An Ethernet motion-sensor based alarm system for epilepsy
Transcription
An Ethernet motion-sensor based alarm system for epilepsy
IRBM 32 (2011) 155–157 Original article An Ethernet motion-sensor based alarm system for epilepsy monitoring EPIMOUV : système de détection de crises nocturnes d’épilepsie par capture de mouvement S. Bonnet a,∗ , P. Jallon a , A. Bourgerette a , M. Antonakios a , R. Guillemaud a , Y. Caritu b , G. Becq c , P. Kahane c , P. Chapat d , B. Thomas-Vialettes e , F. Thomas-Vialettes e , D. Gerbi f , D. Ejnes f a CEA LETI, MINATEC CAMPUS, DTBS, 38054 Grenoble, France b MOVEA SA, 7 Parvis Louis Néel, 38000 Grenoble, France c CHU GRENOBLE, LNPPE, 38043 Grenoble, France d CTR Medical, 38500 Voiron, France e Association EPI, 69290 Craponnes, France f FCEs, 38590 Saint-Étienne de Saint-Geoirs, France Available online 25 February 2011 Abstract The EPIMOUV project aims at providing a place to live and autonomy for drug-resistant epileptic young adults while attempting to provide them a high level of security. In this framework, subjects will be monitored during night using body-worn motion sensors and an alarm will be triggered if an on-going seizure is detected by processing the accelerometer readings. In this paper, we describe an innovative architecture for such alarm system: the wireless inertial sensors are positioned on the subject’s body; the RF transceiver is located in the subject’s room and linked to an Ethernet gateway, and the remote motion capture acquisition system is interfaced with a medicalized alarm system. The EPIMOUV system will be evaluated in a specialized institution from December 2010. © 2011 Elsevier Masson SAS. All rights reserved. Keywords: Accelerometer; Alarm System; Detection; Epilepsy; Ethernet Résumé L’objectif global du projet EPIMOUV est d’augmenter l’autonomie des personnes souffrant d’une épilepsie pharmaco-résistante tout en préservant leur sécurité. Pour cela, le projet vise à développer un système d’alarme en cas de crise épileptique nocturne par analyse du mouvement. Dans ce papier, nous décrivons une architecture innovante de système d’alarme médicalisé : les accéléromètres sans fil sont placés sur le corps ; le récepteur RF est situé dans la chambre du résident et relié au réseau par connexion Ethernet ; et le système déporté d’acquisition de capture de mouvement et de détection de crises est interfacé à un système professionnel d’alarme médicalisé. Le système EPIMOUV sera évalué en institution spécialisé à partir de décembre 2010. © 2011 Elsevier Masson SAS. Tous droits réservés. Mots clés : Accéléromètre ; Alarme ; Détection ; Épilepsie ; Ethernet 1. Introduction Nowadays, body-mounted inertial/magnetic sensors are increasingly used in biomedical applications. Their main advantages are miniaturization, autonomy, low intrusiveness and unrestricted application range. In environment-controlled laboratory setups, they have shown to be reliable for upper-limb (kinematic analysis) and lower-limb (gait analysis), and they do ∗ Corresponding author. E-mail address: [email protected] (S. Bonnet). 1959-0318/$ – see front matter © 2011 Elsevier Masson SAS. All rights reserved. doi:10.1016/j.irbm.2011.01.021 fare well against optical tracking devices. In specialized institutions, there is definitely a need for monitoring the patient in his room and for collecting data remotely at a network scale either for offline analysis, e.g. sleep monitoring or online movement detection. This issue is especially relevant in the EPIMOUV project, dedicated to the monitoring of patients with pharmacoresistant epilepsy. 2. Materials and methods The proposed Ethernet-based solution is build upon watchsized MotionPod® technology from Movea. MotionPod® is a 156 S. Bonnet et al. / IRBM 32 (2011) 155–157 Fig. 1. MotionPod 3-axis accelerometer and 3-axis magnetometer (3a3m) sensors + MotionPod Ethernet Controller (81 × 56 × 40 mm). patented solution of motion sensing technology for accurate human body orientation or motion measurement. This is done by using miniaturized high-tech motion sensing MEMS containing 3-axis accelerometers (3a) and 3-axis magnetometers (3m) in a space the size of a standard wristwatch. Each MotionPod is timesynchronized by a controller and a proprietary TDMA protocol is used for data transmission. In each room, two wireless 3-D accelerometers, located on the subject’s torso and wrist, communicate with a dedicated RF receiver, connected to a server via Ethernet. The MotionPod Ethernet Controller has been developed and validated during the project and is shown in Fig. 1. The main application, named EPILOG, is driven from a remote PC for configuration, data acquisition and online processing. The supervision is limited to five rooms. The seizure detection algorithm is interfaced with a traditional nurse-call system: when an abnormal movement is detected by the software, a message is sent to the nurse’s pager, in order to give assistance to the resident in due time. The EPILOG software is coupled to the CTR Medical alarm system via an ASCOM Fig. 2. Nocturnal epilepsy monitoring system in specialized institution. Rooms under supervision are located in Building B while the acquisition/detection EPILOG system is located in Building A. Ethernet protocol is used to transmit data between both buildings. Wireless motion sensors and RF receiver are located in each resident’s room. platform and relays. Furthermore, for validation purposes, residents’ rooms are equipped with two wireless buttons for the staff to acknowledge the occurrence of a seizure or to indicate a false alarm. The components of the nocturnal seizure alarm system are described in Fig. 2. Different systems have been proposed in the literature for seizure detection. The most widespread ones are based on undermattress sensors but lack of specificity. Accelerometer is a better candidate for this task since tremor and abnormal limb orientation can easily be identified. The seizure detection algorithm has been patented and described in [1,2]. It is based on a Bayesian modelling of patient-specific seizures and also normal nocturnal movements using hidden Markov models. The sensor located at torso is mainly intended to detect a standing posture in order to inhibit the detection and prevent false alarms. The CHU Grenoble LNPPE, has also contributed to this task by a fine characterization of motor epileptic seizures using epilepsy unit Fig. 3. Example of a 3-axis accelerometer waveforms during a seizure. The sensor is positioned on the left wrist. This is a focal seizure evolving to a generalized convulsive one. Evolution of the waveforms can be observed during the succession of events: t0 - start of electroencephalographic seizure; t1 - start of clinical symptoms, t2: end of seizure; ATF - artefacts generated by caregivers; TONIC - Tonic manifestations; TONICO CLONIC (CONVULSIVE) - convulsive manifestations. S. Bonnet et al. / IRBM 32 (2011) 155–157 observations [3,4]. This work was achieved by first collecting synchronized video-EEG and motion data, and then annotating in a supervised manner the temporal distribution of seizures, as demonstrated in Fig. 3. 157 of the approach and the proposed system is now deployed in an epilepsy-specialized institution and validation is currently underway. Conflict of interest statement 3. Results In EPIMOUV project, the seizure detection algorithm has been first evaluated in home environment with the help of EPI association. Good performances have been reported on two epileptic persons, where 46 seizures were acquired during 39 nights. 90% of seizures have been correctly detected with an average number of false alarms per night of 0.7 [5]. Also the system was found acceptable by the epileptic young adults. Data acquisition is being performed from June 2010 to June 2011 in a FCEs medical institution according to a protocol approved by a local Ethical Committee. This institution is located in Saint-Étienne de Saint-Geoirs, Isère, France and it has been inaugurated in November 30, 2009 during the course of the project. Today, the whole acquisition processing chain has been validated from the resident to the caregiver and the medical validation will start from December 2010 to June 2011. Furthermore, a validation protocol has been defined with the Healthcare professionals who will test and validate the seizure detection system at night. The sensitivity and sensibility of the proposed solution will then be assessed in real conditions. 4. Conclusion This paper has presented an innovative remote motioncapture acquisition system. The system is based on wireless motion sensors and an Ethernet-based RF receiver. Data are then centralized into a server where dedicated processing can be efficiently performed in a multi-threading way, here for epilepsy detection. This kind of system is new and can meet other biomedical applications in hospital or nursing homes. Experiments in the first phase of the project have validated the proof-of-concept S. B., employed by CEA-Léti. P. J., employed by CEA-Léti. A. B., employed by CEA-Léti. R. G., employed by CEA-Léti. Y. C., co-funder of MOVEA. G. B., post-doctorant in CHU Grenoble. P. K., employed by CHU Grenoble. P. C., CEO of CTR Medical, network installer for FCEs institution in Saint-Étienne de Saint-Geoirs. B. T-V., no conflict of interest. F. T-V., no conflict of interest. D. G., employed by FCEs. Acknowledgments This research project was financially supported by the National Research Agency (ANR) through the call for proposals TecSan. References [1] Jallon P, Bonnet S, Antonakios M, Guillemaud R. Detection system of motor epileptic seizures through motion analysis with 3D accelerometers, EMBC 2009, Minneapolis, USA, 2009:2466–9. [2] Jallon P, Bonnet S. Dispositif et procédé de détection des crises d’épilepsie, CEA/MOVEA, PCT/EP2009/060740, 2009. [3] Becq G, Bonnet S, Minotti L, Antonakios M, Guillemaud R, Kahane P. Collection and exploratory analysis of attitude sensor data in an epilepsy monitoring unit, EMBC 2007, Lyon, France, 2007:2775–8. [4] Becq G, Bonnet S, Minotti L, Guillemaud R, Kahane P. Classification of epileptic motor manifestations using attitude sensors. Computers in Biology and Medicin 2011;41:46–55. [5] Jallon P.A Bayesian approach for epileptic seizures detection with 3D accelerometers sensors, EMBC, 2010 Buenos Aeres, Argentina 2010:6325–8.