 Ok, dziękuję za krótką introdukcję. Wreszcie, jeśli ktoś się interesuje, to jest na mój website. Dzisiaj rozmawiamy o tym, żeby też dać jakiś rodzaj podróży na ten temat, ponieważ rozmawiamy o detałach. Jak możemy wytrzymać? Być 3%, 4%. Uwielbiamy komputer w porządku, który jest definityczny na mainstreamie. Dla lat temu mieliśmy odgłosy od Facebooku, że oni will easily do 100 words per minute i wytrzymać komputer w porządku, bo to chyba od tego rodzaju optycznego imię. To było wypracowane, ale w jakiś czasach nie słyszałem nic. W południu lat temu słyszałem, że Regina Dagan, który w porządku w porządku, znalazł i nie słyszałem nic więcej. A więc, kiedy Manny mówił, to w porządku w porządku w porządku, bo wczoraj, mówiliśmy o porządku w porządku, w porządku w porządku w porządku w porządku. Kompany zrozumieją, że ten najwyższy komputer w porządku lub znalazł i nie jest najwyższy komputer w porządku, to multi-billion klap. 100 milionów jest zwykle z koncentracyjnym. Więc inna, dobrze znana kompetitora w porządku jest oczywiście nasz ulubiony Elon Musk z New Rallym, który, dwa lata temu, znalazł, że oni będą to zrobić. Nikt nie wie, co. W porządku w porządku, a co oni się pojawią, jest jakiś rodzaj elektrokortygoografii. Oczywiście, to jest bardzo dobrze wyrażone. Ten robot, który znalazł, może odwołać blot wezelski, mały rozszerzanie, mały numer do kanale, ale dlaczego press jest mówić, że to jest dokładnie, co znalazłem, do podlady kontentu z mojego bryna do harddrive. Nie mogę to powiedzieć, bo, kiedy zobaczysz historię, 15 lat temu, to było nawet wcześniej, ale 15 lat temu, mieliśmy tę pierwszą elektrokortygoografię, więc mówimy, że wyrażanie jest znane, ale to jest kwantytatywane, raczej kwalitatywane. Baszne problemy są jeszcze znalazłe. Więc, wracamy do naszej śniadarki, żeby zrobić coś lepszego. Proponowałem, że to jest kontroversialny titul dla jednej z prezentacji, nie jest to moja prezentacja, jest Marian Dov Giało, i zrobię trochę wyrażania dla tego demo, co będzie available dzisiaj. I też pokażę wam, co próbujemy naprawdę wyrazić, a także wyrażanie niektórych hindransów, o których słyszałem, może to być nie tak zakończone, nie tak zakończone, żeby zrobić specyficzny interfejs dla każdego usera, ale jeszcze rzeczy mogą być wyrażane, zresztą do naszego doświadczenia. Więc, prezentacja, demo D14 będzie spowodowała nasz recentny article o BCI w literacji, ten term, który został postawiony w literacji scientificznej o BCI, i ludzie opowiadają, że na koniec coś, jak 20, może 30% userów, są BCI illiteraty, co oznacza, że nie można używać żadnej komputery. I to jest podstawowa pytanie, ponieważ jeśli ktoś nie może używać BCI, no, co to znaczy? Więc, od naszej doświadczenia, od mojego doświadczenia, od tego, co widziałem w konferencji, w projekcie Europejskiej, w laboratoriach, jeśli ktoś nie może używać komputerowej interfejstwa, to to znaczy, że on nie da się skończone, nie da się skończone software, czy nie da się skończone paradigm. Jednak z różnych paradigms, visual, audio, tactile, P300, SSVP i nie spotkaliśmy żadnych, którzy nie mogły używać żadnych interfejstwa. Oczywiście, jest wiele problemów, beyond that, some people need larger spaces between stimuli, some people need to understand, some people need larger calibration scheme, but this should not be terasy. So we concluded the article with the Finetis para phrases that visa-iliterity does not exist and I will also show you now how are we, how do we achieve this kind of effects using solid engineering and implementation of state of the art science. So first hindrance that we heard about a few minutes ago is of course the headcap, the gel and everything. Now in the company we have developed a headcap that you can put on maybe not in one second, but it is general, you don't have to 3D model every head, it's adjustable for water-based electrodes and as you see I was presenting this the men's conference you get alpha wave nicely after. Brief time and the quality of signal is incomparable to toys like a motif or anything else, it's 34-bit amplifier. Now you may say especially now of course I'm a perfect subject on my head you can put EEG in seconds, but we took a real challenge and this is probably the toughest challenge you can get with African hair and it also worked like a charm. Here you see the classifier accuracy and this is another, this gadget that you see here on the phone is showing the crucial information online. That's another observation that we have had going with brain computer interface to hospitals, to clinics, to normal users that, ok, you put on the cap on your laptop you have the stimuli and something does not work. What does not work? Maybe the electrode does not work. Maybe the user is sleeping, maybe something is disconnected. So you turn it off, usually you take the diet at home and you see one electrode did that contact or maybe the user was sleeping, there was huge alpha or the user was not performing at all to keep the poor guy. So we have this what we call BCI control panel which in real time shows you the most important information of course you can tell or what but the important thing is that you have it on a separate device it can be another laptop smartphone or anything and you can immediately see whether during the calibration for example you have improving classifier certainty you have event related potential and during the performance during the session also you have for each, as you have seen before for each stimuli you have online increasing classifier accuracy. So these two things are probably the major technical achievements that on a way to be comfortable unobtrusive brain computer interface and now having this kind of things we were able to go into real world applications and example application I would like to talk about here is research on disorder of consciousness there are patients usually after comma that we cannot communicate with the basic problem of medicine is that without communication we cannot optimize therapy we cannot optimize rehabilitation basically we do not know if the treatment works there are several places when people with disorder of consciousness usually after comma or stroke are kept taken care of cured but we really don't know if it works when we get in a new spiper or those are awakened in the but the truth is that some of those patients would come out of this state also by themselves and we cannot really see whether given therapy works or not and this is one of the example because now I still remember the title where the BCI is heading this is one of the examples where BCI technology helps but frankly it's not exactly a BCI it's just good old EEG and evoked potentials but okay BCI sounds better so we did not have any introduction here about basic introduction on BCI so to cope with it we just prepared the English version of our introductory web page on breitech.pl where we have explained workings of basic paradigms of brain computer interface and we use just one single example of P300 BCI which is basically flashing the most basic thing, the information is bits, yes or no, zero or one the most basic information transfer is if user counts yes or no for children for example for comatose children is flashing the donald duck and the mickey mouse and we ask kid please count how many times have you seek mickey mouse okay and then we take this average potentials and if we average potentials for mickey mouse and have significant P300 that gives us a huge amount of information we know that the patient heard the command we know that the patient can consciously process the command he's able for cognitive tasks like counting etc etc so this, normally we use this for communication, yes we choose left or right or whatever but basically using the same software and the same hardware we achieve diagnosis of states that with classical tools with behavioral tools reach almost half of misdiagnosis and this is the only the only way to get information directly from the brain and there are papers and also we have more on this if somebody is interested of poster P34 to be part of this research that we talk about now the bird eye view again brain computer interface has become something like an umbrella term for a lot of technologies that are used in business science, technology clinics, hospital or whatever the ninth thing about BCA is really the renaissance of sefalografy because EEG was getting less and less popular nice imaging techniques like fMRI or PET were showing things exactly how the doctors learned it it was particular structure it was easier to interpret with EEG it's not so easy to interpret but EEG has several advantages and the umbrella of brain computer interface can be also seen I would say that BCI for EEG is more or less like for example car races to the motor industry it's as we are hearing not many people are using BCI actually there are not many users there are not many direct users of BCI yet the technology developed within this funny race hey I made it 5 bits per minute faster than the other guys time better oh it was extremely motivated PhD students who was training for 5 weeks for some kind of performance hit so if you look from this site it makes very little sense and also car races make very little sense but the technology that is developed over there really can help in road safety efficiency etc and one of the examples of this kind of technology that was developed for BCI is steady state visual potentials stimulator which we started developing some years ago when we were in a project that involved BCI based on steady state visual event potentials as you probably know SSVP is the fastest paradigm but you should not use it in the low frequencies because it can induce photoepileptic attack so it's definitely better to use high frequencies high starting 30 Hz 40 Hz this is safe and also significantly less annoying however the problem it was quick here this is how originally the guys from philips started to work on this high frequency it was classical by then you had hardware controls LED flashing why hardware control of course on a computer screen you cannot efficiently render frequencies like 32 and 33 Hz because of the refresh rate because of the operating system etc etc so we have hardware showing you these frequencies and this is of course not very efficient this is like a bank teller approach so we actually made it before like this but we just put these LEDs behind the computer screen and there was the first device of this kind from CBIT in 2012 that was at that time the fastest brain computer interface owing to this owing to the natural speed of steady state visual about potential because this is a parallel interface with many things flashing at the same time to choose from and also it was safe BCI because it was high frequency not annoying and not giving a danger of inducing photoelectric attack now after some time we developed this into way more fields with better electronics better synchronization and now we have this 320 segments of a screen that can be independently highlighted with frequencies of your choice and here probably the calibration method that we heard before based on multi-arm band it would be a very nice approach to choose the frequencies which are most right for a given user but surprisingly this device opens new ways, new possibilities of research and maybe also entertainment because if you highlight given section of a screen with given frequency and you see this frequency in EEG then you know that the user looking at the screen and on top of this flashing LED's you can have a movie you can have advertisement you can have text, you can have whatever and if we detect in EEG the frequency related to this particular corner of the screen we know that the user was not just looking at it like it was like it is the case with eye tracking but we also know that the user or we even know that the user was concentrating his or her attention on this part of the screen on this symbol on this image so this opens completely new possibilities of investigating attention focus as opposed or correlated with eye tracking technology that show you what the user was looking at and this this technology will be presented today evening at DEMO D14 of course it does not it doesn't mean that one of you comes and cannot use the BCI that contradicts our main thesis because we don't have here first of all we don't have here all the paradigms implemented, we'll be having as far as remember only visual P300 and SSVP maybe so this is not the complete test if you fail you're still not BCI illiterate you shouldn't worry yet the the head cap and blinker will be available and since I finished two minutes before we have time for question thank you this approach using very high frequency SSVP is very interesting what is the maximum limit of frequency you know from physiology we know that the brain should response to 100 Hz but EEG cannot pick up the gamma so usually you know the power is 1 over f so they go so I wonder ok this is very important what you emphasize that high frequency in SSVP is extremely important because not only reduce annoying but also epilepsy problem and other things which you already mentioned but I understand that this is you want above 30 or 35 Hz what is the highest limit and how how is possible to increase this upper bound if anyway well first of all you should ask why would you want to increase the upper bound and it would be in rare cases if you want to have I don't know very many buttons on the screen it's of course it's very user specific and that's why in the calibration we're choosing particular frequencies that given user responding to and also basically you replied to yourself partly thank you for this mentioning the 1 over f characteristic of EEG signal and of course in higher frequencies EEG can pick up higher frequencies indeed but that would require probably very good skin preparation low impedances so traditionally we usually don't go with this European upper bound of 50 Hz where you have network and between this 30, 35 or 50 Hz we usually find enough frequencies because the response of course the lower frequency the stronger the response and of course the fastest the fastest SSVP BCIs that you are seeing are in low frequencies they take in the risk but they have way way stronger response and if you don't hit the alpha of course you have best responses over there but basically we usually don't go above 50 and the technical details you can get from Marian on the presentation ok Maria to follow up on this same topic does a blinker allow to do code based visual evoke potential which are not at a constant frequency do you use barcodes type of slashing technically yes we did not yet program this but there is a driver for this so we can program anything and this is very good thing apart from this very elegant multi-arm bandit approach to calibration which i thank you for mentioning it's very nice, apart this is probably the next thing that we should be trying or anybody who wants to use the device again technical details Marian is the best source