Category Archives: body area networks

Google wil nu ook data uit je lichaam

Google wil nu ook data uit je lichaam

Bio-elektronica Techbedrijven verzamelen met farmareuzen zeer gevoelige informatie over medische aandoeningen. Ligt die straks bij je baas of je verzekeraar?

MRI-scan van een jongen van 9. Aan het verzamelen en verwerken van medische data kleven privacyrisico’s. Foto ANP

Wouter van Noort, NRC Handelsblad, 7 augustus 2016

Gadgets die werken als medicijnen. Het is de toekomst als een groeiende groep farmacie- en technologiebedrijven zijn zin krijgt. Googles moederbedrijf Alphabet kondigde vorige week een samenwerking aan met farmareus GlaxoSmithKline (GSK) op het gebied van zogeheten bio-elektronica, minuscule implanteerbare apparaatjes die via elektrische signalen ziektes kunnen genezen en voorkomen. Ook Apple en Samsung werken al een tijdje aan bio-elektronica en biosensoren: meetapparaatjes voor lichamelijke functies die je zowel buiten als binnen in je lijf kunt dragen.

Apple en Samsung hebben, net als Google, bovendien steeds nauwere banden met de farmaceutische industrie. Googles zusterbedrijf Verily, dat zich helemaal richt op farmaceutische toepassingen, sloot op andere gebieden al eerder samenwerkingen met Johnson & Johnson en Novartis. Apple werkt ook samen met GSK, en Samsung investeert veel om zelf meer een farmaceutisch bedrijf te worden.

Dat juist bedrijven uit de consumententechnologie ineens zo geïnteresseerd zijn in de farmacie, en vooral in de bio-elektronica, roept interessante vragen op. Met name over privacy: behalve informatie óver mensen, kunnen technologiebedrijven dankzij bio-elektronica straks namelijk ook data verzamelen ín mensen.

„Ik ben er niet gerust op”, zegt Wouter Serdijn, hoogleraar bio-elektronica aan de TU Delft en London University College. „Een bedrijf als Google weet al heel veel van je, en juist als je gegevens uit bio-elektronica combineert met grote hoeveelheden andere data, ontstaan mogelijk interessante inzichten over de gezondheid van individuen.”

Die inzichten kunnen nuttig zijn voor de genezing van bepaalde aandoeningen, maar er zitten ook privacyrisico’s aan. Ook de Haagse technologiedenktank Rathenau Instituut spreekt al jaren zijn zorgen uit over privacygevolgen van geïmplanteerde elektronica.

Wat kunnen techbedrijven nou precies te weten komen ín een lichaam? „Het gaat met de huidige bio-elektronica vooral om de communicatie tussen cellen of bijvoorbeeld informatie over de zuurtegraad in je darmen”, zegt Serdijn. Volgens hem is de informatie die uit bio-elektronica en -sensoren komt op zichzelf commercieel nog niet direct heel bruikbaar. Maar gecombineerd met andere informatie, over bijvoorbeeld lichaamsbeweging, zijn daar mogelijk wel interessante patronen in te ontdekken. „Dan zou je er mogelijk zaken als epileptische aanvallen mee kunnen voorspellen, en misschien wel andere ernstige aandoeningen”, zegt Serdijn. En dat is informatie die je niet altijd wilt delen met je werkgever of verzekeraar.

GSK wil geen details geven over hoe het informatie uit bio-elektronica precies gaat delen met Googles zusterbedrijf Verily. Wel zegt woordvoerder Carien Mulder: „Wij staan honderd procent voor het waarborgen van vertrouwelijke patiënteninformatie, en dat is ook een prioriteit in de nieuwe samenwerking met Verily.” Ze geeft echter geen antwoord op wat er precies is afgesproken over de data die er worden verzameld in het lichaam.

Een woordvoerder van Alphabet kon niet op tijd reageren op vragen van NRC. Wel zei Brian Otis, de technologiedirecteur van Alphabet-dochter Verily, vorige week tegen het Amerikaanse Forbes Magazine dat het zijn bedrijf bij deze samenwerking vooral te doen is om de data. „De uitdaging met bio-elektronica zit ’m uiteindelijk in data. Het uitlezen en interpreteren van de signalen. Natuurlijk heeft Google expertise in het omgaan met grote hoeveelheden data, beslissingen nemen op basis van data en feedback geven aan de gebruiker.”

Google beschikt over enorm veel gegevens over menselijk gedrag. Via de Android-smartphones van het bedrijf verzamelt het ook veel informatie over bijvoorbeeld lichaamsbeweging. Het blijkt bij dit soort big data-toepassingen vaak erg lastig om verbanden tussen gegevens te ontdekken die ook echt bruikbaar zijn. Maar juist een bedrijf als Google is daar heel goed in.

Informatie uit bio-elektronica zou ook bepaald niet de eerste medische data zijn die Google de laatste tijd verzamelt. Via de zoekmachine ziet het bedrijf al jaren welke medische vragen bezoekers stellen. Via zusterbedrijf 23andMe, dat genetische tests ontwikkelt, heeft Google de laatste jaren daarnaast van vele duizenden mensen DNA-informatie verzameld. Onlangs sloot het een samenwerking met de Britse National Health Service voor het analyseren van grote hoeveelheden patiëntengegevens van Britse burgers.

Dat zijn zeer uiteenlopende projecten, met ook zeer uiteenlopende privacyvoorwaarden. Het is niet automatisch zo dat Google met die informatie allerlei gedetailleerde profielen opbouwt die het zomaar kan herleiden tot individuen. Laat staan dat het die zomaar kan doorverkopen, als het bedrijf dat al zou willen. Er gelden voor medische gegevens strengere privacywetten dan voor andere soorten informatie.

Maar de Amerikaanse technologiereus verdient wel veruit het meeste van zijn geld met op maat gemaakte advertenties. En als je advertenties op basis van je zoekgeschiedenis krijgt voorgeschoteld, waarom dan niet op basis van data uit een biosensor die uitwijst dat je binnenkort misschien behoefte krijgt aan een bepaald medicijn?

„Zover is het voorlopig waarschijnlijk nog niet,” zegt hoogleraar Serdijn. „Maar het is wel zaak om dit heel goed in de gaten te houden.”

HOE BIG PHARMA EN BIG TECH SAMENWERKEN

Apple sloot in juli een samenwerking met Glaxo Smith Kline (GSK) om behandelingen te ontwikkelen voor reuma. GSK gaat daarvoor Apples onderzoekssoftware ResearchKit gebruiken. Dat platform brengt allerlei gegevens samen die Apple over zijn gebruikers verzamelt, bijvoorbeeld over lichaamsbeweging. Die is te meten via de bewegingssensoren in iPhones. Dergelijke sensoren kunnen volgens de twee bedrijven ook worden gebruikt om nauwkeuriger in kaart te brengen hoe reuma het leven van patiënten beïnvloedt.

Googles zusterbedrijf Verily werkt samen met Novartis om een slimme contactlens te ontwikkelen die bloedsuiker meet in het oogvocht van diabetespatiënten. Zo’n lens zou in de plaats kunnen komen van andere manieren om bloedsuikers te meten, bijvoorbeeld van de bloedprikken die nu gebruikelijk zijn.

Telefoonmaker Samsung investeert ook fors in biotech en farmacie, onder meer via Samsung Bioepis en Samsung Biologics.

Lecture on Electroceuticals: getting better with electricity

Lecture on Electroceuticals: getting better with electricity

Lecture on Electroceuticals: getting better with electricity

On May 6, 2015, Collegerama of TU Delft made video recordings of the lecture I gave on Electroceuticals.

Electroceuticals are the electronic counterparts of pharmaceuticals and are miniature electronic devices that interact with the body in an electrical fashion.

In this talk I discuss: neurostimulation and the need to make neurostimulators smaller, more power efficient and more intelligent; optogenetic neuromodulation and the need to make this new neuromodulation modality operate in a closed-loop fashion; neurosensing devices to make neurostimulators intelligent and thereby adjust themselves to the therapeutical needs of the patient; autonomous wireless sensor nodes that can measure temperature or the electrocardiogram without the need for a battery; an outlook into the future of electroceuticals with the promise to treat a larger variety of neurological and brain disorders better.

Click here to start watching the video and slides:

https://collegerama.tudelft.nl/Mediasite/Play/cc7888beb88349c1a60c1414476b577a1d?catalog=528e5b24-a2fc-4def-870e-65bd84b28a8c

Injectable Electronics: dawn of a new era in electroceuticals?

Injectable electronics still need to become smaller

Frequent readers of this weblog may still remember a previous post, entitled “And the paralyzed will walk again“. This phrase comes from a Discovery Channel movie/documentary, called “2057: the body”, in which it is predicted that by the year 2057 you will be able to survive a three story fall and even be able to walk again as there will be tiny microstimulators attached to your muscles, which can be injected.

Injectable electronics, how fascinating would that be! No more lengthy surgeries, during which only a single, bulky device is implanted, but rather a procedure that takes less than a couple of minutes, during which multiple micro-stimulators are inserted via a seringe. Once done, these stimulators will form a wireless network and will provide the motory neural pathway with well-timed electric stimuli necessary to evoke the correct contraction of the multiple muscles involved in a delicate movement or even seemingly simple posture control.

But how feasible is this idea of injectable electronics? If you search for the term injectable electronics, you will most likely find a lot of references to the work of John Rogers, professor at the University of Illinois in the US, who built “an electronic LED device so tiny it can be injected into delicate tissue, such as in the brain, without harming it“.
Other links that can be found refer to work done on silk implants or even magnesium implants that are either stretchable or can easily dissolve into the body once the good work has been done.

I personally believe that we only can create injectable electronic devices if they have at least some intelligence in them. For this, the good old silicon would be an excellent candidate. Silicon is a nice and friendly biocompatible material, can be made bendable (by thinning the substrate) or stretchable (by removing the substrate altogether at some points). And what’s more, silicon can accommodate stimulation circuitry, sensors, signal processing, communication electronics, antennas, battery foils, all the good stuff needed to make a good injectable.

Of course, in order not to damage the tissue that the electronic device is injected in, it needs to be small, i.e., thin and narrow. It is however allowed to make it long, e.g., a couple of millimeters up to one or two centimeters. These unconventional dimensions raise very exciting technological challenges, such as:

  • how can we create electronic integrated circuits (ICs) that are merely one-dimensional, i.e., are not wider than one, maximally two, bondpads?
  • how can we transfer information and energy to an implant that has virtually no area?
  • what kind of material should we use for the antenna and electrodes?
  • will a Li-Ion battery foil have enough capacity to provide successful stimulation of the tissue, or should we refrain from using batteries altogether?

There obviously is still a lot to do. Exciting stimes ahead, if you ask me.

Wouter

A new name, but Biomedical Electronic remains

Biomedical Electronics Lab

Dear Reader,

The Biomedical Electronics Group underwent a small name change. From now onwards, the group is called “The Biomedical Electronics Laboratory”.

Its mission is “to provide the technology for the successful monitoring, diagnosis and treatment of cortical, neural, cardiac and muscular disorders by means of electroceuticals.”

To this end it conducts research on, provides education in and helps creating new businesses in neuroprosthetics, biosignal conditioning / detection, transcutaneous wireless communication, power management, energy harvesting and bioinspired circuits and systems.

Mission Possible

In order to present the Biomedical Electronics Group of Delft University of Technology to a couple of companies, it made sense to reveal our mission statement. So here it goes…

The mission of the Biomedical Electronics Group of Delft University of Technology is "to provide the technology for the successful monitoring, diagnosis and treatment of cortical, neural, cardiac and muscular disorders by means of electricity." In order to reach this goal we investigate and design circuits and systems for electrical stimulation, ExG readout, signal specific analog signal processing, power management/conversion, energy harvesting and wireless communication, to be applied in future wearable and implantable medical devices, such as hearing instruments, cardiac pacemakers, cochlear implants and neurostimulators.

So how about that? Reactions are welcome via this blog.

Wouter

New way of data conversion

Analog-to-digital converters (ADCs) are indispensable building blocks of wearable and implantable biomedical data acquisition systems. Ultra-low-power ADCs for biomedical signal sensing have witnessed a dramatically reduced power consumption in recent years, but we have to admit that our biomedical systems need more breakthroughs than just squeezing harder in conventional ways.

As is known to all, many biomedical signals are born with a sparse nature. A large amount of redundant digital samples will be thus generated if we use Nyquist-rate ADCs to convert such signals. Most likely, ADC power savings are not a major concern in a system in which transmission power dominates the overall power consumption. However, if this is not the case, from a signal point of view, new ways of sampling or sensing are necessary to further improve the performance of the whole system.

A new and promising ADC approach for biomedical data acquisition is based on so-called level-crossing (LC) sampling, in which samples are generated only when the input signal crosses the threshold levels, so there is no redundant sample in this case. However, the conventional LC-ADC utilizes power hungry comparators and DACs, which causes the LC-ADC to consume much more power than ultra-low-power Nyquist ADCs (e.g., SAR ADCs). In our new approach (mentioned by Wouter earlier in the weblog), innovations at both system level and circuit level enble us to design a more power-efficient LC-ADC. Power consumption is now in the range of hundreds of nanowatts. We are currently investigating the possiblity to further improve its performance and reliability.

Yongjia

Slides Hermes Partnership Workshop “Visions Towards ICT Supported Health” have been posted

HermesStill in shock by the post below? Don’t be. As always there’s hope on the horizon. The slides of the Hermes Partnership Workshop "Visions Towards ICT Supported Health" of last week have been posted online. If you want to find out more about one or more of the topics below, don’t hesitate to click here or on the links below.

“E-health in practice, business opportunities”, prof. dr. Felix Hampe, University of Koblenz, Germany

“Present experiences and future perspectives of Telerehabilitation”, prof. dr. ir. Hermie Hermens, University of Twente, The Netherlands

"Moving diagnostic, monitoring and therapeutic wireless medical devices into the homes and into the body",  dr. ir. Wouter Serdijn, Delft University of Technology

"Nanoscale smart communication components and systems”, a research proposal of Hermes partners, dr. Jean Benoit Pierrot, CEA LETI France  

"Status e-health and  telehealth in Poland, prof. dr. Łukasz Januszkiewicz, University of Lodz, Poland

"e-health systems developments and business opportunities at SME-companies”, dr.ir. Piet Verhoeve, Televic, Belgium

"Energy harvesting in e-health applications”, dr. Paul Mitcheson, Imperial College, UK

Wouter

 

Biomedical Group Meeting today

lunchWhile enjoying lunch, the Biomedical Electronics Group gathered in the Davidse room (named after the former head of the Electronics Research Lab and also my "promotor", Jan Davidse) to listen to three presentations. The first one was by Duan Zhao, on an interesting new way of bridging the gap to low-power software radios by means of subsampling. After an introduction on the operation of a subsampling receiver, he explained to us a technique to remove the jitter originating from the sampling clock by using a harmonically related reference. Currently Duan is working hard on a manuscript to be submitted to GlobeCom.

The second presentation was by Neil Yongjia (as we call Yongjia because he will perform a song by Neil Young at the ELCA festival) on the correspondence and differences of successive approximation (SA) analog-to-digital converters (ADCs) and level-crossing ADCs. There is an interesting paradigm shift involved in the latter and many issues, such as DC sampling, bandwidth and slope limitations need to be investigated still. Nevertheless, it looks like the level-crossing ADC is a natural candidate for the conversion of physiological signals such as those that are generated by the body.

The third presentation was by Yours Truly, and was about how to turn your profession into the best job in the world. We touched upon cultural aspects, organizational aspects, academic aspects and personal aspects and things like procrastination, drive, bosses and the fun-factor. Probably in June, I will give a similar presentation to my colleagues of our faculty. 

Tomorrow will be the ELCA festical. Don’t miss it, as the world will never be the same…

Wouter

Rats go wireless in an analog fashion

Rat carrying wireless systemThis little image on the left shows a rat carrying a wireless system, partly mounted on his head, partly realized as a kind of a backpack. It has been developed by researchers at Harvard University in close collaboration with colleagues at California Institute of Technology and is being used for neurological research on rats in the wild. According to Nature (Febr. 25, online) the entire systems comprises "a tetrode microdrive, for chronic positioning of electrodes in the brain; an integrated circuit for high channel-count neural recordings; and a radio-frequency wireless transmitter. The device takes up to 64 analog voltage signals from neurons in the brain and muliplexes them into one signal that appears in a temporally interleaved fashion, one after the other. Then that signal is transmitted by analog FM radio to a receiver."

The article further reports that good old FM (frequency modulation) transmission has been used as it outperforms digital wireless communication on weight, power drain, throughput and distance.

So why does analog FM outperform its digital counterparts, such as FSK, QPSK, QAM and OFDM? Before answering this question it is important to realize that from a channel-capacity perspective (as defined by Shannon) there is no preference for analog modulation over digital modulation. The answer thus has to follow from practical considerations. Digital modulation implies that the information is transmitted over the wireless radio channel in a digital fashion. As all information in nature, also neural signals are analog in nature and thus, in order to prepare the neural signals for their wireless journey, they have to be converted to the digital domain by analog-to-digital conversion (ADC). This thus requires at least one ADC. Often, depending on the digital modulation type used, channel coding is performed prior to the digital modulation. As a consequence, with these additional blocks, the entire transmitter becomes more complex, which, in turn, entails a larger power consumption and, when battery-operated, a larger battery and thus a larger weight on the head or back of the rat.

Another reason why analog FM may outperform its digital counterparts lies in the frequency spectrum of the transmitted radio signal. FM produces an almost flat frequency spectrum. As a consequence, it is relatively immune to frequency-selective fading, which is good for radio communication over relative long distances. Also, FM transmission does not require a highly linear power amplifier. This is good for its power efficiency and thus for the overall energy efficiency of the transmitter.

One final remark. From the picture it looks like the rat is loaded with a transmitter that has been implemented using discrete components, rather than with a single chip (or integrated circuit, IC). Many of the blocks needed for digital transmission could be implemented with a much smaller form factor and consuming less power when realized on-chip, rendering the antenna the largest component and the largest power consumer. In such a case the choice for either analog or digital could have just as well turned upside down.

Wouter