- Email: [email protected]
P24-9 Extended seizure detection algorithm for intracranial EEG recordings
S246 P24-7 Pharmaco-resistance temporal epilepsy. An analysis in 33 neurosurgical patients B. Alemany1 Clinical Neurophysiology Department, Gregorio Maranon Hospital, Madrid, Spain
Posters and perform adequate signal preprocessing such as linear detrending of each subsegment. Support: CInAPCe-FAPESP.
1
Objective: Analysis of clinical manifestations, critical, intercritical and intraoperative EEG monitoring in temporal epilepsy’s patients with medical treatment resistance successfully controlled surgically. Methods: Retrospective analysis, 33 patients. In statistical average terms, the age was 36 years, with 9 monthly seizures, 22 years of disease’s course and 3 antiepileptic polytherapy treatment. The average of days of revenue was 6.42 days, on video EEG Unit Care for Intensive Electroencephalography Monitoring. The record fulfilled by means of surface electrodes placed according to 10 20 International System with additional electrodes based on 10-10 System adapted ones for record of the temporal basal region and sphenoidal ones. In 12 patients, subdural electrodes were necessary. Results: On critical clinical analysis, 158 seizures were recorded of which, the most frequent type was focal complex. Speaking on percentages, 82 of patients presented aura (36.1 gastric). The level of conscienceaccents disorder appeared in almost 100 (94.3), being more precocious than the automatisms in the majority of the sample. For automatisms, the most frequent and precocious was the oromandibular. On having analyzed our information we meet a low percentage of distonyc contralateral position that it might turn explained by the criterion used to consider as such to the sign. On bioelectrical analysis: Intercritical: epileptiform activity 87.9. Sharp wave 93.1. Increase with sleepness 97. Polyspikes in dream 21.2. Bilateral activity 21.2. Critical activity: more precocious EEG activity 69.1. Type beginning: levelling 46, rhythmic slow activity 41.7. Focal 48.9 (posterior lateral temporary 16.5, frontotemporal 15.1). Conclusions: Detailed knowledge of clinical and electrical manifestations of temporal lobe epilepsies would allow a precocious diagnosis and a reduction on accessing time of these patients to surgical treatment in case of farmacological resistence. P24-8 Connectivity characterization of mesial temporal epileptic seizures using generalized partial directed coherence with asymptotic statistics
P24-9 Extended seizure detection algorithm for intracranial EEG recordings T.W. Kjaer1 , L.S. Remvig3 , J. Henriksen1,2,3 , C.E. Thomsen4 , H.B.D. Sorensen2 1 Department of Clinical Neurophysiology, Rigshospitalet University Hospital, Copenhagen, Denmark, 2 Denmarks Technical University, Elektro, Building 349, Orsteds Plads, DK-2800 Kgs. Lyngby, Denmark, 3 Hypo-Safe, Diplomvej 381, DK-2800 Kgs. Lyngby, Denmark, 4 University of Copenhagen, Norre Alle 20, DK-2200 Copenhagen N, Denmark Objective: We implemented and tested an existing seizure detection algorithm for scalp EEG (sEEG) with the purpose of improving it to intracranial EEG (iEEG) recordings. Method: iEEG was obtained from 16 patients with focal epilepsy undergoing work up for resective epilepsy surgery. Each patient had 4 or 5 recorded seizures and 24 hours of non-ictal data were used for evaluation. Data from three electrodes placed at the ictal focus were used for the analysis. A wavelet based feature extraction algorithm delivered input to a support vector machine (SVM) classifier for distinction between ictal and non-ictal iEEG. We compare our results to a method published by Shoeb in 2004. While the original method on sEEG was optimal with the use of only four subbands in the wavelet analysis, we found that better seizure detection could be made if all subbands were used for iEEG. Results: When using the original implementation a sensitivity of 92.8% and a false positive ratio (FPR) of 0.93/h were obtained. Our extension of the algorithm rendered a 95.9% sensitivity and only 0.65 false detections per hour. Conclusion: Better seizure detection can be performed when the higher frequencies in the iEEG were included in the feature extraction. Our future work will concentrate on development of a method for identification of the most prominent nodes in the wavelet packets analysis for optimization of an automatic seizure detection algorithm. P24-10 On emotional effects of odors of squeezed organic kale leaf based on EEGs and heart rate variability
K. Sameshima1 , C.S.N. de Brito1,2 , D.Y. Takahashi3 , C.L. Jorge4 , L.H. Castro4 , L.A. Baccala5 1 Department of Radiology, Faculdade de Medicina, University of S˜ ao Paulo, S˜ ao Paulo, Brazil, 2 Programa de Pos-graduacao em Bioinformatica, University of S˜ ao Paulo, S˜ ao Paulo, SP, Brazil, 3 Instituto de Matematica e Estatistica, University of S˜ ao Paulo, S˜ ao Paulo, SP, Brazil, 4 Epilepsy Service, Hospital das Clinicas, Faculdade de Medicina, University of S˜ ao Paulo, S˜ ao Paulo, SP, Brazil, 5 Department of Telecommunications and Control, Escola Politecnica, University of S˜ ao Paulo, S˜ ao Paulo, SP, Brazil
Y. Okita1 , H. Nakamura2 , I. Takahashi1 , T. Takaoka3 , K. Kouda4 , M. Kimura5 , T. Kobayashi6 , T. Sugiura7 1 Graduate School of Science and Technology, Shizuoka University, Japan, 2 Graduate School of Human Development and Environment, Kobe University, Japan, 3 Enseki Aojiru Co. Ltd, Japan, 4 Department of Public Health, Kinki University School of Medicine, Japan, 5 Faculty of Engineering, Shizuoka University, Japan, 6 Graduate School of Engineering, Kyoto University, Japan, 7 Research Institute of Electronics, Shizuoka University, Japan
Objective: Our aim is to apply the generalized partial directed coherence (gPDC) for multichannel time series recorded from a standard 10 20 EEG system electrodes during the episodes of mesial temporal epileptic seizure based on rigorous statistical asymptotic properties of the gPDC estimator. Methods: PDC is the representation of Granger causality in the frequency domain, which allows looking at frequency-wise interaction between brain areas. We introduced an adaptation to PDC definition, which can better deal with those cases of large power or prediction error disparity among the multivariate channels, called gPDC, for which rigorous asymptotic confidence intervals and statistical tests for the null hypothesis of zero gPDC using multivariate autoregressive modeling were derived. gPDC analysis routines were implemented in Python 2.5 (www.python.org) using NumPy, SciPy and Matplotlib packages. We sought to study mesial temporal epileptic patient’s EEG data. The midline derivations (Fz, Cz, Pz and Oz) were not considered in the analyses. Results: Confirming our previous work, the graph connectivity representations showed variable degree of laterality, an asymmetry of directed connections distribution toward the focus site hemisphere. Conclusions: The gPDC showed to be a robust measure of connectivity between time series in the frequency domain and its asymptotic statistics. Great care is necessary to exclude signal epochs with artifacts,
Objective: It is well known that odor influences human emotion and the status of the mind. The odor is also the most important sensation confirming the safety of food. Recently, it is reported that smelling of specific odors, such as green leaf and lavender, etc., has a relaxing effect on humans. Emotion researches which examine the effects of odors often use electroencephalogram (EEG) or heart rate variability (HRV). However, there are few reports which combined EEG and HRV in the evaluation of the odors. In this study, we measured EEGs at the frontal regions together with heart rate (HR) to investigate emotional effects of odor of squeezed vegetable leaf (Organic Kale). Methods: EEGs and HRV before and during odor presentation were analyzed by the Tone-Entropy method and a quantitative evaluation method of emotion based on the frontal alpha wave fluctuations (Yoshida Method). Eight non-smoker university students (8 men, mean age 21.4±0.5 years) with normal sense of smell participated in the experiments. EEGs were recorded at 19 electrode positions (10 20 system) before and during inhalation of the vegetable odor and ECGs were recorded according to CM5 induction. Results: Tone estimated from the heart rates has varied widely during the odor presentation compared with that before presentation. In contrast, the entropy before and during odor presentation showed no change. Furthermore, the state of mind estimated by the Yoshida method showed rapid changes before and during odor presentations.
Posters and perform adequate signal preprocessing such as linear detrending of each subsegment. Support: CInAPCe-FAPESP.
1
Objective: Analysis of clinical manifestations, critical, intercritical and intraoperative EEG monitoring in temporal epilepsy’s patients with medical treatment resistance successfully controlled surgically. Methods: Retrospective analysis, 33 patients. In statistical average terms, the age was 36 years, with 9 monthly seizures, 22 years of disease’s course and 3 antiepileptic polytherapy treatment. The average of days of revenue was 6.42 days, on video EEG Unit Care for Intensive Electroencephalography Monitoring. The record fulfilled by means of surface electrodes placed according to 10 20 International System with additional electrodes based on 10-10 System adapted ones for record of the temporal basal region and sphenoidal ones. In 12 patients, subdural electrodes were necessary. Results: On critical clinical analysis, 158 seizures were recorded of which, the most frequent type was focal complex. Speaking on percentages, 82 of patients presented aura (36.1 gastric). The level of conscienceaccents disorder appeared in almost 100 (94.3), being more precocious than the automatisms in the majority of the sample. For automatisms, the most frequent and precocious was the oromandibular. On having analyzed our information we meet a low percentage of distonyc contralateral position that it might turn explained by the criterion used to consider as such to the sign. On bioelectrical analysis: Intercritical: epileptiform activity 87.9. Sharp wave 93.1. Increase with sleepness 97. Polyspikes in dream 21.2. Bilateral activity 21.2. Critical activity: more precocious EEG activity 69.1. Type beginning: levelling 46, rhythmic slow activity 41.7. Focal 48.9 (posterior lateral temporary 16.5, frontotemporal 15.1). Conclusions: Detailed knowledge of clinical and electrical manifestations of temporal lobe epilepsies would allow a precocious diagnosis and a reduction on accessing time of these patients to surgical treatment in case of farmacological resistence. P24-8 Connectivity characterization of mesial temporal epileptic seizures using generalized partial directed coherence with asymptotic statistics
P24-9 Extended seizure detection algorithm for intracranial EEG recordings T.W. Kjaer1 , L.S. Remvig3 , J. Henriksen1,2,3 , C.E. Thomsen4 , H.B.D. Sorensen2 1 Department of Clinical Neurophysiology, Rigshospitalet University Hospital, Copenhagen, Denmark, 2 Denmarks Technical University, Elektro, Building 349, Orsteds Plads, DK-2800 Kgs. Lyngby, Denmark, 3 Hypo-Safe, Diplomvej 381, DK-2800 Kgs. Lyngby, Denmark, 4 University of Copenhagen, Norre Alle 20, DK-2200 Copenhagen N, Denmark Objective: We implemented and tested an existing seizure detection algorithm for scalp EEG (sEEG) with the purpose of improving it to intracranial EEG (iEEG) recordings. Method: iEEG was obtained from 16 patients with focal epilepsy undergoing work up for resective epilepsy surgery. Each patient had 4 or 5 recorded seizures and 24 hours of non-ictal data were used for evaluation. Data from three electrodes placed at the ictal focus were used for the analysis. A wavelet based feature extraction algorithm delivered input to a support vector machine (SVM) classifier for distinction between ictal and non-ictal iEEG. We compare our results to a method published by Shoeb in 2004. While the original method on sEEG was optimal with the use of only four subbands in the wavelet analysis, we found that better seizure detection could be made if all subbands were used for iEEG. Results: When using the original implementation a sensitivity of 92.8% and a false positive ratio (FPR) of 0.93/h were obtained. Our extension of the algorithm rendered a 95.9% sensitivity and only 0.65 false detections per hour. Conclusion: Better seizure detection can be performed when the higher frequencies in the iEEG were included in the feature extraction. Our future work will concentrate on development of a method for identification of the most prominent nodes in the wavelet packets analysis for optimization of an automatic seizure detection algorithm. P24-10 On emotional effects of odors of squeezed organic kale leaf based on EEGs and heart rate variability
K. Sameshima1 , C.S.N. de Brito1,2 , D.Y. Takahashi3 , C.L. Jorge4 , L.H. Castro4 , L.A. Baccala5 1 Department of Radiology, Faculdade de Medicina, University of S˜ ao Paulo, S˜ ao Paulo, Brazil, 2 Programa de Pos-graduacao em Bioinformatica, University of S˜ ao Paulo, S˜ ao Paulo, SP, Brazil, 3 Instituto de Matematica e Estatistica, University of S˜ ao Paulo, S˜ ao Paulo, SP, Brazil, 4 Epilepsy Service, Hospital das Clinicas, Faculdade de Medicina, University of S˜ ao Paulo, S˜ ao Paulo, SP, Brazil, 5 Department of Telecommunications and Control, Escola Politecnica, University of S˜ ao Paulo, S˜ ao Paulo, SP, Brazil
Y. Okita1 , H. Nakamura2 , I. Takahashi1 , T. Takaoka3 , K. Kouda4 , M. Kimura5 , T. Kobayashi6 , T. Sugiura7 1 Graduate School of Science and Technology, Shizuoka University, Japan, 2 Graduate School of Human Development and Environment, Kobe University, Japan, 3 Enseki Aojiru Co. Ltd, Japan, 4 Department of Public Health, Kinki University School of Medicine, Japan, 5 Faculty of Engineering, Shizuoka University, Japan, 6 Graduate School of Engineering, Kyoto University, Japan, 7 Research Institute of Electronics, Shizuoka University, Japan
Objective: Our aim is to apply the generalized partial directed coherence (gPDC) for multichannel time series recorded from a standard 10 20 EEG system electrodes during the episodes of mesial temporal epileptic seizure based on rigorous statistical asymptotic properties of the gPDC estimator. Methods: PDC is the representation of Granger causality in the frequency domain, which allows looking at frequency-wise interaction between brain areas. We introduced an adaptation to PDC definition, which can better deal with those cases of large power or prediction error disparity among the multivariate channels, called gPDC, for which rigorous asymptotic confidence intervals and statistical tests for the null hypothesis of zero gPDC using multivariate autoregressive modeling were derived. gPDC analysis routines were implemented in Python 2.5 (www.python.org) using NumPy, SciPy and Matplotlib packages. We sought to study mesial temporal epileptic patient’s EEG data. The midline derivations (Fz, Cz, Pz and Oz) were not considered in the analyses. Results: Confirming our previous work, the graph connectivity representations showed variable degree of laterality, an asymmetry of directed connections distribution toward the focus site hemisphere. Conclusions: The gPDC showed to be a robust measure of connectivity between time series in the frequency domain and its asymptotic statistics. Great care is necessary to exclude signal epochs with artifacts,
Objective: It is well known that odor influences human emotion and the status of the mind. The odor is also the most important sensation confirming the safety of food. Recently, it is reported that smelling of specific odors, such as green leaf and lavender, etc., has a relaxing effect on humans. Emotion researches which examine the effects of odors often use electroencephalogram (EEG) or heart rate variability (HRV). However, there are few reports which combined EEG and HRV in the evaluation of the odors. In this study, we measured EEGs at the frontal regions together with heart rate (HR) to investigate emotional effects of odor of squeezed vegetable leaf (Organic Kale). Methods: EEGs and HRV before and during odor presentation were analyzed by the Tone-Entropy method and a quantitative evaluation method of emotion based on the frontal alpha wave fluctuations (Yoshida Method). Eight non-smoker university students (8 men, mean age 21.4±0.5 years) with normal sense of smell participated in the experiments. EEGs were recorded at 19 electrode positions (10 20 system) before and during inhalation of the vegetable odor and ECGs were recorded according to CM5 induction. Results: Tone estimated from the heart rates has varied widely during the odor presentation compared with that before presentation. In contrast, the entropy before and during odor presentation showed no change. Furthermore, the state of mind estimated by the Yoshida method showed rapid changes before and during odor presentations.