Press Release

Similar documents
-To become familiar with the input/output characteristics of several types of standard flip-flop devices and the conversion among them.

DO NOT COPY DO NOT COPY DO NOT COPY DO NOT COPY

Measurement of Capacitances Based on a Flip-Flop Sensor

10. Water tank. Example I. Draw the graph of the amount z of water in the tank against time t.. Explain the shape of the graph.

Lab 2 Position and Velocity

Workflow Overview. BD FACSDiva Software Quick Reference Guide for BD FACSAria Cell Sorters. Starting Up the System. Checking Cytometer Performance

4.1 Water tank. height z (mm) time t (s)

IN THE FOCUS: Brain Products acticap boosts road safety research

application software

EX 5 DIGITAL ELECTRONICS (GROUP 1BT4) G

Enabling Switch Devices

First Result of the SMA Holography Experirnent

application software

G E T T I N G I N S T R U M E N T S, I N C.

Besides our own analog sensors, it can serve as a controller performing variegated control functions for any type of analog device by any maker.

Advanced Handheld Tachometer FT Measure engine rotation speed via cigarette lighter socket sensor! Cigarette lighter socket sensor FT-0801

Monitoring Technology

2015 Communication Guide

Overview ECE 553: TESTING AND TESTABLE DESIGN OF. Ad-Hoc DFT Methods Good design practices learned through experience are used as guidelines:

Student worksheet: Spoken Grammar

SAFETY WITH A SYSTEM V EN

Adaptive Down-Sampling Video Coding

TUBICOPTERS & MORE OBJECTIVE

Press Release. Dear Customers, Dear Friends of Brain Products,

Digital Panel Controller

Removal of Order Domain Content in Rotating Equipment Signals by Double Resampling

Personal Computer Embedded Type Servo System Controller. Simple Motion Board User's Manual (Advanced Synchronous Control) -MR-EM340GF

LATCHES Implementation With Complex Gates

MULTI-VIEW VIDEO COMPRESSION USING DYNAMIC BACKGROUND FRAME AND 3D MOTION ESTIMATION

Computer Graphics Applications to Crew Displays

CE 603 Photogrammetry II. Condition number = 2.7E+06

A Turbo Tutorial. by Jakob Dahl Andersen COM Center Technical University of Denmark

UPDATE FOR DESIGN OF STRUCTURAL STEEL HOLLOW SECTION CONNECTIONS VOLUME 1 DESIGN MODELS, First edition 1996 A.A. SYAM AND B.G.

Solution Guide II-A. Image Acquisition. Building Vision for Business. MVTec Software GmbH

Physics 218: Exam 1. Sections: , , , 544, , 557,569, 572 September 28 th, 2016

The Art of Image Acquisition

AUTOCOMPENSATIVE SYSTEM FOR MEASUREMENT OF THE CAPACITANCES

THERMOELASTIC SIGNAL PROCESSING USING AN FFT LOCK-IN BASED ALGORITHM ON EXTENDED SAMPLED DATA

Solution Guide II-A. Image Acquisition. HALCON Progress

Nonuniform sampling AN1

MELSEC iq-f FX5 Simple Motion Module User's Manual (Advanced Synchronous Control) -FX5-40SSC-S -FX5-80SSC-S

TEA2037A HORIZONTAL & VERTICAL DEFLECTION CIRCUIT

On Mopping: A Mathematical Model for Mopping a Dirty Floor

DIGITAL MOMENT LIMITTER. Instruction Manual EN B

LABORATORY COURSE OF ELECTRONIC INSTRUMENTATION BASED ON THE TELEMETRY OF SEVERAL PARAMETERS OF A REMOTE CONTROLLED CAR

The Impact of e-book Technology on Book Retailing

The Art of Image Acquisition

Trinitron Color TV KV-TG21 KV-PG21 KV-PG14. Operating Instructions M70 M61 M40 P70 P (1)

LCD Module Specification

THE INCREASING demand to display video contents

Drivers Evaluation of Performance of LED Traffic Signal Modules

(12) (10) Patent N0.: US 7,260,789 B2 Hunleth et a]. (45) Date of Patent: Aug. 21, 2007

Diffusion in Concert halls analyzed as a function of time during the decay process

TRANSFORM DOMAIN SLICE BASED DISTRIBUTED VIDEO CODING

A Methodology for Evaluating Storage Systems in Distributed and Hierarchical Video Servers

SMD LED Product Data Sheet LTSA-G6SPVEKT Spec No.: DS Effective Date: 10/12/2016 LITE-ON DCC RELEASE

H3CR. Multifunctional Timer Twin Timer Star-delta Timer Power OFF-delay Timer H3CR-A H3CR-AS H3CR-AP H3CR-A8 H3CR-A8S H3CR-A8E H3CR-G.

LCD Module Specification

I (parent/guardian name) certify that, to the best of my knowledge, the

Coded Strobing Photography: Compressive Sensing of High-speed Periodic Events

Real-time Facial Expression Recognition in Image Sequences Using an AdaBoost-based Multi-classifier

Communication Systems, 5e

Novel Power Supply Independent Ring Oscillator

Computer Vision II Lecture 8

Computer Vision II Lecture 8

Telemetrie-Messtechnik Schnorrenberg

ZEP - 644SXWW 640SX - LED 150 W. Profile spot

SAFETY WARNING! DO NOT REMOVE THE MAINS EARTH CONNECTION!

Determinants of investment in fixed assets and in intangible assets for hightech

SOME FUNCTIONAL PATTERNS ON THE NON-VERBAL LEVEL

A ROBUST DIGITAL IMAGE COPYRIGHT PROTECTION USING 4-LEVEL DWT ALGORITHM

Type: Source: PSU: Followspot Optics: Standard: Features Optical Fully closing iris cassette: Long lamp life (3000 h) Factory set optical train:

R&D White Paper WHP 120. Digital on-channel repeater for DAB. Research & Development BRITISH BROADCASTING CORPORATION.

f, I f, f, t t A Tale of Two Cities : A Study of Conference Room Videoconf erencing I ELEI{TE)ENLE 0nulo Telepresence Project

SC434L_DVCC-Tutorial 1 Intro. and DV Formats

Commissioning EN. Inverter. Inverter i510 Cabinet 0.25 to 2.2 kw

Q = OCM Pro. Very Accurate Flow Measurement in partially and full filled Pipes and Channels

Lancelot TS. Grand W HTI. Followspot. Type: Followspot Source: 4000 W HTI PSU: Magnetic - hot restrike Optics: 2 to 5 zoom.

Video Summarization from Spatio-Temporal Features

Ten Music Notation Programs

Region-based Temporally Consistent Video Post-processing

Circuit Breaker Ratings A Primer for Protection Engineers

LCD Module Specification

VECM and Variance Decomposition: An Application to the Consumption-Wealth Ratio

The Measurement of Personality and Behavior Disorders by the I. P. A. T. Music Preference Test

Flo C. Compact W MSR. Followspot

AJ- P. Operating Instructions. Digital Video Cassette Recorder. Printed in Japan VQT S0699W3119 A OFF CH1 CH2 CH2 RESET COUNTER CH3 CH4

TLE7251V. 1 Overview. Features. Potential applications. Product validation. High Speed CAN-Transceiver with Bus Wake-up

You can download Mozart s music. You can t download his genius!

TLE Overview. High Speed CAN FD Transceiver. Qualified for Automotive Applications according to AEC-Q100

CHEATER CIRCUITS FOR THE TESTING OF THYRATRONS

Evaluation of a Singing Voice Conversion Method Based on Many-to-Many Eigenvoice Conversion

MELODY EXTRACTION FROM POLYPHONIC AUDIO BASED ON PARTICLE FILTER

And the Oscar Goes to...peeeeedrooooo! 1

Automatic location and removal of video logos

TLE6251D. Data Sheet. Automotive Power. High Speed CAN-Transceiver with Bus Wake-up. Rev. 1.0,

Sartorius Combics Series

United States Patent (19) Gardner

Marjorie Thomas' schemas of Possible 2-voice canonic relationships

TLE7251V. Data Sheet. Automotive Power. High Speed CAN-Transceiver with Bus Wake-up TLE7251VLE TLE7251VSJ. Rev. 1.0,

Transcription:

Press Release 3.2008 Ediorial Analyzer 2 is released, he mos complex and difficul sofware upgrade in he hisory of our company. No only he developmen eam, he whole saff gave heir bes o make his release a success. And he congraulaions and posiive feedback we received from you immediaely afer he release showed us, ha our work was successful indeed. I wan o hank our whole eam, our consulans and also our Bea Tesers for heir conribuion o his sory. Also immediaely afer he release firs quesions appeared abou, wha will happen now? Wha are he fuure plans of our sofware developmen? Which feaures will come nex in Analyzer? Sure i is oo early o promise new feaures, when he crew is jus aking breah and cusomers jus sar heir experience wih he release. Bu one promise holds, ha I wroe already in a press release before: his is no he end of he developmen bu he beginning of a new era on a new and flexible sofware plaform. We will for example use he power of he new implemened MATLAB inerface, o bring a new kind of soluions and AddIns wih los of new possibiliies ino he analyzing process. Anoher frequenly asked quesion was wha now will happen o he Analyzer 1. Also his produc will be suppored and mainenanced, so a sof migraion is guaraneed. Dr. Achim Hornecker Brain Producs General Manager Impressum Brain Producs GmbH, Zeppelinsrasse 7, 82205 Gilching, Germany Phone +49 (0) 8105 733 84 0, Fax +49 (0) 8105 733 84 33 Find us in he World Wide Web a: www.brainproducs.com To repor errors, please send a noe o markeing@brainproducs.com General Managers: Dr. Achim Hornecker, Alexander Svojanovsky Copyrigh 2008 by Brain Producs GmbH Noice of Righs All righs reserved in he even of he gran of a paen, uiliy model or design. For informaion on geing permission for reprins and excerps, conac markeing@brainproducs.com. Unauhorized reproducion of hese works is illegal, and may be subjec o prosecuion. Noice of Liabiliy The informaion in his press release is disribued on as As Is basis, wihou warrany. While every precauion has been aken in he preparaion of his press release, neiher he auhors nor Brain Producs GmbH, shall have any liabiliy o any person or eniy wih respec o any loss or damage caused or alleged o be caused direcly or indirecly by he insrucions conained in his book or by he compuer sofware and hardware producs decribed here. All Brain Producs hardware and sofware devices are for research purposes only, no for clinical use! Conens IN THE FOCUS Erich Svojanovsky: A ribue for many years service in he field of naional and inernaional research in Pyschophysiology Analyzer 2 in pracice An inroducion o he ICA in he Anlyzer 2 way EASYCAP New Cap Shape: AsiaCu beer fi, beer daa New Produc Imporan safey aspecs for combined MR recordings New Produc Brain Producs launches long awaied GSR module for fmri Produc Developmen BrainVision Recorder 1.03.0005 Produc Developmen acicap ConrolSofware II Brain Producs producs in real life Taking Brain-Compuer Inerface (BCI) research ono he rack. Inside Naional Regulaory Affairs Bes pracice Taiwan! Deparmen of Healh (DOH) User Research BOLD Correlaes of coninuously flucuaing epilepic aciviy isolaed by independen componen analysis. BP Young Scienis Award And he Award goes o Dr. Cornelia Kranczioch BP produc in real life Our small conribuion o a Gold Medal a he Olympic Games in Beijing Cerified Dealers Isvan Roll: Farewell and Thank You Cerified Dealers The company Qualiis New Brain Producs dealer in Hungary Conference Promoion And he Winners of 2222$ are Ediors Pinboard Brain Producs in he Public Eye BP Inside New Team Members: Dipl. Psych. Parick Briz, Inernaional Sales & Suppor Dr. Davide Riccobon, Produc & Risk Managemen News in Brief Downloads, Programms and Updaes for Analyzer BrainVision Workshops & Exhibiions

page 2 of 19 IN THE FOCUS Erich Svojanovsky: A ribue for many years service in he field of naional and inernaional research in Pyschophysiology by he Brain Producs Team During he 34h Conference of he German Associaion for Applied Psychophysiology (DGPA, held in Magdeburg from July 15-17h 2008) an official ribue was made o Erich Svojanovsky for his naional and inernaional achievmen in he field of Psychophysiology. Prof. Dr. Paul Pauli, Chairman of he Associaion said: As a resul of your iniiaives, your own moivaion and openness o new scienifically based ideas, ogeher wih your developmens and producs, Psychophysiology has reached a poin a which is effeciveness has been significanly enhanced. Erich Svojanovsky has devoed himself for over fory years o neurophysiological research. His own innovaions, his undersanding of he requiremens of neurophysiological research and is applicaion, his many years of coninuous, ireless engagemen in problem solving and especially his incomparable alen for making he correc decisions a he righ ime have all enabled him o provide a simulus which has had a considerable influence in he field of brain research. The honorific speech, o which Prof. Dr. Paul Pauli, Prof. Dr. Erich Schroeger and Prof. Dr. Chrisoph Herrmann all conribued, included he following words: The German-speaking conribuion o Psychophysiology makes i now one of he leaders in he inernaional field; your producs and your suppor have enabled hundreds of publicaions o see he ligh of day. You are responsible for he careers of a few of us presen here. Erich Svojanovsky was an iniiaor of counless producs, such as acousic and visual simulaors, DC Research amplifiers and oher diverse programmes for aquisiion and analysis. Wihou him he company, Brain Producs, would never have come ino being; wihou him here would no have been he BrainVision Analyzer or he BrainAmp MR. He was responsible for furhering research and creaing livelihoods which, in urn, opened up oher avenues of research. Many people hroughou he world was influenced, helped and suppored by Erich Svojanovsky, alhough i is cerainly he case ha mos of hem do no know i because he never pu himself forward; he was oo busy looking afer wha he had creaed, devoing ireless effor o he compleion of he ask in hand, and bringing abou innovaion. Because of his fair-minded and seady applicaion, he was held in high eseem and affecion, even amongs his compeiors, boh as a human being and as a businessman. In he name of everyone whom he encouraged, influenced and suppored in he course of his career, we would like o express our profound hanks o Erich Svojanovsky for his remarkable services o Psychophysiology. As long as he has srengh and happiness, as long as he knows ha his acions bring joy o ohers, he will remain in business for a long ime. I was said of him a he conference: Your rapid response o problem solving is even more well-known han your red pullover. When answering he quesion as o how he would envisage he period following Erich Svojanovsky, he replied: The company is in good hands. We wish him coninued joy in his creaive effors; heir impac will reach a wider audience and live a longer life han his red pullover. Erich Svojanovsky is an inspiraion o us, o he field of research and o all hose who know him.

page 3 of 19 Analyzer 2 in pracice An inroducion o he ICA in he Anlyzer 2 way The ICA ransform is as easy as a walk in he park by Parick Briz, Sales & Suppor In he las Press Release you heard a lo abou he new Analyzer 2 wih is easy-o-use inerface, enhanced workflow as well as is exended capabiliies. Today I will show you how his affecs your daily work by aking a closer look a one of he numerous ransforms in Analyzer 2. I have chosen he ICA ransform (Independen Componen Analyses) for his purpose, no only because i is such a versaile and widely used mehod bu also because i is an analyzing sep in which a lo of exper knowledge is considered necessary. How do you spo an exper ICA user? Well, i would seem ha an exper has he abiliy o ell a a glance: This IC is clearly a myogen arifac, his IC is eye-aciviy and his one looks like a real EEG. This leads us o he quesion: how do you become an Exper ICA user? Firsly, of course, i s imporan o read he relevan informaion and o undersand he background of he mehods, bu mos imporanly, you have o ge a feel for he ool. Play around wih your daa and possibly someone else s. I know he word play can sound like somehing requiring no effor and he ICA ransform is definiely no considered o be he easies hing o come o grips wih. In fac, i has somewha of a repuaion for being very complex and a drain on a compuer s resources - bu le s see. Below I will ell you how o really explore your ICs using Analyzer 2. Bu firs we should ake a quick look a he idea of he ICA ransform and briefly discuss seing up an ICA analysis o improve performance. The ICA ransform is used o spli he EEG signals up ino independen componens using informaion heory mehods. I is assumed here, ha EEG signals are a linear combinaion of independen componens (independen as defined in informaion heory). The purpose of he ICA ransform in wha is known as blind source separaion is o reconsruc source signals from a combinaion of hese signals. Boh he source signals and he combinaion are unknown, so assumpions are made abou he signals and how hey are combined. I is assumed ha he signals o be reconsruced are saisically independen of each oher and accordingly, he ICA ransform is a purely saisical mehod. I is based on he above assumpions and does no use any addiional physiological informaion. No, i does no have any informaion abou he channel locaions nor does i know anyhing abou head shape. Therefore, i never ceases o amaze me o see he resuls being so inuiively linked o physiological processes or arifacs. The resul of he ICA ransform is a se of componens ha are defined analogously o EEG channels in he ime domain. In addiion, he ICA ransform provides a ransformaion marix by means of which he componen acivaions can be calculaed from he channel daa. A more deailed descripion of he mehod and all he parameers you can se for your ICA ransform can be found in he Analyzer manual. The manual is definiely worh reading and I promise i is clear and easy o undersand. You will be able o find ou: Why he fas ICA is no always faser han he Infomax ICA (because he fas ICA requires daa wih good separabiliy of componens in order o be fas) and wha s he difference beween he various opions you can selec while seing up your analysis (e.g. some ICA algorihms are deerminisic and some are no and will herefore lead o slighly differen resuls every ime you run hem). In addiion he manual will guide you nicely hrough all he seps you need in order o se up an ICA analysis which I don have ime or space o alk abou now. Once we ve addressed he heoreical elemens we ll ake a quick look a he seps needed o run an ICA. There is only one oher hing we need o discuss before ha: he daa on which you wan o run he ICA, because ICA resuls srongly depend on he qualiy of he daa. The ICA ransform is somewha like a person saring a new job. If you rain someone for a job you ensure hey re rained on relevan hings and ge a sufficien amoun of raining. If he person doesn ge enough raining or if hey are rained incorrecly hey will almos definiely do a poor job. Is here any wonder, you hink. Well he ICA has exacly he same resricions. So wha informaion does your ICA need o know? I has o be EEG composed ou of he sources you like o separae and wihou oher informaion. Tha means you have o avoid severe or a ypical arifacs. This is because he ICA would ry o process his daa and because he arifac is no usually relaed o oher daa some of your IC s would only ry o explain he arifac. I is easy o see how his compromises your resuls. In simpler erms, even hough he ICA is a prey good mehod i does no overcome he GiGo rule (garbage in garbage ou). Le s say we have daa from a session where we firs recorded a resing EEG (relax, open your eyes, close your eyes) followed by a very demanding ask (a speeded reacion ime ask). If we now rain our ICA on he firs 100 seconds of our daa i will consequenly fail o indenify he myogen aciviy of he ensed neck muscles even hough i could be a very prominen aciviy, simply

page 4 of 19 because his kind of aciviy was no presen in he resing EEG which we used o rain he ICA. Here he conclusion is clearly ha having he correc daa o rain your ICA wih is crucial. When i comes o he amoun of daa one should use for he ICA, he Analyzer 2 suggess aking he inerval of 0 o 1 second. You ve probably already figured ou ha his value is no inended o be a defaul value. Now you know ha Analyzer 2 brings good defaul values wherever his is possible. However, his isn as simple for he ICA, le me explain. There is a rule of humb regarding how much daa you should give o he ICA ransform in order o obain he bes resul. For up o 64 channels he rule is: daa poins per channel should be he number of channels squared and hen muliplied by 20, for more han 64 channels he rule is: daa poins per channel should be he number of channels squared and hen muliplied by 30. However, his rule only works as long as as long as your sampling rae corresponds roughly o he specral conen you wish o decompose. To be more precise, he relaion is given by he Nyquis- Shannon Theorem (or WKS-Sampling-Theorem). For our case he heory basically saes ha you have o sample a leas wice as fas as your highes frequency. If he fases frequency you are ineresed in is around 100 Hz, you would sample your signal a somehing like 250 Hz. However, if you sample a 5000 Hz, hen he fases frequency conained in he signal is 2500 Hz and hus 25x higher han he 100 Hz maximum you are really ineresed in. In shor, he ICA ransform depends on informaion and jus because you sample your EEG wih a higher frequency here is no more informaion in i. Also, please noe ha EEG is a ime based biological signal and ha he above rules of humb do no conain any sample rae correcion facor. A 250 Hz sample rae, he rule of humb for 64 channels would give you (64*64*20)/250 or 327 seconds worh of daa. However, if your sampling rae is 5000 Hz, hen he above would give you only 16.3 seconds worh of daa. Ask yourself wha will happen if you rain your ICA on only 16.3 seconds worh of daa... Firs, you should mark he arifacs and hen ensure you have enough daa. Be aware ha hese wo hings ogeher can cause problems. The Analyzer 2 does no ake arifacs for he raining of he ICA. This means ha if you wan o rain your ICA on 82 seconds of daa and your chosen inervals are 30 seconds marked as arifacs, your ICA will lack a lo of raining, which is no a good hing. Finally wih all hese precauions in mind, selec he ICA from he Transformaions menu. Make all he relevan selecions. Afer you give he ICA ransform a sufficien amoun of good daa (please selec semiauomaic mode), sar he ICA and ake a shor ea or coffee break. Even hough he new Analyzer 2 is much faser han is predecessor, he ICA ransform will sill need some ime. When you reurn from your break, you ll see a window like he one shown in figure 1. The firs hing I would do is give your IC s some room. You have already read abou how o do his in he las newsleer. And now, le s have a play around! So firs I ll describe he window our playground. Then I ll inroduce some of he ools you can use and finally we ll explore some ICs and see wheher we can ell anyhing abou hem a a glance. Ok, le s sar wih he window. The firs hing o cach he eye is he IC aciviy on he lef side of he window bu please ake a look a he righ hand side firs. A he op you ll see he lis of your ICs. If you click on one of hem i becomes blue and now a he boom you ll see he IC map. By he way, I know you probably expec o be able o cusomize he map jus like any oher map in Analyzer 2 and of course you re righ, you can do so he usual way (by Righ-Click, choose Seings). If you double click on he green field in he firs column i will urn red I will ell you wha his is for in a minue. If you look a bi furher down here is a dropdown menu which you see in figure 2. The menu conains Figure 1

page 5 of 19 Figure 2 spend some ime on hem. he following four iems: ICA- Componens, Correcion, Subraced and Topographies. These four views allow you o look a your ICs in a really exhausive manner. So le s informaion you are used o seeing i wih your bare eyes - i can easily be idenified by he ICA. Moreover, i is likely o be he firs IC on your lis of IC s, as is he case in figure 3. Analyzer 2.0 has a more sophisicaed way of dealing wih ocular arifacs, ha is, ICA ransform ocular correcion. Though, he advanages and resricions menioned above are basically he same for his ransform. If you selec ICA Componens, he componens are displayed firs, as shown in figure 1. Now you can use he ransien view ool o find ou he specra of your ICs. Mark a ime inerval by righ-clicking FFT from he ransien views. You will ge a frequency specrum of he componens. This is where Analyzer 2 s real power can be seen. Slide he marked inerval over your daa. The Analyzer 2 will coninuously display he changing frequency specra of he chosen ime poin as you slide over he daa. By he way, if you have wo screens on your compuer you can even undock windows like he FFT-ool and drag i ono your second screen o have a beer view of your daa. Do you wan o see your correced EEG wih and wihou his or ha suspicious componen? If you selec Correcion, he reconsruced daa se is displayed wihou he componens highlighed in red. If you wan o view he componen in he correced EEG or no, or see he EEG wih or wihou a componen - i is jus a click away. Bu wha was he original EEG like? Check overlay wih original daa and you will be shown how your daa will change. Easy, righ? And wha if I like viewing wha informaion I m going o remove from my daa? If you selec Subraced, he reconsruced daa se is displayed using only he componens highlighed in red. See how he subraced daa changes depending on wha componens are checked or unchecked. Overlay his wih he original EEG o see wheher you ve removed he correc informaion. Now do you undersand why I alked abou playing around? Ok, here s one more bi of imporan informaion ha describes an IC. If you selec opographies, he associaed map view of all componens is displayed. As you can really find a scaling facor o fi all maps I recommend displaying normalized maps (check he box in conrol panel). This view is very helpful o ge a firs general impression. This paern of figure hree s hird (F02) and fourh (F03) opography wih heir fronal and clear opposing aciviy make hem candidaes o conain he aciviy caused by horizonal eye movemen. Because he ocular arifac is so prominen and wih such disinc However, he simple approach of jus looking a he ICs can lead o very good resuls. Figure 4 shows he correced EEG overlaid wih he original daa (in red). Wha do you hink prey good huh? Bu we wan o become expers, so le s alk abou some oher, no so easy o spo arifacs. Someimes i jus so happens ha an elecrode is prepared badly, or some elecrodes may no funcion properly. I have heard ha he laer does happen occasionally. Anyhow, he resul of his will be ha one of your channels conains mainly noise. Because his noise is no linked o any EEG informaion, he ICA will make a separae componen ou of his channel. A bad channel can easily be found on he ICA Maps. If you find a map wih only one channel acive, well, you guessed i. Bu wach ou, if you look a a spline-inerpolaed map, a bad channel on he edge of he cap does look a bi like a componen. Therefore i is someimes helpful o use riangulaion as inerpolaion o display he opographies. Le s come back o he experimen we used o discuss our subjec. Possibly here is severe muscle aciviy in our daa we can spo immediaely. Bu wha abou daa where we can see he arifac? Does his mean ha here is none? Well, we should assume ha he arifac is here. Now we can make hings easier and filer i ou, however, if cases where we are ineresed in higher frequencies, filering is no an opion. So, wha can we do? Are you sill wih me, or are you already looking a he FFT of your ICs? Afer deermining he ICs, which have high power in he higher frequency bands, you check heir opographies. You are looking for componens, wih raher local aciviy on he edges of he cap. Wih an infra-cerebral-cap i s ofen quie amazing. You can even connec he IC aciviy o a specific muscle. Moreover, he opography reflecs he orienaion of he elecrical field he muscle produces. Ok, by now you re almos an exper bu le s search for one more kind of arifac. You usually do no see any arifacs from he hear in your EEG. I m afraid he bad news is ha here is always some ECG relaed aciviy in your EEG. This ime even he filer opion won save you. So, how do you find an IC wih ECG-aciviy? As he Figure 3 F00 F01 F02 F03

page 6 of 19 hear-relaed-aciviy has in many cases no clear opography you have o look a some oher informaion. Le s scale he IC aciviy so you see a leas 20 seconds. This will make i easier o find aciviy wih such a low frequency. And of course, look a he FFT ransform of he ICs (o ge a finer frequency resoluion choose a wide inerval (e.g. 4 second) o slide over he ICs). Is here an IC wih a Peak in he range of 0.7 o 2 Hz? I have o confess, ha I Figure 4 do no always find his kind of aciviy. Clearly having more han 32 channels, le s say 64 channels, helps a lo. I lef his ou when I discussed he heory of he ICA ransform wih you. However, he ICA assumes ha here are more channels han sources. If his assumpion is violaed, he ICA has o mix he aciviy of some sources ino one IC. The direc consequence is ha you should record enough channels if you wan o obain really good resuls wih he ICA ransform. By now we have looked a some ICs and we have seen how o idenify some arifacs wih a few clicks and by aking a quick look a hem. Clearly his is no he only hing he ICA ransform can do in fac, ha s jus he beginning. The Analyzer 2 provides you wih all he informaion you need o calculae any kind of saisic on he componens, or on he Marices he ICA ransform uses o calculae he ICs from he EEG and he EEG form he ICs. Of course, I have no covered everyhing here is o know, however I hope you ve now go enough enhusiasm o wan o explore your daa wih he new Analyzer 2 ICA ransform. And if you like working wih i, I have some good news for you: We are enhancing he ICA ransform even more. Say uned for a surprise... EASYCAP New Cap Shape: AsiaCu beer fi, beer daa by Falk Minow, EASYCAP General Manager The basic componen of all he caps in he EasyCap, BrainCap, acicap family has always been one and he same cap emplae. Alhough he anaomy of he human head varies significanly among individuals, he fi has generally been very saisfacory, as long as he caps are worn by subjecs wih a European head shape. This is differen for Asian subjecs: In he pas, we ascerained ha no only did he caps no fi sufficienly well, bu here was also a quesion as o wheher he elecrodes sill sa over he correc cerebral areas, for example did he C3/C4 elecrodes sill record he aciviy in he sensorimoor cenre wih sufficien precision? In order o provide a saisfacory cap shape also for our Asian cusomers, we measured he head shapes of more han 60 Asian voluneers (32 male and 29 female) in cooperaion wih our Asian disribuors. For comparison we measured 22 European voluneers (13 male and 9 female). The following mehod was applied: The 10/20 posiions were defined in relaion o he anaomical landmarks nasion, inion and audiory canals. Taking hese as our saring poin, we operaionalised all he dimensions required for a cap emplae as inermediae disances beween hese anaomical poins and beween he elecrode posiions derived from hese. The cap secions were correspondingly alered for hose variables wih significan deviaions beween he Asian and European mean values. As a resul, he caps adjused o he Asian head anaomy received op marks in he pracical esing, and hus no furher correcion was required. This new AsiaCu cap model does no only conribue o a beer fi and comfor for subjecs, we also hope for clearer, even more accurae daa. The AsiaCu is now available for all EASYCAP and Brain Producs caps. We would like o ake his opporuniy o sincerely hank our disribuors for all heir suppor!

page 7 of 19 New Produc Imporan safey aspecs for combined MR recordings by Alexander Svojanovsky, General Manager, and Dr. Davide Riccobon, Produc & Risk Manager Le us sar his aricle wih a rip ino he pas, a ime when he BrainAmp MR was no available. Only a few scieniss dedicaed heir research ino evaluaing he echnical possibiliies of performing simulaneous EEG & fmri recordings. As far as we know, he firs aemps o combine boh echniques a he same ime were published in he lae 80s. (von Klizing, L.: Do saic Magneic Fields of NMR Influence Biological Signals?, Clin. Phys. Physiol. Meas. 7, 157-160 1986, von Klizing, L.: Saic magneic fields increase he power inensiy of EEG of man, Brain Research, 483 201-203 1989). The firs working amplifier was placed inside he scanner room was developed exclusively for he inended use of performing combined recordings, and again as far as we know, his was consruced in he early 90s by K. Krakow, L. Lemieux, P. Allen. Anoher device, an analog pre-amplifier sysem o ransfer EEG daa acquired in he scanner o a conneced amplifier ouside he shielded room was developed by J. Ives. A US paen Apparaus and mehod for recording an elecroencephalogram during magneic resonance imaging. And was filed 1995 (and wen hrough a re-examinaion process). An addiional paen in his field filed 2004 FMRI compaible elecrode and elecrode placemen echniques deals wih he invenion of analog EEG daa ransfer because of he inabiliy o use digial circuis in he design, due o broadcas inerference from inernal clock circuis. Boh paens do need separae amplifiers ouside he shielded room, or a leas, way ouside he MRIscannerbore. Since he beginning of combined EEG & fmri recordings, he saic magneic field srengh of MRI scanners wen up from 1.5T o 7T and even higher. Scanner sequences have been opimized and he RF power has increased. In he year 2000, Brain Producs released is firs fmri compaible digial amplifier wih opical daa ransmission o he recording compuer ouside he scanner room. By ha ime, recording EEG in an inerleaved mode was he common procedure. Pulse arifac removal was he firs problem and was resolved by implemening a correcion module based on he average subracion mehod ino he BrainVision Analyzer sofware. (Idenificaion of EEG evens in he MR scanner: he problem of pulse arifac and a mehod for is subracion, Allen PJ, Polizzi G, Krakow K, Fish DR, Lemieux L., Neuroimage. 1998 Oc;8(3):229-39.) In a shor amoun of ime, he scienific benefis of combining boh mehods were demonsraed and increased demand o expand he possibiliies. Brain Producs released improved models of he BrainAmp MR amplifier which allowed running coninuous fmri sequences and enhanced correcion echniques were added. (Gradien correcion, ICA, Clock synchronizaion). Since ha ime, i is possible o ransfer EEG lab paradigms ino he scanner and o benefi from boh echniques in a real simulaneous way. Numerous papers have been published evaluaing he mehods and subsaniaing he benefis of he simulaneous EEG & fmri echnique (e.g. S. Debener, A. Lavric, D. Brandeis, C. Muler, ec.). The plaform o a new scienific area was creaed and researchers around he world are now able o shif heir paradigms ino he scanner and o profi from he early days of EEG & fmri invesigaion. I is he righ ime for us o offer our experience o make your recordings safe and o avoid such hings as burning your voluneers for example. Figure 1: High resoluion MPR of a subjec wih EEG elecrode assemblies applied on he scalp. The MR-compaible elecrodes cap did no compromise he corical signal. (Lef side Abralye HiCl wih Supervisc, righ side Abralye 2000) Jürgen Brinkmeyer, Rheinische Kliniken Düsseldorf Kliniken der Heinrich-Heine-Universiä Düsseldorf As menioned above, he echnique of combined EEG & fmri recordings has been explored and improved over he las couple of years. Since he release of our firs producs for his purpose we have improved our devices in erms of qualiy, handling and safey. This means herefore ha he elecrode cap and elecrodes, which are conneced direcly o he subjec, have been coninuously changed due o our experience in his field.

page 8 of 19 BrainCap MR - he new design Safey Elecrodes mus no ouch he skin. Our new elecrodes are pin ype sensors which are placed inside a plasic holder mouned on he cap. The holder is filled wih gel o reduce skin conducance and o esablish a conac beween he sensor and he subjec s skin. Every elecrode conains safey resisors beween he sensor and he connecion wire. Connecion beween he componens is per-formed by gluing, no soldering. Addiional safey resisors are placed inside he cap connecor, acing like an addiional RF-filer. Wires are locaed on he ouside of he cap o ensure isolaion beween skin and wire according o FDA paien safey regulaions (hp://www. youube.com/wach?v=7vkpja1brng). High emperaure isolaing ubes around he ECG elecrode cable avoid conac beween skin and wire. ECG elecrodes do conain higher resisors han normal elecrodes o compensae for he echnical characerisics of longer wires. All wires are fixed ino he cap o avoid loops. Wire lengh from elecrode o he amplifier inpu is fixed o a minimal required lengh. Wire oules for he cable ree a cenral posiions avoid loops due o cable rouing. Exra elecrodes (for e.g. EMG recordings) are surrounded by ubes o avoid conac beween wire and skin as well as o avoid cable loops occurring by misake. Serial number sicker for easy racking and suppor/service issues. Image Arifacs The new pin sensor design reduces he amoun of Ag/AgCl in he cap. Reducing he number of exra elecrodes hanging ou of he cap reduces possible image arifacs. Wire oules for he cable ree a cenral posiions ensure shor cable rouing disances hrough he head coil. Comfor Exra fla elecrode holders for exra comfor, especially when resing on elecrodes. Exra spare elecrode holders for exra comfor o compensae for gaps beween elecrodes especially when using less han 64-channels. Sof cap fabric o increase comfor and o widen he fi o accommodae differen head sizes. Serviceabiliy Serial number sicker for easy racking, assembly and layou quesions. Colored elecrode holders for easy access/clarificaion of assembly affiliaion (e.g. 10-20, 10-10). Name labels on every elecrode for easy recogniion. Special gel adds elecrode visualizaion capabiliies (similar o viamin E capsules), as shown in figure 1. Wha happens o old elecrode caps ha do no mee he curren safey sandards? If you are sure ha your scanner equipmen has no changed and ha scanner sequences and relaed procedures are exacly he same as when you firs received your equipmen, here is no urgen need o change your caps. Oherwise, afer a couple of years of use i may be a good idea o adap your equipmen o he laes safey sandards. Please verify if your scanner has been upgraded or sequences have been added by he manufacurer. The innovaive BrainCap MR sandard will be available for all radiional cap sizes and assembly from Ocober 1s, 2008. We ask cusomers o wach ou for fuure announcemens: special offers for exchanges awai you! Addiional safey insrucions abou scanner sequences, posiioning of equipmen and prevenive measures can be found in our new BrainAmp MR manual also available soon. I is worh reading o avoid damaging your equipmen or huring subjecs. Trus our experience, make your recordings safe. Safe recordings can only be assured by combining our BrainCap MR, BrainAmp MR and by adhering closely o our safey insrucions. Please find safey insrucions in he BrainAmp MR manual.

page 9 of 19 professional BRAINCAP MR 5 10 11 8 1 Sof fabric 2 Ear sli 3 Colour coded elecrodes 4 Lable on wire 7 6 2 4 3 9 5 Wires ouside cap, no loops 6 Tubes around wire 1 7 No direc conac sensor-skin 8 Exra fla adaper 9 Exra fla adaper 6 10 Oule for comfor & safey 11 Shor safey connecion 7

page 10 of 19 New Produc Brain Producs launches long awaied GSR module for fmri by Dr. Rober Sörmer, Technical Suppor Skin conducance (SC) has for decades been one of he mos employed measures in psychophysiologic research. As sensiive parameer for emoional and cogniive saes, sress and pain SC has also been widely used in psychiaric research. Wih growing availabiliy of EEG for funcional MRI, we ve seen increasing demand for he esablished psychophysiologic parameers in addiion o EEG and funcional imaging. A he very op of he wish lis came EDA/GSR. The iniial marke survey revealed ha here are already some so called MRI compaible GSR devices on he marke: These consis in essenially MR incompaible amplifiers, placed in he conrol room wih elecrodes sreched o reach he subjec hough a wall breakhrough. Based on our EEG/fMRI experise, we are convinced ha he concep of shor resisive elecrode wires maches safey requiremens beer han he compeiive producs. As we are a EEG company, we know abou EMC and noise problems caused by galvanicaly conneced devices locaed ouside he scanner room. Based on his consideraions we sared he developmen of a compeely RF poof sensor for use inside he scanner. Our MR amplifier developmen eam decided o se up an exremely compac bu heavily shielded DC insrumenaion amplifier measuring conducance direcly using.5v consan volage. Due o he insrumenaion amplifiers high CMRR properies, he gradien arifac is already r e m a r k a b l y suppressed so ha in many cases smoohing is sufficien o ge a laboraory-like GSR signal. The sensor inerfaces wih he bipolar BrainAmp ExG MR via he ExG AUX inerface box and he GSR is recorded synchronously wih he EEG. EMC and IEEC 60601-1 are confirmed by exernal esing laboraories. The new sensor has been exensively esed in various 3T scanners. The firs publicaions have already been submied. I is now possible o add he GSR sensor o your already exising BrainAmp ExG sysem and record GSR in laboraory-like qualiy even during funcional MRI. Produc Developmen BrainVision Recorder by Dr. Rober Sörmer, Technical Suppor The Recorder s main module as well as all amplifier specific componens have undergone a general overhaul during he las monhs. This includes fixes, general usabiliy enhancemens, Visa x32 suppor and driver adjusmens. Recorder main module, all amplifiers: Added: Toggling of display filers online Added: Parameer of AUX-channels sored in header file Added: Visa x32 suppored USB inerfaces (BrainAmp family, V-Amp) Added: Confirmaion promp before workspace is saved Fixed: Workspace now correcly displayed for Asian OS Fixed: Offse of AUX-channels correced Fixed: Trigger debouncing for very low sampling raes Fixed: Monages suppor >64 Channels BrainAmp driver and ACC: Added: 250Hz low pass for BrainAmp AUX-channels (MR!!) Added: SyncBox saus logged on-change, SyncSaus indicaor Added: BrainAmp DC-Correcion by TTL rigger Fixed: Impedance view now shows correc elecrode idenifier for bipolar channels Fixed: Correc elecrode labels for bipolar impedance readings (vhdr) QuickAmp driver (New Recorder Version for QuickAmp will be released laer): Fixed: Some muliprocessor PCs caused QuickAmp o sop in very rare cases Fixed: No adminisraor privileges needed any more V-Amp / FirsAmp driver: Fixed: Marker offse a 20kHz and oher minor changes The complee insallaion package is available for regisered users from our websie.

page 11 of 19 Produc Developmen acicap ConrolSofware II by Bernd Pich, Technical Coordinaor Research & Developmen Free bu no for nohing. This pun describes he new acicap Conrol Sofware which is immediaely available o download from our websie or is available on CD. In addiion o a compleely redesigned menu display which is self-explanaory, paricular aenion was paid o he hree caegories: Edi, Elecrodes LED and Impedance measuremen. These caegories have been fundamenally changed. The faciliy in Edi Elecrodes was exended so ha single elecrodes can be inroduced insead of groups of 32. This faciliaes he possibiliy of generaing subjec-specific conrol groups which are no subdivided ino 32 sages. In he folder Elecrodes LED, all LEDs of he elecrodes and all hree colours can be esed manually or auomaically. The clearly defined visual arrangemen makes i possible o deermine which elecrode is being riggered a he seleced locaion. The Elecrode group acivaion is also new. This display makes i possible o conrol and bring single elecrodes ino groups wih jus one mouse click. The individual groups are displayed in differen colours in line wih he user s requiremens. I should also be menioned ha he impedance measuremen has been compleely revised and offers improved auomaic measuremen of impedance ogeher wih a clear display of he impedance values. Brain Producs producs in real life Taking Brain-Compuer Inerface (BCI) research ono he rack. Poenial and challenges for increasing raffic safey. by Sebasian Welke, Cener of Human-Machine Sysems, GRK promeei; Deparmen of Psychology and Ergonomics, Chair of Human-Machine Sysems, Team PhyPA, Technische Universiä Berlin, Germany The year 2007 saw a number of 335.845 injured and 4949 faaliies as a resul of raffic accidens in Germany [1]. In comparison o he las decade, he decreasing number of injured and faaliies indicaes ha vehicle-safey research has proven o be very effecive. Neverheless, his good rend suffers from he consan high number of raffic accidens and also from he fac ha here are sill faaliies. Consequenly, he necessiy for a new holisic approach in vehicle-safey research becomes eviden. Research in he area of vehicle-safey and Brain-Compuer Inerfaces (BCIs) is conduced by several researches of he TU Berlin. The projec, presened in his aricle, is relaed o projecs of he research raining group promeei, he chair of Human-Machine Sysems (HMS) and he Team PhyPA. The research raining group promeei develops and inegraes mehods, procedures and ools in order o invesigae he human-machine ineracion already in he early developmen phases of echnical sysems. This group is siuaed a he Cener of Human- Machine Sysems. The chair of HMS focuses on he analyses and design of human-machine sysems wih special emphasis on he invesigaion of human informaion processing and percepion. Team PhyPA is par of he chair HMS and aims a he ransfer of echnologies from BCI research ino he conex of human-machine ineracion. One possible approach ha is focused in his work-environmen is an absoluely holisic driver cenred poin of view. Subjec wih 32-Channel acicap by Brain Producs The idea is o use he driver s inenion o improve and rigger exising drive-dynamic and driver suppor sysems. Therefore menal processes before an acion, paricularly movemen preparaions and decisions, can be idenified as a possible source for such a rigger-suppor sysem. In his manner he elecroencephalogram (EEG) allows projecing menal processes o measurable elecrical recordings which reflec brain aciviy wih a precision in he range of milliseconds. So, changes in he EEG-daa could be examined which are induced by inended acions (e.g. emergency braking)

page 12 of 19 and also decision-making processes during driving. Drive-dynamic exreme siuaions are a maer of paricular ineres, since hey represen pre-crash condiions o some degree. This ambiious exreme EEG-projec deals wih possibiliies and challenges using an EEG-based Brain-Compuer Inerface (BCI) by PhyPA for deecing he driver s inenion and behaviour in drivedynamic exreme condiions and pre-crash siuaions respecively. The major goal is he robus discriminaion of EEG paerns correlaed o differen menal processes in his conex. Differen neuropsychological consolidaed findings are suiable as a possible source for projecing he absrac erm of driver s inenion o feaures in EEG daa. Baseline - subjec on fron passenger s sea double lane change The quesion is wha kind of feaures and mehods can be used o correlae EEG wih psycho-physiological findings? For feaure selecion a lo of well known aspecs of human informaion processing and movemen preparaion can be applied. The accompanying brain aciviy can be examined by even-relaed poenials (ERPs) and even-relaed desynchronisaions (ERD). In he field of he ERPs he following anicipaing poenials seem o be a useful pool for approaching he represenaion of he driver s movemen inenion in EEG daa: This way, curren BCIs allow conrolling devices by discriminaing or classifying differen menal saes. Firs, hese saes could be inenional houghs, e.g. movemen imaginaion, mahemaical calculaions or menal roaions, as used in acive BCIs (see [3] for a definiion). Second, hese saes could be aspecs of informaion processing, movemen preparaion and planning as used in passive BCIs as well as aspecs of associaion. Boh ypes of BCIs offer a lo of poenial for applicaions in he conex of driver suppor for increasing raffic safey and for decreasing he number of accidens. Challenges In conras o EEG recordings in laboraories, recordings under real life condiions are a big challenge. This leads o special requiremens regarding he EEG measuremen, especially he qualiy of daa. For single rial paern recogniion in he EEG daa, machine learning is used. According o machine learning heory, he machine has o be rained on differen movemen relaed EEG paerns in a raining phase [4]. Especially in his phase, he machine learning approach suffers from aspecs of overfiing on he recorded raining daa. EEG daa ypically consiss of underlying cogniive processes bu also coincidenal noise paerns and arefacs. This addiional informaion has o be removed from he recorded EEG daa before he BCI can be rained. For a real life BCI applicaion an approach mus be developed which no only allows o remove arefacs by e.g. eye movemens, eye blinks, cardiac signals and muscle noise, bu also drivedynamic influences. Experimens In a firs sep, direc movemens (e.g. braking, gearing and seering) of he driver will be considered. Therefore, differen real life experimens were carried ou. To examine wheher i is in fac possible o measure EEG while driving in exreme siuaions, and o idenify robus EEG paerns o planning and possibly decision processes in he brain. Coningen negaive variaion (CNV) and Pre-moor or readiness poenial (BP). Aciviies of neural srucures are also accompanied by specral variances in EEG signals (ERD) and hus suiable as feaure for movemen relaed cogniive processes. Brain areas of ineres are: Pre-moor area (PMA), Primary moor corex (M1), Supplemenary moor area (SMA) and Associaion areas. The applicaion of mehods from he area of saisical machine learning for paern recogniion and feaure selecion based on ERD has proven o be very pracical and efficien in BCI research [2]. Inside he race car Renaul Clio II RS CUP by RenaulSpor

page 13 of 19 While driving in he real world beside he EEG he longiudinal and laeral acceleraion, speed, engine speed, seering angle, acceleraor posiion, emperaure, ec. as well as muscle movemens (neck) are measured. This offers he possibiliy o decompose EEG recordings for arefac removal. The independen componen analysis (ICA) offers one soluion for revealing for insance muscle arefacs by sabilisaion of he head in EEG daa [5]. The second sep deals wih aspecs of decision-making and approaches o projec sraegies and inenions o exraced feaures based on EEG-daa. Therefore, differen real life and also simulaor EEG-experimens will be carried ou. Resuls which are validaed will be presened soon. BCI research combines knowledge and mehods of psychology, mahemaics, compuer science and engineering. This accenuaes he muli disciplinary of his research domain. This holisic approach presened here enables he embedding of menal processes ino applicaions of human-machine ineracion. As saed in his aricle, he applicaion of ideas from machine learning combined wih susained neuropsychological knowledge is a promising approach for implemening driver adapive suppor. The usage of a BCI for predicing he driver s movemen inenion can help o rigger well sophisicaed drive-dynamic and driver suppor sysems for increasing raffic safey and o avoid faaliies. References: [1] Saisisches Bundesam Deuschland. Unfälle und Verunglücke im Sraßenverkehr 2007 [2] B. Blankerz, F. Losch, M. Krauleda, G. Dornhege, G. Curio, and K. R. Mueller. The berlin brain-compuer inerface: Accurae performance from firs-session in bci-naive subjecs. IEEE Trans. Biomed. Eng, 2007. [3] T. O. Zander, C. Kohe, S. Welke, and M. Roeing. Enhancing human-machine sysems wih secondary inpu from passive brain-compuer inerfaces. In: 4h In. BCI Workshop and Training Course, 2008. [4] Richard O. Duda, Peer E. Har, and David G. Sork. Paern Classificaion (2nd Ediion), Wiley-Inerscience, November 2000. [5] T.-P. Jung, S. Makeig, C. Humbphries, T.-W. Lee, M.-J. Mckeown, V. Iragui, T. J. Sejnowski (2000). Removing elecroencephalographic arifacs by blind source separaion. Psychophysiology, 37, 163 178. Exreme EEG-Projec Equipmen: BrainAmp DC 32-channel acicap BrainAmp ExG AIM EVO3 Pisa Daalogging (8 configurable channels, CAN and infrared beacon) Racecars: Renaul Clio II RS CUP Renaul Megane CUP Sreecar: VW Touran Suppored by: Brain Producs GmbH HFC Human-Facors-Consul GmbH Deparmen of Psychology and Ergonomics, Chair of Human- Machine Sysems, Team PhyPA Technische Universiä Berlin, Germany ADAC Fahrsicherheiszenrum Berlin Brandenburg Conac: sebasian.welke@u-berlin.de Inside Naional Regulaory Affairs Bes pracice Taiwan! Deparmen of Healh (DOH) by Anne Vogel, Regulaory Affairs Manager DOH - Taiwan s equivalen of he U.S. FDA proved o be a very posiive exempion from he usual process of naional produc regisraion. Alhough all impored medical devices mus obain a regisraion cerificae from heir Bureau of Pharmaceuical Affairs (BPA), we received he paper in he shores amoun of ime in black and whie aesing ha our producs are approved. Taiwan s medical device marke is he 4h advanced marke in Asia and i coninues o be increasingly sophisicaed. Thank you for he excellen suppor of our disribuor Gez Bros & Co. (BVI), especially Penny Lin!

page 14 of 19 User Research BOLD Correlaes of coninuously flucuaing epilepic aciviy isolaed by independen componen analysis. Jann e al., Neuroimage. 2008 Aug 15;42(2):635-48. Epub 2008 May 8. by Kay Jann, Universiy of Bern, Universiy Hospial of Psychiary, Deparmen of Psychiaric Neurophysiology In epilepsy he correc diagnosis of he syndrome and he idenificaion of he irriaive zone are of grea ineres since i is necessary for correc reamen and evenually surgical inervenion. This irriaive zone is no only involved in he epilepic crises, bu also produces flucuaing inerical aciviy. Ierical aciviy flucuaes sponaneously, has ypical EEG signaures, and is assumed o mosly resul from aciviy in sharply delimied regions. Therefore, a mehod ha deecs random evens wih high emporal and spaial resoluion is needed. This is provided by simulaneous EEG/fMRI recordings. Combined EEG/fMRI has indeed been demonsraed o be a valuable ool o delineae he irriaive zone defined as he area wih alered BOLD during epilepiform aciviy (Krakow e al., 1999 or Salek- Haddadi e al., 2006). The idenificaion of epilepiform aciviy has so far been based on subjecive selecion of suspicious evens, simplifying he epilepiform aciviy o an on/off phenomenon. While his has been successful in a reasonable amoun of cases, his approach is cerainly loosing sensiiviy because i subjecive, does no ake ino accoun he variance in he lengh and ampliude of he single inerical epilepic discharges (IED) and canno accoun for inermien rhyhmic discharges. Moreover, subhreshold IEDs are no accouned for. Our sudy has herefore used an Independen Componen Analysis (ICA) o idenify epilepiform aciviy coninuously. The decomposiion of he EEG ino ICA componens separaes normal physiological EEG aciviy from componens represening arifacs (e.g. eyeblinks, scanner noise, cardioballisogram, ec.) and, mos imporan for our purpose, i isolaes facors represening epilepiform aciviy (Iriare e al., 2006; Jung e al., 2000). The emporal evoluion of hese epilepiform ICA facors can hen again be correlaed wih he fmri BOLD signal. This has he advanage ha he epilepiform ICA facors include also subhreshold epilepiform aciviy and accoun for he lengh and he ampliude of separae IEDs. Moreover, he ICA based idenificaion eliminaes he need o manually selec single evens, which is cumbersome and error-prone. We expeced ha his should increase he sensiiviy of he mehod as compared o he convenional approach using a binary (on-off) predicor for he BOLD signal. Tweny paiens wih differen ypes of focal and generalized epilepsy syndromes were invesigaed in a 3T Siemens Magneom Trio MR scanner. Simulaneously a 92 channel EEG was recorded wih 3 MR compaible BrainAmp amplifiers. To avoid aliasing of he scan pulse arifac, he EEG sampling was locked o he clock of he MR scanner using he Brain Producs SyncBox (Mandelkow e al., 2006). Addiionally, a second EEG daase was recorded ouside he MR environmen. MR scan pulse arifac correcion was performed according o he average arifac subracion (AAS) mehod described by Allen e al. (2000). Aferwards he EEG recorded ouside he scanner and he MR scan pulse correced EEG were combined and decomposed using an exended infomax ICA algorihm (Delorme and Makeig, 2004). Thereby he ouside recorded EEG provides a baseline for he ICA algorihm o learn he unconaminaed EEG. This improved he separaion of facors and yielded a small se of ICA facors wih lile specral ampliude ouside he scanner and markedly increased ampliude inside he scanner. This is he behavior ha we expeced for facors relaed o scanner and cardioballisic arifacs. Two experienced neurophysiologiss seleced he facors coding for epilepiform aciviy based on heir emporal dynamics (aciviy a he same imepoin as IEDs presen in he original EEG) and heir load on he elecrodes (e.g. scalp disribuion). The facor raed as mos accuraely meeing he crieria was considered for furher analysis. The absolue ampliude of he seleced ICA facor was convolved wih a double gamma hemodynamic response funcion (HRF) o accoun for neurovascular coupling. This convolved signal used o compue voxelwise correlaions wih he BOLD signal. As covariaes moion in six direcions (ranslaion and roaion along he X, Yand Z-axis) were used. To validae our hypohesis ha our more physiological predicor may increase he sensiiviy, we also analyzed each paien wih he convenional approach. The ICA separaed arifacs from physiological and pahological aciviy in each paien. In seveneen of weny paiens, he BOLD correlaes of epilepic aciviy were in accordance wih he suspeced EEG sources, he clinical semiology and, if presen, he srucural lesions. In hose cases ha were clinically equivocal, he BOLD correlaes helped o esablish he proper diagnosis of he underlying ype of epilepsy. In one paien who suffered from emporal lobe epilepsy, he affeced hippocampus could be idenified by he BOLD correlaes of rhyhmic dela aciviy. This is a case where he convenional approach ypically fails. Overall, we could demonsrae ha he ICA based approach increased he sensiiviy for proper laeralizaion in our very heerogeneous group of paiens wih diverse ypes of epilepsies form 50% (convenional approach) o 80% (ICA approach). The ICA EEG/fMRI approach is a safe, non-invasive and easily applicable echnique, which can be used o idenify regions wih alered hemodynamic effecs relaed o IEDs as well as inermien rhyhmic discharges in differen ypes of epilepsy.

page 15 of 19 References Allen, P.J., Josephs, O., Turner, R., 2000. A mehod for removing imaging arifac from coninuous EEG recorded during funcional MRI. Neuroimage 12, 230-239. Delorme, A., Makeig, S., 2004. EEGLAB: an open source oolbox for analysis of single-rial EEG dynamics including independen componen analysis. J Neurosci Mehods 134, 9-21. Iriare, J., Urresarazu, E., Arieda, J., Valencia, M., Levan, P., Vieri, C., Alegre, M., 2006. Independen componen analysis in he sudy of focal seizures. J Clin Neurophysiol 23, 551-558. Jung, T.P., Makeig, S., Weserfield, M., Townsend, J., Courchesne, E., Sejnowski, T.J., 2000. Removal of eye aciviy arifacs from visual evenrelaed poenials in normal and clinical subjecs. Clinical Neurophysiology 111, 1745-1758. Krakow, K., Woermann, F.G., Symms, M.R., Allen, P.J., Lemieux, L., Barker, G.J., Duncan, J.S., Fish, D.R., 1999. EEG-riggered funcional MRI of inerical epilepiform aciviy in paiens wih parial seizures. Brain 122 ( P 9), 1679-1688. Mandelkow, H., Halder, P., Boesiger, P., Brandeis, D., 2006. Synchronizaion faciliaes removal of MRI arefacs from concurren EEG recordings and increases usable bandwidh. Neuroimage 32, 1120-1126. Salek-Haddadi, A., Diehl, B., Hamandi, K., Merschhemke, M., Lison, A., Frison, K., Duncan, J.S., Fish, D.R., Lemieux, L., 2006. Hemodynamic correlaes of epilepiform discharges: an EEG-fMRI sudy of 63 paiens wih focal epilepsy. Brain Res 1088, 148-166. Graphical overview of analysis procedure and an example of resuls for one paien wih emporal lobe epilepsy.

page 16 of 19 BP Young Scienis Award And he Award goes o Dr. Cornelia Kranczioch (Universiy of Porsmouh) by Sefanie Rudrich, Markeing Several hundred scieniss aended he annual meeing of he German Sociey for Applied Psychophysiology (Deusche Gesellschaf für Psychophysiologie und ihre Anwendung; DGPA) Psychologie und Gehirn which ook place a he Universiy of Magdeburg from May 15h - 17h, 2008. And jus as in previous years he DGPA presened hree awards for young scieniss - one of which was sponsored by Brain Producs. In he evening of May 15h, Professor Paul Pauli (Presiden of he DGPA) and Alexander Svojanovsky (Brain Producs General Manager) presened he Brain Producs Young Scienis Award for a Disinguished Conribuion o EEG research o Dr. Cornelia Kranczioch. Dr. Kranczioch is a senior lecurer a he Universiy of Porsmouh, UK. Her research work focuses mainly on behavior and brain dynamics inheren in visual awareness. She also works on ineracions beween visual and moor sysem in humans. Las year she became Head of he Affecive and Cogniive Neuroscience Research Group a Porsmouh s Deparmen of Psychology. More recenly, she received an offer from he Universiy of Jena, making i more likely ha she will be back in Germany very soon and will coninue her EEG and fmri work. Dr. Kranczioch has received he Young Scienis Award 2008 for her paper Temporal Dynamics of Access o Consciousness in he Aenional Blink, published in he journal NeuroImage (vol 37, 3, 947-955). She and her colleagues sudied even-relaed poenials and oscillaory brain aciviy in order o beer undersand why physically idenical informaion is someimes consciously perceived and someimes los o awareness. They found ha conscious percepion is heralded by changes in coherence in he alpha and bea frequency ranges as early as several hundred milliseconds before he relevan simulus is even presened. They sugges in heir paper ha hese changes are crucial in deermining access o consciousness and for he disribuion of processing resources amongs compeing informaion. The picure shows Dr. Kranczioch holding he Brain Producs crysal cup wih her son Jonahan Arie. In addiion o he cup, he award also includes a sum of 1.000 for research rips and a nice cerificae. Congraulaions once again! Brain Producs producs in real life Our small conribuion o a Gold Medal a he Olympic Games in Beijing by Pierluigi Casellone, Sales & Suppor Gaze and posural seadiness paired wih srong concenraion are he main goals ha any ahlee compeing in Air Rifle shooing should pursue. A fac ha he very alened Indian ahlee Mr. Abinhav Bindra and his coach knew very well. Therefore Mr. Bindra decided o ake a combined recording by using he Simi Moion Sysem for moion analysis and he Brain Producs equipmen for EEG acquisiion during his las raining camp in Germany. Mr. Bindra during he recording session wih Pierluigi Casellone (Brain Producs), Andi Russ (Simi Moion) and Dr. Ferdinand Tusker (TU Munich). Mr. Bindra perfecly knew ha small imperfecions in his performance could be deeced and correced only wih he help of an insrumenal invesigaion. The daa recorded from he wo sysems were merged already during he acquisiion, allowing o analyze and correc Mr. Bindra s performance in realime. Looking a he resul Abhinav Bindra has jus achieved a his year s Olympic Games in Beijing winner of he firs Olympic Gold Medal in he hisory of India we can be very proud of he possible help we have provided him wih. Brain Producs congraulaes Mr. Bindra for he grea vicory and hanks him for having aken a bi of Brain Producs o he Olympic podium.

page 17 of 19 Cerified Dealers Isvan Roll: Farewell and Thank You by Alexander Svojanovsky, General Manager Some people have o reire o make us aware of how indebed we are o hem. Tha is no he case wih Isvan Roll, beer known as Oecsi. We ve always known we would miss him grealy - as a person and as a business parner (see he aricle Cerified Dealers ) - even before his reiremen.. He was wih us for many years as a colleague in sales in Hungary and was incredibly popular wih his cusomers. Bu he does no like using he word cusomer because his cusomers are his friends. Tha fis in well wih our philosophy and is surely a reason why he many years of collaboraion wih him proved o be so enormously producive, naural and unique. Oecsi Roll undersands research, sees soluions and implemens hem wih his inimiable charm and in doing so is able o draw on hiry years experience. He is a one off - boh as a person and in he field of business. I go o know him personally and o value him highly many years before our collaboraion. I would describe him as somebody of he old school, of honourable and high characer, refined in an incredibly naural way, bu a he same ime asserive in a friendly way, always acing in a rusing, responsible manner and open o fuure developmens. He even carefully prepared his own deparure from working life ino a well deserved reiremen; his planning showed sensiive consideraion for all concerns. Oecsi Roll seleced his successor wih grea care and nurured him over he years. He will be available o us and his successor as an advisor for a furher year so ha he ransiion is as smooh as possible. Afer ha we hope he will be able o enjoy he fruis of his endeavours for a long ime o come. We would like o hank him for his suppor over many years and for he excellen collaboraion we have enjoyed. Oecsi Roll s reiremen represens a grea loss for everyone who has ever worked wih him; and a loss for everybody who has no had he pleasure of working wih him. All he bes o our friend, who is mos welcome here a any ime o inspire us wih his presence. Cerified Dealers The company Qualiis New Brain Producs dealer in Hungary by Pierluigi Casellone, Sales & Suppor When Mr. Isvan Roll informed us abou his decision o sep down from business and enjoy a well deserved reiremen, we were all very sad. No only were we going o say goodbye a grea friend of ours bu also a very compeen business parner. Neverheless, our concerns abou fuure business in Hungary were very quickly resolved by Mr. Roll himself, who already had a proposal for he successor: The company Qualiis. This company, which in he meanime had also became our cerified dealer, was esablished some years ago by a group of young professionals: Mr. Nagy, Mr. Solmár and Mr. Tóh. The hree young and dynamic founders and heir colleagues offer producs and cusomer services in differen fields: Conracual sysem adminisraion for small and medium-sized companies, inerne services as well as sales and suppor for biomedical equipmen for neurophysiology. Mr. György Tóh, who is in charge of Qualiis biomedical branch, saed: The company Qualiis brings op qualiy services and producs o is cusomers by possessing high grade experise in all fields in which we are involved. We are very proud o be he exclusive Hungarian dealer of Brain Producs, as hey sand for sae-of-he-ar echnology in he neurophysiology research field. We are very confiden ha his business relaionship will be a grea success for many years o come. (Qualiis is he proud sponsor of he Hungarian Himalaya Expediion since 2006 - he company flag has been brough up o he 8047 meer high Broad Peak on he 11h Augus 2007. Now we can lierally say ha wih he Qualiis company, Brain Producs is on op of he world )

page 18 of 19 Conference Promoion And he Winners of 2222$ are by Sefanie Rudrich, Markeing Wha a hrill On June 23rd, he brand-new version of our analysis sofware, he BrainVision Analyzer 2, was finally released. As we waned you o join in our exciemen wih his big sep, we ran a special promoion in he conex of his year s ISMRM (Torono / Canada) and Human Brain Mapping (Melbourne / Ausralia) giving he aendees a chance o win 2,222 Dollars. A boh conferences we se up a safe and invied he conference paricipans o ry and crack i by seing a four digi code of heir choice. Alhough hundreds of aendees ried heir luck in Torono (May 3rd - 8h), none of hem managed o find ou he combinaion ha opened he safe. Neverheless we waned o make sure ha he money wen o good use for research and science and herefore decided o donae he 2,222 Canadian Dollars o he ISMRM Suden Sipend Fund. ISMRM represenaives Jennifer Olson (Associae Execuive Direcor) and Sandra Daudlin (Direcor of Meeings) happily acceped he check on behalf of he ISMRM (visi www.ismrm.org/08/bp_gran.hm). The money will now be used o enable more young sudens, posdocoral and clinical rainees o aend he Annual ISMRM Scienific Meeing & Exhibiion and presen heir absracs here. Abou one monh afer he ISMRM, we ran he same promoion a he HBM in Melbourne / Ausralia (June 15h -19h). And his ime here was a lucky safe cracker! Mr. Bjoern Scheler from he Freiburg Cener for Daa Analysis and Modeling se he correc code of 2789, opened he safe and won 2,222 Ausralian Dollars o be spen on research purposes. Of course we d like o ake his opporuniy o congraulae Mr. Bjoern Scheler again on being he lucky winner of our Analyzer 2 Promoion: Well done, Bjoern! From he lef: Pierluigi Casellone (Brain Producs Sales Manager), Sandra Daudlin (ISMRM Direcor of Meeings), Dr. Ingmar Guberle (Brain Producs Scienific Suppor Manager), Sefanie Rudrich (Brain Producs Markeing Assisan), Jennifer Olson (ISMRM Associae Execuive Direcor). From lef: Dr. Achim Hornecker (Brain Producs General Manager) and Bjoern Scheler (Universiy of Freiburg) - he lucky winner of 2,222 Ausralian Dollars. Ediors Pinboard Brain Producs in he Public Eye by Sefanie Rudrich, Markeing Again and again, Brain Producs and is producs are menioned in various newspaper or magazine aricles. To give you an example: Wihin he las few monhs GEO kompak (German Magazine), Focus (German Magazine) and BBC News published picures and / or aricles demonsraing how our producs work in real life. To ge an updae on hese as well as furher publicaions, please visi he Media Coverage secion on our websie a: www. brainproducs.com/references.

page 19 of 19 Brain Producs Inside New Team Member: Dipl. Psych. Parick Briz, Inernaional Sales & Suppor by Alexander Svojanovsky, General Manager Saring from June 2008, Dipl. Psych. Parick Briz sared working a our sales and suppor eam. Parick is an experienced user of our producs undersanding scienific needs and requiremens, formerly working a he Universiy of Trier, Deparmen of Psychophysiology. Brain Producs Inside New Team Member: Dr. Davide Riccobon, Produc & Risk Managemen by Alexander Svojanovsky, General Manager Saring from April 2008, Dr. Davide Riccobon is responsible for he coordinaion of produc developmen as well as risk and qualiy managemen in accordance wih applied harmonized sandards. Davide is an experienced user, formerly working a he Ludwig Maximilians Universiy Munich, Deparmen of Psychology. News in brief: Downloads, Programms and Updaes for BrainVision Analyzer and Recorder *08-01-2008 BrainVision Analyzer 2: Updae for LORETA Transform (Version 2.0.0.2801) 08-21-2008 BrainVision Analyzer 2: Second bugfix for OS wih asian fons 09-01-2008 BrainVision Recorder: Version 1.10. For more informaion please visi our websie a www.brainproducs.com/news.php * News in brief: BrainVision Workshops BrainVision Analyzer User Workshop on EEG / fmri Bangor (Wales), Sep. 17h / 18h, 2008 BrainVision Analyzer User Workshop Shanghai (China), Sep. 22nd - 24h, 2008 BrainVision Analyzer User Workshop on EEG / fmri Cardiff (Wales), Oc. 23rd / 24h, 2008 For more informaion and regisraion please visi our websie a www.brainproducs.com/workshops.php News in brief: Exhibiions *ICON X. Bodrum (Turkey), Sep. 1s - 5h, 2008 IOP 2008 S. Peersburg (Russia), Sep. 8h - 13h, 2008 5h ISNIP & ECNS Join Conference Frankfur / Main (Germany), Sep. 10h - 13h, 2008 4h Inernaional BCI Workshop Graz (Ausria), Sep. 19h / 20h, 2008 5h ISNIP & ECNS Join Conference Göingen (Germany), Oc. 1s - 5h, 2008 Annual Meeing of he SPR Ausin (USA), Oc. 1s - 3rd, 2008 Furher informaion on hese exhibiions is available on our websie a: www.brainproducs.com/evens.php