Design of a PC-based multimedia telemedicine system for brain function teleconsultation

International Journal of Medical Informatics 61 (2001) 217–227

Design of a PC-based multimedia telemedicine system forbrain function teleconsultation

Sun K. Yoo a,*, Sun H. Kim b, Nam H. Kim a, Y.T. Kang a, K.M. Kim a,S.H. Bae a, Michael W. Vannier c

a Department of Medical Engineering, Yonsei Uni6ersity College of Medicine, 120-140 Sudaemoon-Gu, Shinchon-Dong 134,Seoul, South Korea

b Department of Neurosurgery, Yonsei Uni6ersity College of Medicine, Seoul, South Koreac Department of Radiology, The Uni6ersity of Iowa School of Medicine, 200 Hawkins Dr., 3984 JPP, Iowa, IA 52242, USA

Abstract

During time-critical brain surgery, the detection of developing cerebral ischemia is particularly important becauseearly therapeutic intervention may reduce the mortality of the patient. The purpose of this system is to provide anefficient means of remote teleconsultation for the early detection of ischemia, particularly when subspecialists areunavailable. The hardware and software design architecture for the multimedia brain function teleconsultation systemincluding the dedicated brain function monitoring system is described. In order to comprehensively support remoteteleconsultation, multi-media resources needed for ischemia interpretation were included: EEG signals, CSA,CD-CSA, radiological images, surgical microscope video images and video conferencing. PC-based system integrationwith standard interfaces and the operability over the Ethernet meet the cost-effectiveness while the modular softwarewas customized with a diverse range of data manipulations and control functions necessary for shared workspace andstandard interfaces. © 2001 Elsevier Science Ireland Ltd. All rights reserved.

Keywords: Brain function monitoring; Teleconsultation; Multimedia

www.elsevier.com/locate/ijmedinf

1. Introduction

There are many kinds of telemedicine sys-tems that have been developed to allow med-ical personnel to share and exchangeinformation between hospitals. Recently,multimedia has been included to support awide variety of medical information with the

advance of computer and communicationtechnologies [1–5]. Particularly, the trend to-ward specialization in medicine and the con-centration of specialists in certain hospitalshas created the demand for a multimediatelemedicine system to provide expert consul-tation over a network [4]. When subspecial-ists are not available, telemedicine plays animportant role in enhancing the quality ofpatient care, and in reducing the morbidity ofpatients at a critical time like brain surgery.

* Corresponding author. Tel.: +82-2-3615403.E-mail address: [email protected] (S.K. Yoo).

1386-5056/01/$ - see front matter © 2001 Elsevier Science Ireland Ltd. All rights reserved.

PII: S1 386 -5056 (01 )00143 -5

S.K. Yoo et al. / International Journal of Medical Informatics 61 (2001) 217–227218

During brain surgery, the patient’s physio-logical status is dynamic and subject to rapid,life threatening changes because the function-ing of the cerebral cortex is exquisitely sensi-tive to its environment. Insufficient cerebralblood flow or inadequate partial oxygen pres-sure is reflected within seconds in the elec-troencephalograph (EEG). However, EEGwaveforms are so complex that a brain-func-tion monitoring system called compressedspectral array (CSA) system [6–11] is oftenused in the early detection of ischemic eventduring carotid artery cross-clamping in brainsurgery.

The detection of a developing cerebral is-chemia is particularly important becauseearly therapeutic intervention may reduce pa-tient mortality [1–6]. However, the numberof subspecialist surgeons or anesthesiologistswho can quickly evaluate and interpret theischemia by observing the brain functionmonitoring system is very limited. In thissituation, various forms of medical datashould be gathered and transmitted simulta-neously allowing the subspecialist to performthe medical procedures with the proper inter-pretation. This includes not only video con-ferencing in terms of face-to-faceconversation between a demanding medicalsurgeon and a consulting subspecialist, butalso supplementary facilities that affect theconsultation directly or indirectly, such asEEG waveforms, CSA data, radiological im-ages, and surgical microscope video images.

In this study, we designed a multimediatelemedicine system to provide brain functionteleconsultation. The design criteria werebased on low-cost implementation and open-standard realization, as well as the inclusionof as much medical multimedia as possiblefor brain function teleconsultation. Finally,technical and clinical experiments were per-formed to evaluate the operability and effi-cacy of this system.

2. Materials and methods

2.1. The design of the brain-functionmonitoring system

A brain-function monitoring system cansupport three modes of display: raw EEGsignals, CSA and Color Density CompressedSpectral Array (CD-CSA) display modes [6–11]. Our brain function monitoring systemwas specially designed to allow simultaneoustransmission of three modes of display in realtime. As shown in Fig. 1, the designed brainfunction monitoring system consisted of 2-channel analog EEG amplifiers, analog-digi-tal converter and dedicated software routines.The analog EEG amplifier designed by ourteam [12] was adjusted to accommodate anoisy environment such as the operatingroom. It also had the capability to adjust thegain and filter setting for low and high cutthrough a computer control. An analog-to-digital converter was connected to thetelemedicine system to acquire the sampledEEG data at the rate of 128 Hz, which wasappropriate for the computation of CSA andCD-CSA [6–11]. The software routinely per-formed pre-processing steps (including digitalhigh pass filter, and hamming window), andthe computation of EEG power spectrumswas based on overlapped Fast Fourier Trans-form (FFT). The display of CSA and CD-CSA was accomplished by successivelylayering the computed spectrum and bymatching the pseudo color to the amplitudeof the power spectrum, respectively.

2.2. The hardware architecture of thebrain-function teleconsultation system

The personal computer (PC) was selectedas the platform, since it offered a cost-effec-tive solution and a wide range of inexpensivemulti-media devices, including teleconferenc-

S.K. Yoo et al. / International Journal of Medical Informatics 61 (2001) 217–227 219

ing and motion picture expert group (MPEG)[13,14] board. The brain function teleconsul-tation system as in Fig. 2 consisted of anetwork, media interfacing devices and com-puter. Pentium PC with 32 MB of memoryand auxiliary storage is currently used.

The equipment at the transmitting site in-cludes: (1) teleconferencing video CODEC(Coder/DECoder) with video camera, micro-phone and speaker for face-to-face conversa-tion, (2) analog-to-digital converter for rawEEG signals, CSA and CD-CSA, (3) MPEG-1 board for the on-line transmission of surgi-cal microscope video sequence images, (4)Small Computer System Interface (SCSI)board for acquiring radiological imagesthrough scanner and (5) network card forcommunication over the network.

CODEC board (Intel Co., Proshare) basedon H.320 videoconferencing standard [15]performed the compression and decompres-sion of video and audio signals with a low bitrate of 128 kbps, which allocated enoughmargin to other media resources to operateeven over the low bandwidth Ethernet.

To allow for the on-line transmission of

microscopic video sequence images, aMPEG-1 board (Darim Co.) based onMPEG-1 motion picture standard [13] wasinterfaced with the video signal source of asurgical microscope to compress the time-se-quenced surgical microscope video images(moving picture). The quantizer step-size andframe-drop rate of a MPEG-1 board werecontrolled to allow for a low bandwidth net-work like the Ethernet.

The radiological images for transmissioncan be obtained from two different sources:scanner and picture archiving and communi-cations systems (PACS). Analog images suchas X-ray films are digitized by scanner, whiledigital images such as computed radiography(CR), computed tomography (CT), and mag-netic resonance imaging (MRI) are directlytaken from the PACS archive as a digitalimaging and communications in medicine(DICOM) compliant format [16].

The equipment at the receiving site is sim-pler than at the transmitting site because allthe multi-media devices, apart from the tele-conferencing board, are only required at the

Fig. 1. Schematic diagram of brain function monitoring system. It consists of 2-channel analog EEG amplifiers,analog-to-digital converters and software routines to calculate and display CSA and CD-CSA. The computerremotely controls the gain and filter setting of analog EEG amplifier and sampling frequency of analog-to-digitalconverter.


Fig. 2. Hardware configuration of multimedia brain function telemedicine system. Transmitting site is equipped withteleconferencing CODEC, A/D converter, MPEG-1 board, SCSI, and LAN card, but only teleconferencing CODECand LAN card are required at receiving terminal.

transmission. Our current PC platform isequipped with an Ethernet network card tohandle a bandwidth of 10 Mbps, becausehigh speed backbone networks such as asyn-chronous transfer mode (ATM) or fiber dis-tributed data interface (FDDI) can supportmulti-connection through HUB (Backbone toEthernet connection) distribution. Moreover,Ethernet is almost universal in hospitals dueto its cost effectiveness.

2.3. The software architecture of brainfunction teleconsultation system

The software was designed to customize adiverse set of data manipulations and controlfunctions needed for shared workspace andstandard interfaces. Using object orientedprogramming language, Visual C+ + [17],and running under the Windows 95/98 envi-ronment, it employs a modular designapproach.

The program structure puts an emphasison standardization, hardware independence,and expandability, which are defined by a set

of classes organized in different layers. Asshown in Fig. 3, the lower layer classes areused as device driver interfaces which arespecific to individual physical interface com-ponents. These include A/D driver for EEGsignals acquisition, GDI driver [18] for screen

Fig. 3. Software components are defined by a set ofclasses organized in different layers. The lower layerclasses based on Windows kernel are used as devicedriver interfaces which are specific to individual physi-cal media components of the integrated multimediasystem, while processing and presentation of variousresources are performed on highest layer classes.


display, TWAIN driver [18] for scanner inter-face, DICOM message driver for PACS inter-face, H.320 audio/video driver forteleconferencing CODEC interface, and net-work driver [19] for TCP/IP network proto-col, and Active movie driver [18] forMPEG-1 interface.

The highest layer modules are functionallydivided into four separate modules: mediamanipulation, data acquisition, graphicaluser interface and telecommunicationmodule.

The communication module will allow thissystem to link with other sites through localarea network (LAN) and wide area network(WAN). Communication service based on theTCP/IP protocol, the ‘de-facto’ standard pro-tocol, has been implemented by using aWinsock 1.1 Application Program Interface(API) library [19]. It also assigns the trans-mission priority. Radiological images aretransmitted first before the session starts.During session, teleconferencing and realtime EEG signals have higher priority thanmicroscopic video-sequenced images. Also,the system handles the DICOM message ex-change to interface the DICOM gateway ofPACS.

A data acquisition module based on under-lying device driver interfaces will be able tomake this system as diverse as possible fordata acquisition. It also remotely controls thegain and filter settings for the EEG amplifi-ers, as well as adjusts the quantizer step-sizeand frame-drop rate for the MPEG-1 boarddepending on the available network band-width. Alternatively, the data acquisitionmodule at the receiving site can perform thedecoding process instead of a MPEG-1 hard-ware board. The module processes theMPEG-1 bit stream of microscopic video-se-quenced images by using an active moviedriver, and this reduces the overall cost of thereceiving site system. Moreover, it includes

implemented DICOM decoding functionsbased on the DICOM data dictionary [16].

In order not to require extensive skill inoperating the brain function teleconsultationsystem, the graphical user interface module isbased on Windows icons, buttons and menus.As the Windows 95/98 kernel provides a richset of graphical user interface functions[17,18], user interface control classes are com-posed of API functions. Those functionsprovide tools for accessing, controlling, andmanipulating the multi-media data on sharedworkspace for teleconsultation.

The data manipulation module enablescomputation of CSA and CD-CSA fromtransmitted EEG signals, as well as providesfunctions to manipulate data in a sharedworkspace. Shared workspace is designed toallow for all actions initiated at differentterminals (either transmission or receiving) tobe performed in the same order and samefashion at each terminal. Working in ashared workspace is done by telepointer.Control of the telepointer is handled by giv-ing the privilege to the user moving it first,but the multiplex time scheme permits thetelepointer operator only a small amount ofreaction time in maintaining a seamless oper-ation. Shared workspace controls (1) the se-lection and display of multimedia windows,(2) the maximization, minimization, restora-tion, and destruction of window, and (3) theadjustment of window levels and zoom fac-tor. In addition, the system performs imageprocessing functions such as smoothing, edgeenhancement, histogram processing, as wellas flipping, rotating and reversing images.

3. Results

3.1. Experiments

Two stage experiments were carried outintra hospital and inter hospital, respectively.


Fig. 4. Six different measurements were randomly per-formed on the 10 Mbps Ethernet to measure the intrahospital transmission performance at three typical peaktime periods; 10:00–11:00, 14:00–15:00, and 16:00–17:00 h.

was so small that they could be ignored, andthe transmission of radiological images wasonly required once before starting the session.

At the second stage, the inter hospitalsystem performance was tested betweenYonsei University Medical Center (ShinchonSeverance Hospital, Seoul) and it’s affiliatedhospital (Yongdong Severance Hospital,Seoul) as shown in Fig. 5. The distancebetween them is about 30 km. Brain functionteleconsultation systems at each site wereconnected to 10 Mbps Ethernet LAN throughan ATM-to-Ethernet switching HUB, whilethe two hospitals were connected through 155Mbps ATM WAN. Because the Ethernetinfluenced on the overall performance of theactual transmission rate due to higherbandwidth of the ATM WAN than that ofEthernet, inter hospital tests focused on theoperation of shared workspace and thehandling of EEG, CSA, CD-CSA, radiologicalimages, teleconferencing and surgicalmicroscope video images. As in Fig. 6, theexperiment shows that diverse media windowsdisplayed simultaneously on the monitorenables the subspecialist to make a comparisoncomprehensively during consultation. Realtime display of EEG signals, CSA, CD-CSAand teleconferencing also gives thesubspecialist critical information forinterpreting the ischemic event in a timelymanner. In addition, the telepointer supportsthe manipulation of suspicious data on theshared workspace, as well as selecting thedisplay windows among many types of mediawindows. Throughout the whole testing periodfrom 08:00 to 17:00 h, real time display ofEEG, CSA, CD-CSA and teleconferencingwere maintained, while the frame rate for thetransmission of the surgical microscope videoimages were changed from 8 to 30 frames/s.Although the full frame display rate ofMPEG-1 was not supported due to the band

At the first stage, intra hospital transmissionperformance over the Ethernet (10 Mbps) wastested to evaluate the actual transmission rateunder the low performance Ethernetenvironment, since the Ethernet is widely usedas the cheapest hospital network. As shown inFig. 4, the actual transmission rate rangedfrom 0.94 to 2.1 Mbps depending on thenetwork load. Even the measured minimumtransmission rate of 0.94 Mbps was capable oftransmitting 2-channel EEGs andteleconferencing data in real time, because 2and 128 kbps (for both audio and video) wererequired, respectively. The remainingbandwidth ranged from 0.81 to 1.97 Mbpscorresponding to measured minimum andmaximum transmission rate. It sometimes wasnot enough to transmit the surgical microscopevideo images at 30 frames/s due to the lack ofMPEG-1’s 1.5 Mbps bandwidth requirement[13]. However, the purpose of the transmissionof the surgical microscope video images is toprovide the supplementary information tomatch CSA interpretation to the overallsurgical procedure. Therefore, the remainingbandwidth could be allocated to transmit thesurgical microscope video images with reducedframe rate, because the number of control bits


width limitations of the Ethernet, continuallychanging images of surgical microscope gavethe subspecialist supplementary informationto understand the surgical procedure whileinspecting the CSA display.

3.2. Clinical trial

The 15 cases of clinical consultation werecarried out to evaluate the usefulness of thissystem between the two hospitals. Fig. 7shows the typical example that this systemwas successfully applied to. The consultationwas initiated before the start of surgical oper-ation. The transmitted EEG signals and cal-culated CSA and CD-CSA were continuouslydisplayed on the sub-specialist’s receiving ter-minal while the operation was underway. Asshown in Fig. 7, symmetrical electrical activ-ity (normal state) was monitored on bothtemporal areas (T3 and T4) before the leftintra carotid artery (ICA) was clipped. How-ever, 15 min later after left ICA clipping, thesub-specialist found that electrical activityhad decreased. Since ischemic events are indi-

cated by those activities, an immediate con-sultation began through teleconferencing.The subspecialist consulted that the carotidclamp should be removed to prevent cerebralischemia. Since the electrical activity on bothsides slowly recovered after operation, thisconsultation prevented severe patient damageduring the operation. Therefore, this pilotclinical trial shows the usefulness of this sys-tem in case of detecting the ischemic eventwhen time-critical decision making is re-quired during an operation. Real-time trans-mission of EEG signals and real-time displayof CSA and CD-CSA can provide criticalsupport in interpreting the causes of cerebralischemia.

4. Discussion

The evolution of high-speed computer andcommunication technologies will enable thedevelopment of a multimedia system capableof supporting a wide variety of media such astext, audio, image and video. In the medical

Fig. 5. Experimental systems were set up between Yonsei University Medical Center (Shinchon Severance) and it’saffiliated hospital (Yongdong Severance). The distance is about 30 km. Each terminal was connected to the Ethernet(10 Mbps) via ATM-to-Ethernet switching HUB (Catalist 5000). Each hospital was equipped with ATM switch (ForeSystem Co.) connected to ATM WAN (155 Mbps).


Fig. 6. Monitor screen captured at the receiving site. Inter hospital performance was tested between two hospitalsfrom 08:00 to 17:00 h. Real time display of EEG, CSA, CD-CSA and teleconferencing were maintained throughoutthe whole testing period, while the frame rate of on-line video images were changed from 8 to 28 frames per sdepending on the network traffic and the number of opened media.

environment, the multimedia telemedicinesystem will lead to a situation where subspe-cialists among widespread hospitals can ex-change multimedia data including patientdemographic data, audio, video, radiologicalimages and even surgical video images. Al-though many multimedia telemedicine sys-tems have been developed in the past fewyears, very few systems for brain surgeryhave been developed [1–5]. Moreover, a PC-based system over Ethernet has not yet beenconsidered to meet the cost-effectiveness ofthe multimedia system for brain functionteleconsultation.

During brain surgery, cerebral ischemiadue to arterial vasospasm is a major source

of morbidity in patients with cerebrovasculardisease. There are many causes of cerebralischemia, which may occur in operatingroom. These include carotid artery surgery,stroke, subarachnoid hemorrhage and headtrauma [6–11]. Early detection of the devel-oping cerebral ischemia enables early thera-peutic intervention, which may reduce themortality. Routine EEG measures correlatewith the changes in cerebral metabolic func-tion over a wide range of cerebral blood flowvalues. However, EEG data measured byconventional EEG machine is so complexthat most clinicians have severe difficulty indetecting cerebral ischemia during theoperation.


The brain function monitoring system(CSA system) is a new tool that aids theneurosurgeon to detect critical changes inblood flow at an early stage to avoid cerebralischemia, because quantitative computerizedanalysis correlates well with conventionalEEG interpretation. The CSA system pro-cesses the EEG data and transforms it intomore easily readable form by applying FFTand pseudo three-dimensional display tech-nique. Although brain function monitoringsystems are commercially available now,those systems are not adequate for the simul-taneous transmission of three modes of dis-played data: EEG, CSA, and CD-CSAdisplay mode. If a screen capturing methodwas used to transmit three modes of data

digitally, at a refreshment rate of 30 framesper s with 256-color mode and pixel resolu-tion of 1024 by 768 for real time transmis-sion, then the network bandwidth of 180Mbps would be required. However, if thesame brain function monitoring software wasinstalled at both the transmitting and receiv-ing sites to help synchronize both systemsand 2-channel EEG signals, with 8-bit signalresolution and a sampling rate of 128 Hz,then the total bandwidth could be reduced to2 kbps because CSA and CD-CSA can bederived from EEG signals. Therefore, thededicated brain function monitoring systemshould be designed to maximize the efficiencyof a given communication channel.

For a brain function teleconsultation sys-tem to succeed, many types of multi-mediamust be transmitted together to permit aremote subspecialist to make a proper inter-pretation of an ischemic event. Those includenot only EEG signals and related CSA andCD-CSA for interpreting brain function, butalso supplementary data that directly or indi-rectly affects consultations such as radiologi-cal images for localizing the lesion, on-linesurgical microscope video images for moni-toring surgical procedure, and video confer-encing for face-to-face conversation betweenconsulting and requesting doctors.

In order to support the various tasksneeded for consulting the ischemic event dur-ing operation, the following criteria should beconsidered to design the procedure. Real-timetransmission of EEG signals is prerequisitefor the early detection of ischemic event. Thelatency due to the transmission and the calcu-lation of the CSA and CD-CSA should beminimal for synchronizing the display at boththe transmitting and receiving sites. The la-tency for radiological images are less critical,but DICOM based implementation is im-portant because of compromising the con-nection between PACS and multi-vendor

Fig. 7. Clinical trial was carried out to demonstrate theusefulness of the implemented system. Fifteen minutesafter left ICA clipping, the electrical activity was de-creased and frequency shift had occurred at the lefttemporal region. The while line represents the energy ofthe electrical activity. The remote subspecialist con-sulted the carotid clamp should be removed to preventcerebral ischemia.


telemedicine equipment [20]. Moving picturesfor a surgical microscope should be sup-ported to understand the current surgicalprocedure when interpreting the ischemicevent. The inclusion of the MPEG-2 insteadof MPEG-1 allows the high quality display ofsurgical microscope images, but the band-width requirement of 15 Mbps should notcompromise the operability over the mostpopular Ethernet network. Both audio andvideo should be supported in teleconferencingfor efficient consultation. Audio latency andsynchronization problems compared withvideo should be minimized to allow the ex-change of opinion during consultation. Par-ticipants must be able to view andmanipulate a common set of multimedia be-cause the subspecialist typically views manykinds of media simultaneously during theconsultation process. Moreover, a telepointershould be provided to mark the suspiciousdetails of information displayed on theshared workspace. The software packageshould be designed to meet the requirementsfor accessibility, expandability and modular-ity. Also, it is designed to allow the unifiedcustomization of data processing and controlfunctions, the systematic interface of hard-ware components and easy graphical inter-face for non-computer oriented medicalpersonnel.

Many different types of local and widearea networks ranging from low performanceEthernet to high performance ATM (155Mbps) has been installed within and betweenhospitals, which are supposed to be satisfiedwith the minimal bandwidth requirement formulti-media transmission. However, from thepoint of fiscal practicality, operating on thelow performance network is important be-cause the Ethernet is the cheapest and mostpopular LAN. Nevertheless, TCP/IP basedimplementation can cover networks rangingfrom low to high performance regardless of

the underlying network bandwidth. Operabil-ity over 155 Mbps ATM means that multipletelemedicine equipment can be installedsimultaneously at the multiple places withinhospital.

In conclusion, multimedia brain functiontelemedicine system has been designed andtested for remote brain function teleconsulta-tion, especially for the early detection of is-chemic event during brain surgery. In orderto support remote teleconsultation in a com-prehensive fashion, multi-media resourcesneeded for teleconsultation were integrated:EEG signals, CSA, CD-CSA, radiologicalimages, surgical microscope video images andvideo conferencing. PC-based system integra-tion, dedicated design of brain function mon-itoring system and standard interfaceprovided this system with cost-effectivenessand open-standard implementation. Experi-ments carried out intra and inter hospitalresulted in usability for improved patient carein a timely fashion, even in the low perfor-mance network environment. Although weconcentrated here on the work carried out inbrain surgery, the system developed can beapplied to many other telemedicine areas re-quiring support for cooperative consultation.

References

[1] K. Ahmed, O.B. Luis, D.G. Licolas, A multimediamedical communications system, IEEE J. SAC 8(1990) 325–339.

[2] J.E. Cabral, Y. Kim, Multimedia systems fortelemedicine and their communications require-ments, IEEE Comm. Mag. 34 (7) (1996) 20–27.

[3] L. Kleinholz, Supporting cooperative medicine:the Bermed project, IEEE Multimedia 1 (1994)44–53.

[4] L. Orozco-Barbosa, A. Karmouch, N.D. Geor-ganas, M. Goldberg, A multimedia interhospitalcommunication system for medical consultations,IEEE J. SAC 10 (1992) 1145–1157.


[5] W.G. Tohme, W.S. Hayes, H. Dai, D. Komo, J.Pahira, D.R. Abernethy, W. Rennert, K.S. Kuehl,G.J. Hauser, S.K. Mun, Technology assessmentfor an integrated PC-based platform for threetelemedicine applications, SPIE PACS DesignAppl. 2711 (1996) 335–344.

[6] E.I. Gusev, A.V. Pokrovskii, Y. Volynskii, L.I.Pyshkina, O. Erokhin, V.V. Goloma, A.B. Gekht,V.B. Levtova, P.V. Mal’tsev, Compression spec-tral analysis of the EEG in patients with occlusivelesions of the carotid and vertebral arteries, Neu-rosci. Behav. Physiol. 19 (1989) 51–56.

[7] E.I. Gusev, A.I. Fedin, O. Erokhin, I.P. Shvarts,Compressed spectral analysis of the EEG in pa-tients with acute cerebrovascular disturbance,Neurosci. Behav. Physiol. 15 (1985) 144–151.

[8] T. Morioka, T. Matsushima, K. Fujii, M. Fukui,K. Hasuo, K. Hisashi, Balloon test occlusion ofthe internal carotid artery with monitoring ofcompressed spectral arrays (CSAs) of electroen-cephalogram, Acta Neurochir. 101 (1989) 29–34.

[9] G. Pfurtscheller, G. Schwarz, O. Schroettner, G.Litscher, H. Maresch, L. Auer, W. List, Continu-ous and simultaneous monitoring of EEG spectraand brainstem auditory and somatosensoryevoked potentials in the intensive care unit and theoperating room, J. Clin. Neurophysiol. 4 (1987)389–396.

[10] E.R. Steele, M.S. Albin, J.L. Monts, P.K. Har-man, Compressed spectral array EEG monitoringduring coronary bypass surgery in a patient withvertebrobasilar artery insufficiency, Anesth.Analg. 66 (1987) 271–273.

[11] R. Tempelhoff, P.A. Modica, R.L. Grubb, K.M.Rich, B. Holtmann, Selective shunting duringcarotid endarterectomy based on two-channelcomputerized electroencephalographic/compressedspectral array analysis, Neurosurgery 24 (1989)339–344.

[12] S.K. Yoo, N.H. Kim, J.R. Kim, S.H. Kim, Thedevelopment of high precision EEG amplifier forthe computerized EEG analysis, in: Proceedings ofAnnual International Conference of IEEE-EMBS,1995.

[13] MPEG-1 CD. Committee Draft ISO/IEC 11172Working Group 11, ISO, IPSJ, 1991.

[14] MPEG-2 CD Committee Draft ISO/IEC 13818JTC1/SC29/WG11, ITU, 1994.

[15] ITU-T Rec. H.320. Narrowband Visual TelephoneSystems and Terminal Equipment. rev.1, 1994.

[16] American College of Radiology, National Electri-cal Manufacturers Association, Digital Imagingand Communications in Medicine(DICOM): Ver-sion 3.0, Draft Standard, ACR-NEMA Commit-tee, Working Group VI, Washington, DC, 1993.

[17] D.J. Kruglinski, Inside Visual C+ + , fourth ed.,Microsoft Press, 1995.

[18] A. Peter, J. Scott, Visual C+ + Multimedia,Coriolis Group, Arizona, 1995.

[19] A.K. Sinha, Network programming in WindowsNT, first ed., Addison-Wesley, New York, 1996.

[20] G. Orphanos, Kanellopoulos, S. Koubias, Com-munication software for physicians’ workstationssupporting medical imaging services, SPIE PACSDesign Eval. 1899 (1993) 476–485.

.

Design of a PC-based multimedia telemedicine system for brain function teleconsultation

Documents