the degree of freedom to the angular distance between injection electrodes [1]. The results indicate that the effect of EEG noise is minimized with current injection distant from extraction. As a slight enhancement, the injection–ground and ground–extraction distances could be kept smaller. Large deviations from σbrain = σskin, say σbrain = 0.5×σskin, result in about 8% underestimation in σskin and 40% in σskull. In conclusion, these results suggest that it would be advantageous to position the injection and extraction electrodes far away from each other and the ground electrode in between. The next step is to use realistic head shapes and the boundary element method (BEM) to pave the way towards EEG with patient-specific head geometry and conductivity information. As an advantage to BEM, the same model can be used for both current feeding and EEG [2]. The present and future research will give an answer to what the optimal measurement procedure is regarding the EIT/EEG electrode placement, current feeding and measurement time. REFERENCES [1] S Gonçalves, JC de Munck, RM Heethaar, FH Lopes da Silva, BW van Dijk, “The application of electrical impedance tomography to reduce systematic errors in the EEG inverse problem—a simulation study”, Physiol Meas, Vol 21: 379–393 (2000). [2] M Clerc, G Adde, J Kybic, T Papadopoulo, J-M Badier, “In vivo conductivity estimation with symmetric boundary elements”, Int J Bioelectromag, Vol 7(1): 1–4 (2005). 14:00 15 mins SOMATOTOPIC ARRANGEMENT OF HUMAN FINGERS EXPLORED BY HIGH DENSITY EEG Konstantina Kalogianni, Teodoro Solis-Escalante, Yuan Yang, Juhani Dabek, Alfred Schouten, Frans van der Helm, Jan de Munck Abstract: Studying the fine spatial structure of fingers’ representation on the somatosensory cortex is helpful in revealing reorganization of hand representation after stroke[1]. For that purpose commonly EEG responses to mechanical stimulation of the fingers are spatially analyzed and mapped to the cortical surface using source localization techniques. Here we explore the use of electrical stimulation to each fingertip because this allows for precise timing of the response and along with high density EEG, precise estimation of the underlying sources [2]. High-density 256-channel EEG (TMSi, The Netherlands) was recorded from ten healthy right-handed subjects. Electrical stimulation (monophasic electrical pulses with randomly 333 or 250 ms interval) was applied in 2 blocks of 500 pulses to all fingers of the right hand at random order. The intensity of the stimulus was set as twice the subject’s sensation level. Offline processing of the data included: correction of the stimulation artefact, band pass filtering , notch filtering (50Hz), and rejection of the noisy channels and trials contaminated with artefacts by visual inspection. Somatosensory evoked potentials (SEPs) were obtained after stimulus baseline correction by averaging across the remaining trials. SEP responses had small amplitudes, typically in the order of 1 V. In all subjects and all fingers for which data was analyzed a peak was found between 30 and 100 ms after stimulus (indicated as P50 component) with a contra lateral dipolar maximum. The orientation of this dipolar pattern roughly shows a systematic change of orientation of the representation of every finger. These results suggest that 256-channel EEG is sensitive enough to identify the differences between the electrically stimulated fingers. We expect to reach high accuracy (2-3 mm) of the underlying SEP sources with the use of individual MRI when taking into account the spatial-temporal covariance of the background EEG noise [3]. Therefore, the electrical finger stimulation seems a suitable tool to study subtle changes in cortical finger representation. 14:15 15 mins A COMPARISON OF SWLDA AND CPD BASED CLASSIFICATION OF AUDITORY MOBILE EEG DATA Rob Zink, Maarten De Vos, Borbála Hunyadi, Sabine Van Huffel Abstract: To date, most of the studies using EEG are still conducted in an artificial setting. These recordings lack the ability to investigate the full dynamics of our brains in real-life. Although mobile EEG recordings became feasible very recently, many challenges still have to be solved before mobile EEG can be used at home [1]. Data-driven estimations of signal and noise might be beneficial for interpreting mobile ERP data since they are able to deal with unknown types of noise. Here we present preliminary results of such an interpretable approach for classifying single trial ERP data based on Canonical Polyadic Decomposition (CPD).