Brain Producs Press Release Conens Issue N 3/2006 IN THE FOCUS IN THE FOCUS: Brain Producs acicap booss road safey research Brain Producs Inside Brain Producs New General Manager User Research Scienific American Mind Brain Producs Raffle V-Amp Raffle News in brief Downloads, Programms and Updaes News in brief Workshops & Exhibiions more on page 2 more on page 2 more on page 3 more on page 3 more on page 3 more on page 3 User Research Using a combined ERP/TMS proocol o invesigae corico-corical ineracions in aenion more on page 4 Brain Producs Inside Job opporuniies @ Brain Producs more on page 4 Produc Developmens Analyzer Professional goes Educaion more on page 4 Did you know...? MR Arifac correcion in coregisered EEG daa (1) more on page 5 Workshops BrainVision Advanced User Workshop Gilching (Munich), Ocober 4-6, 2006 more on page 8 Brain Producs GmbH office@brainproducs.com www.brainproducs.com Zeppelinsrasse 7 82205 Gilching (Munich) Germany Phone +49 (0) 8105 733 84-0 Fax +49 (0) 8105 733 84-33 VAT ID: DE812343586 General Managers: Alexander Svojanovsky, Dr. Achim Hornecker HRB AG München 118521
page 1 of 7 IN THE FOCUS: Brain Producs acicap booss road safey research by Dr. Rober Sörmer, Sales & Suppor Driving accidens caused by drowsiness and micro sleep or inaeniveness due o oversraining are one of he mos frequen causes of accidens in indusrialized counries. Leading car manufacurers herefore bank on new developmens such as driver assisance and driver warning sysems o enhance raffic safey. Neurophysiology based approaches have he poenial o provide a significan conribuion for he developmen of hese sysems. Currenly, he German Federal Minisry of Educaion and Research paronizes he large scaled FaSor research projec o invesigae new approaches in drivers sae predicion and workload miigaion by novel driver assisance sysems. The FaSor consorium represens he cumulaive research power of four op-class research insiues along wih he Daimler-Chrysler AG and Brain Producs GmbH as indusrial parners. The FaSor projec is embedded wihin he MicroSysems Technology (MST) framework. Expeced resul is he developmen of he SLIM-aciCap: The SLIM-aciCAP will combine key feaures such as acive elecrodes wih inegraed impedance conversion, inegraed impedance display and acive shielding. To mach specific needs of on-road or field EEG applicaions, he SLIM-aciCAP will come in an exremely small form facor design. Brain Producs has provided he enire EEG-equipmen (128-Ch. Video EEG sysem) for he firs pilo es series. In a pilo es he research group a DaimlerChrysler performed he firs measuremens wih he acicap in a es vehicle, which provided very good and promising preliminary resuls. However, robus high qualiy EEG acquisiion is only one opic wihin FaSor. Our parners FhG FIRST (Inelligen daa analysis group), Dep. of Neurology a Campus Benjamin Franklin (Charié, Neurophysics Group), TU Berlin, (Sofware Engineering Research Group) and idalab GmbH (Machine learning dep.) presened encouraging resuls during he firs official FaSor projec review. The FaSor projec is scheduled o run over hree years. We will coninue reporing he progress of his projec. If you are planning similar sudies, please feel free o conac our echnical suppor any ime. Key words: on-road EEG, acive elecrodes, SLIM acicap, driver assisance, driver warning, real ime feaure exracion, single rial analysis. Brain Producs Inside Brain Producs New General Manager by Dr. Achim Hornecker, General Manager When I sudied mahemaics and compuer science a he universiy of Freiburg/Germany, I cam in firs conac wih signal processing algorihms and saisical mehods. In 1997 Henning Nordholz invied me o join he new eam of programmers consrucing he firs version of BrainVision Analyzer for he new founded company named Brain Producs. My firs asks were ransforms like Filers and FFT, where a mahemaical background was needed. Laer I ook over more and more programming asks for he BrainVision Analyzer. In July 2006 Henning Nordholz lef Brain Producs and I became new general manager, responsible for Developmen. So I wen more and more from programming o managemen asks. From ime o ime I suppor Ingmar Guberle, eaching he mahemaical lessons of he BrainVision Analyzer Advanced User Workshop. This helps me o exercise my mah and o keep in conac wih our cusomers.
page 2 of 7 User Research Scienific American Mind by Naoko Ono, Markeing Scienific American Mind is a quarerly publicaion focusing on he funcion of he mind and brain. Vol. 17, N 4, Aug/Sep. 06 conained a very ineresing aricle Coming o Aenion edied by Andreas K. Engel, Sefan Debener and Cornelia Kranczioch abou how he brain decides wha o focus conscious aenion on. Find an amazing experimen a: hp://viscog.beckman.uiuc.edu/djs-lab/demos.hml Brain Producs Raffle V-Amp Raffle by Naoko Ono, Markeing Our V-Amp raffle his year had an overwhelming response worldwide and again showcased our level of appreciaion owards our cusomer! Mrs. Naasha Mauris from he Universiy of Groningen (NL), won he fabulous weekend-rip o New York. Enjoying he vibran and urban amosphere of his meropolis Mrs Mauris saed a he very end of her rip: we had a grea ime in New York. The luxurious hoel was convenienly locaed in mid Manhaan, beween Fifh Avenue, Times Square and Cenral Park, allowing us o ge around New York s ouris sies easily. We loved he lighs of he Financial disric as seen from Brooklyn! As par of he prize we enjoyed a lovely dinner a he 44 Resauran and go ickes o he laes Disney musical Tarzan. The laer was a breah-aking specacle: we promised ourselves o go and see i again someime, somewhere. Alogeher, his surprise weekend in New York Ciy fel as a complee holiday! We wan o give our special hanks o all paricipans of he V-Amp raffle 2006! Your Brain Producs Team News in brief: Downloads, Programms and Updaes by Dr. Rober Sörmer, Sales & Suppor BrainVision Analyzer BrainVision Analyzer 1.05.0004 (many inernal revisions; BrainVision Analyzer Educaion Ediion and all previous revised modules now inegraed) New BrainVision Analyzer Modules Brainsar Reader 1.05.0001 (added suppor for Brainsar Forma 4) GenericDaaReader 1.05.0004 (Bugfix) InsEP Daa Reader 1.05.0003 (unlimied version) Tes 1.05.0002 (improved) ViaporReader 1.05.0003 (bugfix) BrainVision Soluions Creae MAT File 1.0 (firs release) Slice2Volume Trigger 1.3 (Bug-Fix; now flexible marker names) Percen Change 1.3 (Bug-Fix; suppor ERD/ERS ransformaions added) Wrie Markers 1.5 (much faser, now drop down liss for marker names) ICA Backransform 1.3 (suppors missing channels, improved marice deecion) ICA Topographies 1.6 (suppors missing channels, improved marice deecion) CBC Parameers 1.0 (suppors now CBA-delay and CBA-ICA-componens) Read Markers 1.1 (delees also exising markers) Slice2Volume Trigger 1.2 (sops now afer n volumes) Please noe: For easy and complee mainenance of your BrainVision Analyzer insallaion we recommend (1) he insallaion of BrainVision Analyzer 1.05.0004, (2) he relaed New Modules for Analyzer 1.05.0004 and (3) Soluions for Analyzer 1.05. All componens are freely available on our websie for all regisered users. acicap acicap Operaing Insrucion 1.0 acicap Benuzerhandbuch 1.0 acicap Conrol Sofware 1.0 acicap CAP monages acicap 32 ch acicap 64 ch acicap 128 ch News in brief: Workshops & Exhibiions by Anja Egger, Markeing Mee us a he MEDICA 2006! Brain Producs will be exhibiing a his year s MEDICA in Düsseldorf saring November 14h unil November 18h. New producs such as he acicap and he BrainVision Analyzer Educaion will be showcased. Moreover, we will be available for discussion and o mee you personally. We would like o welcome you in hall 9, booh N A42 we are looking forward o seeing you!
page 3 of 7 User Research Using a combined ERP/TMS proocol o invesigae corico-corical ineracions in aenion by Marin Eimer, School of Psychology, Birkbeck College, Universiy of London, UK I is generally acknowledged ha he aim of cogniive neuroscience o idenify he neural basis of cogniive funcions in he human brain canno be achieved on he basis of a single mehod. Developing join-mehod approaches where differen measuremen echniques wih complemenary srenghs are combined has herefore become a major focus in neuroscience. Transcranial magneic simulaion (TMS) and EEG/ERP are ideal candidaes o be combined in invesigaions of he emporal and funcional organisaion of cogniive processes. TMS can be used o selecively disrup circumscribed corical areas wih high spaial as well as emporal precision, and ERPs can be employed as on-line measures of he effecs of his disrupion a differen processing sages, and wihin brain regions ha are anaomically remoe from he area disruped by TMS. Previous combined TMS/ERP sudies (1) were primarily concerned wih echnical and mehodological issues, and in paricular wih he massive EEG arefac ha is ineviably induced by TMS. This arefac, which conaminaes EEG/ERP measures, and can resul in amplifier sauraion, is especially problemaic when using experimenal proocols wih inerleaved TMS-EEG/ERP recordings (e.g., when single-pulse TMS is applied in close emporal viciniy of sensory evens in order o sudy he impac of TMS on evoked componens). While i is possible o use purpose-buil EEG amplifiers and blocking procedures o preven sauraion, his approach can lead o he loss of EEG daa wihin criical ime windows, and induce oher problems such as DC baseline shifs. We have recenly demonsraed ha he use of such blocking procedures is no required when using BrainAmps DC amplifiers in an inerleaved TMS/ERP proocol (2). We sudied he neural basis of visual aenion in a visual search ask where single-pulse TMS was delivered over righ poserior parieal corex or he verex. Because EEG acquisiion was coninuous in his experimen (i.e. wihou blocking ), EEG waveforms included a brief TMS arefac induced by he magneic pulse. This was removed by cuing ou 40 ms segmens (from 2 ms prior o TMS onse o 38 ms afer TMS onse) from EEG waveforms for all TMS rials and elecrodes. EEG daa poins before and afer each removed segmen were hen simply joined. Alhough his procedure induced a random volage sep beween he joined daa poins for single rials, his was compleely eliminaed by EEG averaging. To furher remove conribuion of TMS-induced audiory and somaosensory aciviy, EEG recorded on oher rials where TMS was presen wihou concurren visual simulaion was averaged and hen subraced from ERPs on visual search rials wih TMS, similar o he procedure described in (1). ERPs were hen filered using 0.01 Hz high-pass, 40 Hz low-pass, and 50 Hz noch filers. Afer applying hese procedures, we demonsraed ha righ parieal TMS no only resuled in delayed response imes o visual arges during conjuncion search, bu also in he eliminaion of he early phase of he N2pc componen (reflecing he focusing of aenion ono arge locaions) over he righ hemisphere when TMS was applied here. In conras, he early N2pc was presen when TMS was delivered o he conrol sie (verex). Our sudy suggess ha righ parieal TMS can inerfer wih aenional seleciviy in remoe visual areas, and also demonsraes how TMS and EEG can be combined in he sudy of cogniive funcions. 1. Thu G, Ives JR, Kampmann F, Pasor MA, and Pascual-Leone A. A new device and proocol for combining TMS and online recordings of EEG and evoked poenials. J Neurosci Mehods 141: 207-217, 2005. 2. Fuggea, G., Pavone, E.F., Walsh, V., Kiss, M., & Eimer, M. (2006). Corico-corical ineracions in spaial aenion: A combined ERP/TMS sudy. Jornal of Neurophysiology, 95, 3277-3280. Brain Producs Inside Job opporuniies @ Brain Producs by Alexander Svojanovsky, General Manager We are always ineresed in adding value o our company and o increase compeence wihin our eam. As a general precondiion you should have experience using our producs and you mus be capable of working independenly and anicipaory. Creaiviy, friendliness and high moivaion o serve our users is required o join our eam and o improve he processes relaed o your daily work wih your alens. Wheher your background is echnically or scienifically, if you have programming or adminisraive skills, whenever you feel ha you fulfil he above aribues/abiliies and you are looking for a career a Brain Producs, please send your shor resume (English or German) o office@brainproducs.com Produc Developmens Analyzer Professional goes Educaion by Alexander Svojanovsky, General Manager In he previous ediion of our press release we have been announcing he new educaion BrainVision Analyzer Educaion Ediion. Now we are offering you a special educaion package conaining: - Analyzer educaion (30 nework licenses) - Recorder educaion - FirsAmp EEG/ERP amplifier Complee educaion package price: 9.950,00 Ediion For more informaion please conac your local disribuor, sales@brainproducs.com or visi www.brainproducs.com
page 4 of 7 Did you know...? MR Arifac correcion in coregisered EEG daa (1) by Dr. Ingmar Guberle, Scienific Suppor Manager For his and he nex wo issues of our Press Release I will depar from he rodden pah and will no highligh one or wo Analyzer Soluions per se. Insead I will be alking abou some daa analysis problems ha made me wrie hese soluions. Le s see where his akes us When Timing maers Many of you will already have seen EEG daa recorded in he MR scanner and he arifacs he consan swiching of he gradiens and of he RF pulses inroduces ino he EEG daa. Depending on he scanner and on he MR sequence and proocol used, he resuling EEG daa will range from slighly obscured o oally indiscernible due o he massive arifacs he scanner sysem inroduces ino he EEG recordings. If you look a he figure 1, you can easily see some characerisics ha are imporan o know and undersand for he correcion of hese arifacs. - For one hing, he ampliudes of he gradien arifacs can be enormously high, wih ampliude values ofen being well wihin he range of he 13000 µv shown below and hus in excess of 100:1 compared o normal EEG. Anoher hing ha is easy o see from he figure below is he enormous speed wih which he volages develop over ime. In his example, he slopes reach speeds of 25000 µv/ms and more, which means ha he gradiens are more han 500x faser han normal sponaneous EEG. Thus, i is only logical ha we can no even see or disinguish he EEG daa ha does of course ride on op of Fig. 1: Typical MR Gradien Arifac during coregisered EEG hese curves. So, how can one even dream of correcing hese arifacs in he EEG daa? Well, he erm ha saves he day is echnical. These arifacs are generaed by he swiching of he moveable gradien fields and by he RF pulse generaed in he scanner bore and since he generaion of he MR or fmri images demands ha hese gradien fields and RF pulses are given wih exreme accuracy emporally wih regards o magniude, we can expec for wo acquisiions of he same slice or volume o generae he exac same echnical arifacs in our EEG daa. Thus, echnical in his sense means invarian and invarian means easily modeled and hopefully also easily correced. However, one prerequisie for being able o ap ino his echnical naure of hese arifacs is ha each of here consecuive occurrences is also recorded invarianly by he EEG amplifier and while he firs impulse is o say of course and o nod his off, his is no so clear. Time, relaive ime and phase shifing in EEG/MR daa Wha a ile, bu don se your sighs o high. I am no going o inroduce a new concep of ime (alhough i someimes feels like I should). Insead, wha I wan o alk abou are he iming requiremens and poenial iming relaed problems when recording EEG daa in he MR environmen. I also wan o alk abou how o idenify and solve hese problems or criical issues wih he help of Soluions. Bu, le s sar from he beginning These days nearly all MR scanners can oupu a TTL signal a he exac ime hey sar a new Slice or Volume acquisiion and he BrainAmp can of course read and record hese even markers. These markers save us a lo of hassle, since each marker indicaes he exac ime poin a which he (echnically invarian) arifac begins and we can use his marker o guide us in building our correcion emplae across all consecuive volumes. Bu wha if he marker is no always se correcly a he very same ime poin wihin he volume s ime course? Impossible? No, sadly, his is quie possible and common, alhough he likelihood of such problems is drasically differen beween scanner manufacurers. And no, I will no menion scanner or manufacurer names, bu insead I will show you how o diagnose his effec and hen how o remedy i. OK, so wha would perfec iming beween consecuive iming mean? Well, firs and foremos i would mean ha all imes beween one volume (or slice) sar marker given by he scanner and he nex such marker would be compleely invarian. If we were o measure hese inermarker imes and would draw hem as virual ampliudes agains heir own marker-omarker progression, his Fig. 2: User Inerface of he Marker Timing Soluion
page 5 of 7 would give us a sraigh line. This ype of analysis can be done wih he Analyzer Soluion Marker Timing (figure 2), which I originally wroe for doing hear-raevariabiliy analyses where one would also wan o draw he ime beween each consecuive pair of wo markers (R-Markers) as an ampliude value. Variabiliy in he hear rae would hen show up as clearly visible deviaions of he line from being sraigh across ime. The same holds rue here and all we need o se in Marker Timing are our arge marker ype and name. The soluion hen reads in he markers and generaes a small daase wih he iming of our MR measuremens, as shown in he figure 3. Fig. 3: Impac of he synchronisaion on volume marker iming There are wo races in his figure, one from a measuremen wih perfec iner-volume iming and one from a more ypical nonsynchronized EEG/MR measuremen. As can easily be seen in he figure, he red race for he perfecly imed measuremen is compleely fla, hus showing no variabiliy in he iming beween consecuive MR volumes a all. The black race on he oher hand does show period gliches of iming, so le s have a closer look a hese. The firs hing we noice is ha here are only wo iming saes in he picure even for he nonsynchronized recording. One ime is 3000.00 ms and he oher is 3000.20 ms, which accouns for he gliches, bu le s see where hey come from Recording EEG daa inside he MR scanner means ha here is EEG being acquired by our BrainAmp MR a a sample rae of 5000 Hz, which makes he iner-daa-poinime 0.2 ms and needless o say he clock used o drive he amplifiers is and has o be exremely sable and accurae. By he way, his high sample rae is only required in his environmen in order o laer be able o deec and correc he enormously fas arifacs he scanner inroduces ino our daa. Afer, or raher during he gradien correcion, he daa can be downsampled o more convenien daa raes. A he same ime of he EEG recordings here is of course also daa being acquired by he MR scanner and as I already said above in doing so he scanner has o swich a series of moveable magneic gradien fields on and off and also excies he issue wih a leas wo srong RF pulse rains during his ime in order o ulimaely creae he MR or fmri images we know. Since he qualiy of he images derived (among oher hings) depends direcly on he accuracy wih which he iming and srengh of hese magneic and RF pulses can be delivered, scanner sysems generally also have excellenly sable and accurae gradien clocks. Ok, so we have wo excellen clocks, now does his mean all s well ha ends well and I res my case? Clearly no, because all of his does no mean ha he onse of each fmri volume o be acquired mus coincide wih he exac ime poin a which a BrainAmp MR daa poin is being acquired. - This can also be seen as a phase shif beween he iming of he wo devices and of course i is his phenomenon ha accouns for he 0.2 ms iming variabiliy we saw above in our Marker Timing display. Plainly saed, for some MR Volumes, he BrainAmp does no see he Volume Sar marker being se unil is nex daa poin acquisiion, because i had jus finished recording one daa poin, while in some oher insances i migh jus be acquiring he nex daa poin when he marker is being oupu by he scanner. Bu, hold on, hese markers acually denoe a ime poin a which he scanner sars some aciviy, ypically for he acivaion of some gradien swiching such as he slice selec gradien and if he BrainAmp jus misses he insan of he seing of he marker by a iny fracion of a millisecond, hen ha means ha i also sars recording he gradien aciviy of his paricular volume wih he nex daa poin only. Needless o say his is imporan because i means ha he marker as he BrainAmp sees and records i is no necessarily perfecly aligned wih he onses of he MR volumes and heir EEG borne arifacs. - By he way, if his offse or phase shif was consan over ime, e.g. because he Volumes are sared in a iming raser ha is evenly divisible by he 5000 Hz of he BrainAmp clock, hen his effec would no maer a all, because each par of he gradien arifac he scanner sequence inroduces ino he EEG daa would always be recorded a he exac same emporal offse from he volume sar even marker and would hus always have he very same ampliude level a each daa poin being acquired by he BrainAmp. This shows ha he wo clocks do no have o ick a he same ime, bu a consan ick phase for synchrony. However, his is normally no he case. The figure 4 shows wha happens if his is no he case. Alhough he gradien arifac clearly has he same shape and also he same ulimae minimum and maximum ampliude, is ampliude a he ime poin a which he BrainAmp MR samples he nex daa poin (dashed verical line) shows a range from -3001 o -4004 µv. Needless o say ha his variabiliy of 1000 µv a his poin will also find is way ino he correcion emplae as emplae variance and variance in our emplae clearly has o be
page 6 of 7 seen as our enemy in our correcion fea. So, wha we need o do is o reduce he variance and he bes hing would be o shif he waveforms around unil hey show opimal saisical correspondence wih one anoher. Since shifing he daa really is neiher feasible nor possible, his means shifing he Volume Sar markers around insead. Bu if you look a he ime scale of he figure above, you will easily Fig. 4: Phase Shif Effecs on Gradien Arifac Measures see ha e firs need o make sure we can shif hem in smaller incremens han he original 200 µs inervals given by he 5000 Hz sample rae, since his resoluion would be much o coarse o achieve opimal fi. Therefore, we need o upsample our daa o creae more room for our markers and we will do so by increasing our sample rae from 5000 Hz o 50000 Hz. We will go for a en-fold upsampling since i can be shown empirically ha he fi of he volumes does no benefi subsanially from Fig. 6: User Inerface of Resampling Transform he mehod for upsampling (figure 6). Here we can choose beween a Spline Inerpolaion and a Sinc Inerpolaion mehod of which we will choose he Spline mehod since i is well suied for he problems and much faser han he (admiedly opimally suied) Sinc mehod. Afer his procedure we have en imes he number of daa poins per millisecond and all we need o do now is o perform he saisical alignmen of he markers. For his we can use he Analyzer Soluion SliceVolume Align. This lile helper calculaes a lagged correlaion beween one reference Volume s daa in he EEG recordings wih all oher Volume based EEG ime sreches. The correlaion is esed for is maximum in he ime from -2 o +2 ms in seps of 20 µs which gives a emporally resolved correlaion curve wih is maximum a he ime poin of opimal fi beween he emplae Volume and he underlying daa wihin each consecuive Volume. This maximum is where he exising marker is moved o. Due o his emporal alignmen, he variance in he daa is minimized and mos of he ime his variance is clearly wihin he range of he variabiliy of normal sponaneous EEG afer his procedure. Figure 7 shows he very same daa srech as in he figure above, bu afer upsampling o 50000 Hz using he buil-in Sinc-Inerpolaion and afer saisical alignmen of he daa. As 450 400 350 Sandard Deviaion 300 250 200 150 100 50 0 0 2 4 6 8 10 12 14 16 18 20 22 Upsampling Facor (sinc) Fig. 5: Sandard Deviaion as funcion of Upsampling Facor higher upsampling daa raes. The figure 5 shows his effec and i is easy o see ha he mos drasic reducion in Sandard Deviaion is achieved by upsampling a all (facors 2,3,4,5) and ha he effec levels of a an upsampling facor of around 10 imes he original daa rae of 5000 Hz. This upsampling can easily be done in BrainVision Analyzer wih he buil-in Transformaion Change Sampling Rae, where we can choose our arge sample rae of 50000 Hz as well as Fig. 7: Gradien Arifac Measures afer Alignmen can easily be seen, he range of values a he ime a which he BrainAmp sampled he curren daa poin is massively reduced o beween -3820 and -3943 µv. Of course his does no only work wih Volume based markers, bu also wih Slice based markers. The resul would always be an opimally aligned se of markers ha can hen be used for he MR gradien arifac correcion.
page 7 of 7 So, wha s nex in line The nex sep o perform would be he acual gradien arifac correcion and his sep can be done very easily and wih excellen qualiy wihou using any of he more sophisicaed opions available. Bu here are some very ineresing conceps and deliberaions aached o his correcion, which is why I will no delve ino he acual correcion in his issue bu insead will coninue his mini-series on MR-Correcion wih a horough look a his opic in he nex issue of our Brain Producs Press Release. Soluions Ready o go All of he Soluions currenly available can be found ready for download on our websie bu more are in he pipeline and will be added consanly. And, if here is somehing ha you and possibly many ohers would love o have in Analyzer, le us know. Download, double click and resar Analyzer if i was already sared and you will find he new module in is corresponding caegory in he Soluions menu. All of our soluions conain online informaion on heir funcion and parameers, readily accessible via Soluions Help and a double-click on any of he enries. Please noe ha he use of Soluions requires Analyzer version 1.05 or beer. If you are sill working wih an earlier version of Analyzer, please consider upgrading. See our websie for a free download of he newes version of Analyzer and any updaes or addiion of new analysis modules ha may be available. The mos curren version of Analyzer available is 1.05.0003. For quesions, commens or suggesions on Soluions or regarding problems ha could use one, conac Ingmar a suppor@brainproducs.com Brain Producs Workshops BrainVision Advanced User Workshop Gilching (Munich), Ocober 4-6, 2006 by Pierluigi Casellone, Sales Manager Afer years of excellen business resuls which lead o an increasing number of saff members, Brain Producs has recenly moved ino a bigger office. In he course of selecing our new working environmen we decided o inegrae a comforable workshop room. The firs in-house workshop ook place beginning of Ocober 2006. Dr. Ingmar Guberle, Scienific Suppor Manager and Dr. Achim Hornecker, General Manager of he Developmen Deparmen, have run a 3-day BrainVision Advanced User Workshop for an inernaional audience consising of highly experienced BrainVision Analyzer users. During he lecures, Brain Producs covered all aspecs of EEG and ERP research, providing a fundamenal heoreical basis of he mos modern EEG analysis aspecs, ranging from FFT o Wavele analysis and from ICA o EEG & fmri daa correcion. The enhusiasic paricipans were alernaing beween heoreical and hands-on sessions o inensify he informaion given during he lecures and o enjoy he proven Brain Producs learning-by-doing concep. As saed by all paricipans, his meeing has been a grea value for hem and we see his as an excellen saring poin for he fuure. We are already working on he organizaion of he nex Brainvision courses o assure highes qualiy service for all paricipans. Brain Producs GmbH Zeppelinsrasse 7 82205 Gilching (Munich) Germany T +49 (0) 8105 733 84 0 F +49 (0) 8105 733 84 33 sales@brainproducs.com www.brainproducs.com