Frederik Nebeker and Michael Geselowitz IEEE History Center, Piscataway, New Jersey

Those Who Have Played Pioneering Roles in Biomedical Engineering Come from Many Fields. In These Excerpts from the Oral-History Project, Some of the Pioneers Tell of Their Involvement in Shaping the Discipline.

The oral-history project, initially funded

as part of the EMBS anniversary project,

was substantially expanded through a

grant from the United Engineering Foun-

dation, which made possible the inter-

viewing of an additional dozen pioneers

of biomedical engineering from other

engineering fields, mainly chemical and

mechanical engineering. This article in-

cludes a large collection of excerpts from

all of the oral histories, hence providing

some indication of the full range of bio-

medical engineering. ©PHOTODISC

0739-5175/02/$17.00©2002IEEE Nebeker: Your work was on control systems. at the highest level of experimental cardiology, Robert Arzbaecher and he found a trained electrical engineer to help Arzbaecher: Yes, I focused my graduate him with instrumentation problems. During that Director, Pritzker work on that. When I finished my degree in 1960 year we did the first isolated human heart map- Institute of Medical and left [the University of Illinois at] Urbana I ping experiment and published in Circulation. Engineering Illinois took a teaching position at the Christian Institute of Technology Brothers College in Memphis, Tennessee. * * * * * (interview by Christian Brothers College had a good under- Frederik Nebeker on graduate engineering school, particularly in the Nebeker: What was the impact of the mini- electrical area. … It was not a small school, but computer on electrocardiography and similar 14 October 1999) not much was going in on the area of control. areas? Certainly not much in my area of reactor control Arzbaecher: There were two impacts. First, was going on in Memphis. I taught electrical en- it made it possible for many laboratories without gineering but was not able to continue my re- major computer resources to acquire significant search. It was in that context that my conversion computing resources. Secondly, it led to digital from electrical to biomedical engineering be- signal processing, which is now standard, and gan. also to on-line real-time signal processing. For a year I looked for a challenging re- Without the smaller computer and its capabili- search opportunity in Memphis to go with my ties for on-line acquisition, analog-to-digital teaching. Then I met Daniel Brody at a cocktail conversion, and processing in the digital do- party. He was a cardiologist from the Univer- main, that wouldn’t have happened so quickly. sity of Tennessee College of Medicine in Mem- My experience before then had been with punch phis. Unknown to me at the time, he was a very card machines, batch processing, and number well-known cardiac theoretician. Brody’s un- crunching. dergraduate training was in physics. … When I found out more about his background, it was * * * * * the beginning of a wonderful relationship. I spent the next seven years working with Brody Nebeker: Was delivering the drug automati- in his laboratory every chance I could get, on cally your objective? Saturday, Tuesday, and Thursday afternoons to Arzbaecher: Yes. That was my dream. It has the extent I could arrange my teaching sched- been a long time in bearing fruit. ule. We worked in what is now called theoreti- cal electrocardiography, which is the physical Nebeker: It sounds very ambitious for the and mathematical basis for interpretation of the 1970s. human electrocardiogram in health and dis- Arzbaecher: Yes, it was. Support came ease. Brody and I wrote a lot of papers together. from the pharmaceutical company G.D. Searle. That’s how I became a biomedical engineer. They were particularly interested in finding out whether their drug Norpace was effective in * * * * * treating atrial arrhythmias. Necessity being the mother of invention, I invented the esophageal Arzbaecher: After six years of theoretical electrode for recording the electrocardiogram mathematical modeling work I had an urge to from the esophagus. get closer to clinical electrocardiography. I had an opportunity to go to Europe for a year on a * * * * * National Science Foundation fellowship, and Dan Brody set up an opportunity for me to work Arzbaecher: It is interesting how I discov- in the laboratory of Dirk Durrer at the University ered the pill electrode. … In the lab, I had a grad- of Amsterdam. That experience was of major uate student of mine push a catheter electrode significance in my career. Durrer had a concept down my nose and into my throat and esopha- for removing the human heart soon after clinical gus. I was able to measure a beautiful p-wave, death and keeping the heart alive long enough to which was an atrial recording in the electrocar- map its electrical activation. This had been done diogram. with the hearts of animals, and Durrer was poised to do the same thing with the human Nebeker: Did you get a much clearer signal heart. The instrumentation requirements of that that way? task were significant. With the isolated heart on Arzbaecher: The p-wave obtained is three a perfusion stand, needle electrodes were placed to four times bigger than QRS. Compared to all over the heart to record on a multichannel ba- routine electrocardiography on the body sur- sis. This was in the mid-60s, when that was not face, the esophagus gives a beautiful p-wave an easy thing to do. My time in Amsterdam was from the atrium. The recording we made was fortuitous for him and for me. I became involved clear, but it hurt. That night I could still feel the irritation in my throat that had been caused by Arzbaecher: One day in my laboratory in the catheter. Iowa City I disconnected the pill electrode from the electrocardiograph, connected it to an exter- Nebeker: You’re in the class of people who nal stimulator, and discovered I could pace my experiment on themselves. heart. Arzbaecher: That was in the days before we Nebeker: More experiments on yourself. were so careful about human investigation. My Arzbaecher: By connecting the electrode lo- throat and sinuses hurt, I realized that the problem cated immediately behind the left atrium to an was that big old stiff catheter. All I really wanted external pacemaker I was able to capture the were the two rings at the end of the catheter. The heart and run it at any rate I wanted. next morning I went into the lab, took a pair of di- We proved that esophageal pacing was both agonal cutters, and stripped away the polyethyl- safe and effective and received premarket ap- ene coating until I was left with the tip of the proval. No longer a substantial equivalence catheter and the pair of wires. Then I buried the thing, this required the full FDA premarket catheter tip in a spoonful of ice cream and swal- route. It was necessary to perform experiments lowed it. My students laugh about this. One does- and prove there would be no damage to the n’t chew ice cream, so the bipolar pair of esophagus. It also had to be proved that this de- electrodes caused no problem. I just swallowed it, vice was effective in pacing the heart. Following and down went the pill. That was the first pill all those studies we appeared before the FDA electrode. That was the business end of a catheter Cardiology Advisory Board at the end of 1986 electrode with a bipolar pair and a pair of wires and got the PMA approved. I developed a small holding it up. Within a week or two I discovered a battery-operated electronic external pacemaker source of very thin and flexible stranded stainless to which the pill electrode could be connected. steel wire with Teflon insulation. That was it. It could pace the heart. This created an opportunity to do a number of * * * * * things that couldn’t be done before. For exam- ple, a lot of patients can’t exercise due to age, in- Arzbaecher: The esophageal electrode gave firmity, obesity, and other reasons. Cardiac us the signal with which atrial arrhythmias could stress testing can be done by pacing the heart be instantly recognized. Atrial arrhythmias are rather than by exercising the patient. That was an not easy to recognize on the body’s surface. important development. Imaging studies such as Nebeker: Is this a recognition that comes echocardiograms could also be done. Tempo- through visual analysis by a cardiologist? rary esophageal pacing could be used to convert Arzbaecher: It’s computer recognition. The arrhythmias. Atrial flutter, especially in chil- algorithm looks at the esophageal signal, ana- dren, is easily converted by over-drive pacing. A lyzes it, and comes up with a rhythm diagnosis in short ten-second burst of temporary pacing from real time. The whole spectrum of cardiac arrhyth- the esophagus could convert the arrhythmia to mia can be detected, including atrial fibrillation, normal rhythm. atrial flutter, atrial tachycardia, premature atrial beats, and ventricular arrhythmias. * * * * * Our interest was in treating atrial arrhythmias with a Searle drug. I had the first half of that Arzbaecher: The dream of the drug pump problem solved, which was how to automati- people is the treatment of diabetes. A drug pump cally recognize the onset of an atrial arrhythmia with insulin to automatically treat the diabetic in a coronary care patient. patient would be a marvelous technology. Bil-

Nebeker: Did you write programs to analyze the p-wave? Arzbaecher: Yes, my students and I wrote them. A student who helped with that was Janice Jenkins. She is now a professor at the University of Michigan and on leave from Michigan serv- ing as the program director of bioengineering at NSF. She took the esophageal signal and wrote the program to analyze the arrhythmia. It was the first dual chamber arrhythmia analysis program ever written.

* * * * * Dr. Robert Arzbaecher swallowing the Pill Electrode, which he invented, for recording the electrocardiogram from the esophagus. lions of dollars have been spent to develop that a pressurized reservoir of drug with a valve. technology, and several companies are in- When you open the valve, drug squirts into a volved. They have been stymied largely by the catheter and then moves into a central vein. That absence of a good glucose sensor. The pump can device is pretty reliable. There are not many deliver insulin, but no way has been found to moving parts, and it is easy to control. All that close the loop. Insulin is not an easy drug to store has to be controlled is the rate at which the valve and deliver in an implanted pump, but the big- is operating. gest problem is the measuring of glucose levels Nebeker: Are these valves controlled elec- to trigger the pump. That technology is still be- tromagnetically? ing developed. I believe implantable pumps will Arzbaecher: Yes. I have a concept I have become more commonplace, and maybe with tested on animals in which I programmed an time the particular notion of an implantable implantable drug pump with a valve to deliver pump for the treatment of atrial fibrillation will pharmacokinetically designed exponentially become more attractive. tapered infusions, which are known to give Nebeker: Is the control system developed? and maintain a good therapeutic response. Arzbaecher: Yes. For my own lab I have Control of the pump is not difficult, and the published protocols for running a pump. The problem of recognition of atrial fibrillation is implantable pumps that are available do not well in hand in my laboratory as well as in have motors in them. They have valves. There is other places.

Babb: [Babb, in charge of the nuclear engi- the year before I came. Suddenly Moulton said neering program at University of Washington, to me, “You’ve always been interested in medi- had just begun using a teaching reactor for neu- cine. Why don’t you go and see him?” Thus we tron activated spectroscopy] … a secretary … made an arrangement. I went down to the Uni- said, “My daughter has cystic fibrosis, and one versity Hospital to a little room off the cafeteria. of the manifestations of it is the abundance of so- It was a daunting situation. There were about six dium chloride in the sweat, fingernails, and toe- physicians in white coats sitting there. They ex- nails. Maybe you should get together with the plained to me how they had just developed a way people at Children’s Hospital and see if you may of permanently implanting plastic tubes in the have a way of screening newborns for the pres- artery and vein of patients, bringing them out ence of cystic fibrosis so that they can get into through the skin and connecting them with a early treatment.” Dr. Stanley Stamm came U-tube so that when they were not on dialysis the down, and we talked about it. The result was that blood flowed at a high enough velocity to pre- we set up a protocol whereby he advertised vent clotting. Then they were clamped off and around the country and overseas that we were connected to the dialyzer with the artery going to Albert “Les” Babb accepting toenails which would be irradiated the inlet port and the vein to the outlet port to Professor Emeritus of with neutrons. start a dialysis episode. Chemical and Nuclear We accepted maybe a thousand samples Nebeker: How long had that kind of Engineering, from all over the world to irradiate. Here is a pic- extracorporeal dialysis been possible? ture of a resident clipping the toenails of a baby Babb: It was first used during World War II. University of Washington at the University Hospital. There was a writer for A physician named Willem Kolff treated pa- (interview by Life magazine who lived in Belleview, just tients in Kampen in the south of Holland. … but Frederik Nebeker on across Lake Washington, who did a story on this they could not treat patients whose kidneys 6 December 2000) novel use of a nuclear reactor. That’s what got would never return to normal function. They me associated with some clinical people. were only able to keep the patency of the blood access to the cardiovascular system for a few * * * * * weeks or a month at the most. If their function were not restored they would die. After the war Babb: Early in 1963 Dr. Moulton called me everyone was trying to develop permanent ac- into his office and said, “I just had a telephone cess. call from Belding Scribner, M.D. He wanted to know if there’s a young faculty member here * * * * * that could work with him in reducing the cost of hemodialysis.” I had never heard of this proce- Babb: I assimilated all this information, dure. The reason Scribner called him was be- looked at the tanks in use, made some mental cause he took care of Moulton when he was sick notes, and came back to my office in the nuclear three 8-hour treatments per week rather than two 12-hour treatments were more optimal and con- vinced the medical staff to switch to this prac- tice.

* * * * *

Babb: Then I received a call early one Satur- day morning [in 1964]. [Dr. Scribner] explained to me that a 15-year-old girl he’d seen many Albert “Les” Babb describes an extracorporeal times had been turned down for treatment by the system for the continuous treatment of patients deciding committee I mentioned earlier. He told with sickle cell anemia in 1981. The design of me she was going to die if she didn’t get treat- the machine was based on the pioneering proto- ment within four months, and the university had type designed and built by Babb and his team prohibited him from taking any more patients. at the University of Washington for over- He said, “Could you take this monster and min- night-unattended home dialysis. iaturize it?” … He said, “It would be the same principle, diluting concentrate, but for one pa- tient and portable so it could be taken to a pa- tient’s home.” I said, “I don’t know. I’ll have to reactor building. I had barely arrived when there talk to my engineers and technicians.” I called was a phone call from Dr. Scribner, who had told them together and put the problem before them. me it cost about $20,000 per patient per year for And I said, “We’d have to make most of the stuff two 12-hour dialysis sessions per week. ourselves.” We couldn’t buy heat exchangers, I said to Scribner, “I take it your are inter- for instance, that were this small, and the instru- ested in reducing the cost,” he said, “Yes, we’d ments weren’t readily available and we’d have like to get the cost down. Do you think you can to modify them. I said, “It really has to be almost help us?” And I said, “I need more information. I automatic.” The patient could not be expected to need to be walked through the whole proce- do too much. We went for an automated system dure.” wherein at the push of a button the machine From this idea we said to ourselves, “Why would pasteurize itself using hot water then. don’t we premix those chemicals in a concen- Then it needed to produce dialysate at the right trate and dilute it in the right proportion continu- temperature and composition and give an indi- ously. In this way we’ll remove the tank from cation when one could hook up the blood lines. each bedside and pipe the dialysate around the My staff said they’d like to do it, and I told walls to maybe five stations where they can be Scribner we’d take a shot at it. connected to the dialyzers. The dialysate would Nebeker: This was a bootleg development. flow through and then come back into a waste Babb: Yes. And I couldn’t have done it if I line.” … it was dubbed “the monster!” Then all hadn’t had control of it. I could not do that here hell broke loose because manufacturers of com- now. And at that time we happened to have mercial machinery besieged us, each wanting to probably the best cadre of electronics techni- be a supplier. cians on campus, and they were all under my [The Milton Roy Company of St. Petersburg, supervision. We had a lot of ex-Navy people Florida] flew me down as a consultant. I was who had been in the nuclear propulsion pro- there at the startup at the old VA Hospital in Coral gram. We couldn’t have done it without them. Gables, Florida. From that beginning these cen- If I had gone to this current chairman and said I tral systems appeared all over the world. wanted to do this, there would be no way. It was Nebeker: Were these multipatient dialysis just the right time. systems? Nebeker: How many people did you have Babb: Yes. Here we could handle five pa- working on this project? tients at once. … it was working and at half the Babb: I would say about eight at the peak of cost; i.e., the annual patient cost was now about it. [The machine was a success.] People came $10,000. A lot of the work we did after that was from literally all over the world—Sweden, Den- mathematical, and we tried to model this process mark and so on—to photograph the machine and to see if it could be optimized. We suggested take away drawings. Nebeker: What was Norman Jeff Holter’s cated analytic programs for diagnosis upon James Bailey and motivation for using this ambulatory ECG? playback, or even better, immediately. Bailey: His idea was that taking a single Rosalie Dunn resting ECG was too narrow a slice in time. * * * * * Bailey: Center for The idea was to capture a much larger block of Information Technology, cardiac activity—throughout the day and Dunn: Hubert Pipberger was interested in de- NIH night—to get a better picture of what was hap- termining the physiological phenomena behind Dunn: National Heart, pening with the patient during his normal the production of the ECG signals, developing Lung, and Blood daily routine. the model that could accurately reproduce such signals, establishing the external electronic sys- Institute, NIH tem, which could detect these signals, and, by use (interview by “Every ten years someone comes of a validated model, properly interpret them with Frederik Nebeker on up with the idea of applying a high level of sensitivity and specificity. 25 April 2000) Nebeker: His way of creating that model harmonic analysis to the ECGs. was going to be on a digital computer? The problem is that harmonic Dunn: He saw the computer as his basic tool, analysis doesn’t tell whether the capable of reading tracings with higher resolu- T-wave is before or after the QRS. tion, consistent accuracy, and greater speed. Nebeker: There were quite a number of peo- That is critically important loss of ple in various areas of science, with, for exam- information.” ple, analog harmonic analyzers, looking for periodicities in data. Nebeker: Whereas a stress-test ECG could Bailey: Every ten years someone comes up be connected and done in a single location? with the idea of applying harmonic analysis to Bailey: Right. the ECGs. The problem is that harmonic analy- Dunn: There was a classic study, from the sis doesn’t tell whether the T-wave is before or 1960s perhaps, monitoring workmen—I can’t after the QRS. That is critically important loss of recall the industry involved, perhaps stevedores information. or electricians—who were wearing these Nebeker: Was the use of harmonic analysis Holters for at least one day, if not more. These ever fruitful with cardiographs? strapping, healthy men were throwing spurious Bailey: I don’t think it has been helpful, ex- arrhythmias like PVCs and PACs. This may cept for the fact that in the early sixties people have been the first study documenting signifi- believed that all the information was contained cant “normal” variation. What did this mean below 60 Hz. prognostically? They didn’t know. Bailey: Later people found important infor- Nebeker: Was this the first time somebody mation at the 100 Hz level. Some investigators had done ambulatory ECGs? have found important information at even Bailey: Holter’s was the first, yes. And it was higher levels than that. all analog when he started. Avionics was still an- Nebeker: How did they learn that these alog even in the late sixties. higher frequencies were important? Dunn: This capability for monitoring would Bailey: They learned that by studying pa- not have been available from the average cardi- tients with cardiac pathology. ologist, but it was used in some hospitals. Nebeker: Earlier they might have filtered Bailey: Yes. Where I did my residency, the out that part of the signal? cardiology department used one. That was in Bailey: That’s right. In the old PHS-D pro- 1966-1969. gram of 1964 was based on the idea that the sig- Nebeker: Was the Holter Monitor some- nal would have been passed under a 50 Hz filter. thing that was manufactured? It really distorted the QRS. Bailey: Yes. Avionics was the company that Dunn: When I began to work in this field in made it. For patients that were having palpita- the seventies, a large percentage of papers pre- tions or seizures or fainting spells, it was appro- sented at conferences dealt with filtering and priate when the cause was not otherwise clear. It massaging the signal. also probably unmasked coronary disease for some patients. * * * * * Nebeker: Has this kind of diagnosis contin- ued to the present? Bailey: I am skeptical of cardiologists uncrit- Bailey: Oh yes. It’s big business. ically taking interpretation from a computer pro- Nebeker: I suppose these Holter Monitors gram. have become transistorized and miniaturized. Dunn: This became a serious issue. Dunn: Hardware has been miniaturized, in- Pipberger himself always warned against “heart cluding compressed software and more sophisti- disease of computerized ECG origin.” Nebeker: “If the computer said it, it must comparable results could be obtained, for the right.” most part. Bailey: A lot of them had that kind of naiveté. * * * * * Dunn: Some preferred a system whereby everything that came out of the computer Nebeker: The way it was actually imple- should be over-read—that some cardiologist mented was that a cardiologist would run this should at least take a look at it and sign off on program, over-read it, and then feel free to dis- it. agree if his interpretation was different? Dunn: And the cardiologist would notate the * * * * * tracing for differences. Computerized rhythm determination was probably the first aspect Bailey: What evolved in the seventies was si- about which the cardiologist became comfort- multaneous collection of multiple leads. In the able. beginning, in the late sixties and early seventies, Nebeker: Maybe it was something the com- there were carts that collected three channels of puter could do a more accurate job of measur- leads simultaneously. In the eighties there were ing. carts that could collect 12 or 15 leads simulta- Dunn: With proper wave recognition—the neously; i.e., the standard 12 and the XYZ. Ei- points of the QRST, the ST-T segment and the ther analytic approach could be taken. baseline—the computer could make better mea- Controversy between the 12-lead and the surements than the cardiologists and save time. XYZ-lead adherents could be studied more ef- The computer could merely print out the ECG fectively when data was collected by a 15-lead measurements, and the cardiologist was already cart. The logical question was, why not use the provided some assistance. highly powerful multivariate techniques that the Bailey: When I first had to read ECGs in my XYZ-lead people were using on 12-lead data? I residency, we had little calipers where we had to think that’s where Jos Willems came into the manually make measurements on magnitudes picture. and intervals and so forth. Then we had to record Dunn: Jos Willems had been a fellow in them before offering an interpretation. The mea- Hubert’s lab when I was there. He was the bridge surements are what the computer produces auto- between the two camps. And he showed that matically.

James Bassingthwaighte Bassingthwaighte: In 1950 I was in Air for Sick Children. The first summer was on the Force officer training at Crumlin Air Base, Lon- instrument development and the second, 1952, Professor, Department don, Ontario. This led to summer research the on using it in the Burn Unit. of Bioengineering, next two summers at the Institute of Aviation University of Washington and Medicine. In 1951 I worked with A.W. * * * * * (interview by Farmer, professor of surgery at the Hospital for Frederik Nebeker on Sick Children, and Prof. Wilbur Frank at the Bassingthwaighte: We began this [in about Banting Institute, both at the University of To- 1960] as an analog computer model [to deter- 5 December 2000) ronto. I was asked to design and develop instru- mine the transfer function of a portion of the vas- mentation to measure the spreading and cular tree, namely that of the artery in the leg disappearance of fluorescein injected along with going from the femoral artery in the groin to the hyaluronidase into the skin of the forearm. It in- dorsalis pedis artery on the top of the foot]. A volved having a source of ultraviolet light and a convolution integration is essentially a digital receiver that excluded the source light and process, unless you can write the transfer func- looked at the fluorescence. In that way the fluo- tion as a differential operator, something that rescence could be looked at as a function to time one cannot do for a lagged normal density func- after injecting a tenth of a milliliter of tion. If you ever saw something strange, it was fluorescein-containing material into the skin. I doing a convolution integration on an analog designed the instrument and the circuitry and computer. You had to have a perfect gain of one contracted out to get the device built, a huge, through a system and then do repeated record- heavy thing. We observed fluorescence inten- ings onto a tape in order to do the integration. It sity to determine whether the rate of disappear- was crazy, and very tedious, but worked. ance of fluorescein was affected by the Nebeker: The purpose of this work was to hormonal level of corticosteroids in the blood. understand the circulatory system better? At that time it was thought that this might be- Bassingthwaighte: One reason is to want to come a useful clinical measure of the stress lev- understand its hemodynamic characteristics. els in children in the Burn Unit at the Hospital Another is to characterize the changes that occur when there are changes in flow through a limb. Bassingthwaighte: Physiology is an integra- Third, potentially this same femoral-to-foot tive discipline, historically speaking, so one is ex- transfer function might be used diagnostically to pected to integrate diverse observations into a assess limb flow in patients with ischemic arte- self-consistent scheme and to look at the time rial disease. courses of events to understand relationships. These The actual mathematical technology that we cannot be very well understood without having developed is not used diagnostically but is used some modicum of control systems analysis. in research. Nebeker: That’s quite a span of years from * * * * * when you began working on it. Bassingthwaighte: Most of 40 years. And Bassingthwaighte: Bob Rushmer had that kind of beginning expanded into other ap- founded the bioengineering program at the Uni- plications of mathematical modeling that we versity of Washington in 1968, and I had come to have used for image interpretation using posi- UW twice as a site visitor on behalf of NIH to see tron emission tomography and MRI, particu- whether his program should be funded. I liked larly for the measurements of flows and his program. Dr. Rushmer got to know that I had metabolism within organs. been in the market for the position at McGill, and when he decided to retire I was recruited here. * * * * * The reason I came here was because UW had one of the Mayo Clinic’s prime attributes, namely, an atmosphere in which people would “One of the intriguing things we collaborate across departmental lines freely and openly in a noncompetitive way. The University discovered very early on was that of Washington had another attribute similar to the heart is very heterogeneous in McGill: a medical school and engineering its regional blood flows. In the school on the same campus, right beside each sixties we had learned that other, and furthermore, the Center for Bioengi- neering was part of both schools. different regions had different flows. In the early seventies we * * * * * learned that the different flows Bassingthwaighte: The central theme is still stay that way when examined at cardiac metabolism, unraveling it in the normal several later times.” state and to a lesser extent in cardiac ischemia. Ischemia has been a secondary mission. In the Bassingthwaighte: I chose to look at cardiac process of getting at these nuances of metabo- metabolism, which was not being explored lism, I started, in collaboration with Harvey much at that time, and to try to develop tech- Sparks at Michigan State University, a program niques for doing that. To start that off, I started to look at purine nucleoside transport and me- doing multiple tracer indicator dilution experi- tabolism in cardiac capillary endothelial cells ments. I needed to use pairs or triples of sub- and myocytes. ATP is the prime energy source stances simultaneously. The idea is to use for cardiac contraction. Therefore looking at the reference tracers along with the tracer-labeled ingredients that make up ATP and play a role in substrate of interest. The three tracers are in- its regulation was important. Harvey Sparks was jected together in a short bolus injection into the an expert in this; I had the technology for look- coronary arterial inflow, and a sequence of sam- ing at the transport and metabolic rates. We ap- ples is taken from the outflow at 1 or 2 second in- plied tracer techniques to that. That was a step tervals. exactly in the direction I needed to take. Nebeker: What were the technological chal- Working with the group at Michigan State and lenges in getting this to work? with Ray Olsson at University of South Florida Bassingthwaighte: One was multiple to develop some new tracer labeled compounds tracer techniques. We were the first people to was a neat thing to do. That has evolved into use triple label beta counting. All my advi- more research in purine nucleosides, and I am sors told me it was impossible. It took us a lit- still doing that. I also started looking at sub- tle while. The reviewers send these [grant strates for energy supply, glucose and fatty ac- proposals] back saying it’s impossible. You ids, characterizing their pathways from the have to do it before they’ll give you the blood through to the cell, an important part of money to do it. cardiac metabolism.

* * * * * * * * * * Bassingthwaighte: The heart is interesting chine that allowed one to use models to check from a structural as well as metabolic functional data and thereby parse the transfer function of point of view. One of the intriguing things we the black box. That turned into a very good mod- discovered very early on was that the heart is eling system with automated ways of fitting very heterogeneous in its regional blood flows. equations and models to data, evaluating the ac- In the sixties we had learned that different re- curacy of the parameters that came out, and dis- gions had different flows. In the early seventies playing the solutions in real time so that one we learned that the different flows stay that way could see that the right result was being ob- when examined at several later times. There are tained. It allowed people to get a feeling for the two problems here that I struggled with for data analysis; to see it as they do it. Our whole years. One was how to measure heterogeneity. philosophy of the resource is centered on that, The other was how to ascertain its cause. With “You should be able to do your analysis and see respect to its measurement, if one uses different it happen before your eyes and understand what scales of measurement, one gets a different mea- you are doing as you are doing it.” The idea is for sure of heterogeneity. people to be able to look at the behavior of these Anyway, the blood flow distribution models and understand the sensitivity of various throughout the heart tissue is a fractal and so is parameters. This then enables the user to explore its vascular system, so we are characterizing the what the models can reveal about the system that network properties of the vascular system in wasn’t previously known. A model is always a terms of fractals. Fractal methods are good sta- mind expander, and so the simulation tool is an tistical tools. augmentation to one’s set of thinking tools. Nebeker: How has this been received? ***** Bassingthwaighte: The grant has received Bassingthwaighte: We have one system we funding for over 20 years. My original experi- developed over the years. Actually in 1967 I put mental grant, starting in 1964 and still running, together a simulation interface system called supported the early SIMCON. The Simulation SIMCON, which was short for simulation con- Resource Facility has supported the advanced trol. It was the interface between man and ma- efforts since 1978.

John Chato Professor Emeritus, Chato: I’ve worked in heat transfer pretty much He wanted to have a brain probe—used for sur- all my life— although it’s what I’d call “oddball” gery—insulated so that he could cool only the Mechanical Engineering heat transfer—such as biological heat transfer and tip. The probes were typically about 20-30 cm Department, electrohydrodynamics with heat transfer. long and 2 mm in diameter. University of Illinois “Oddball” heat transfer is where you really Nebeker: Why did he want to cool the tip? (interview by have to work with another field, such as biology, Chato: He was working with cerebral palsy, Frederik Nebeker on physiology, or electrical engineering, as with which is somewhat similar to Parkinson’s dis- 11 October 2000) electrohydrodynamics, so that another disci- ease. The hypothalamus is deep in the brain, and pline must be combined with heat transfer in or- it has been found that if a small part of it is de- der to do the work well. stroyed, the uncontrollable shaking of the pa- tient ceases. It doesn’t cure the disease but stops * * * * * the shaking typical of both cerebral palsy and Chato: I always tell my students that when I Parkinson’s disease. He was trying to get to that did my initial work I never had any problem with little spot in the brain. When the temperature of data acquisition. I had a potentiometer and wrote the nerves is reduced to below 27 °C, they essen- down the numbers by hand. However now we tially stop functioning—as when fingers ex- consistently have trouble with computerized posed to cold get numb. He wanted to insert the systems. Occasionally a computer will screw up probe, turn on the cooling, and see whether or due to a glitch in the software or an electronic not this would stop the shaking. If it did stop then disturbance. We have to be very careful check- he could apply rf (radio frequency) current to ing over our numbers to make sure that the num- permanently destroy the tissue. bers we get from the computer are correct. The doctor explained that he wanted to insu- late a long tube, not to exceed 2 mm in diameter, * * * * * so that only the tip would be cold. I didn’t think it could be done. When he explained to me why he Chato: In 1960 I took a phone call that came wanted this, I said, “I think what you really need in from a doctor at Massachusetts General Hos- is a new refrigeration system.” I went on to de- pital who was looking for a heat transfer expert. velop for him what I call “the smallest refrigera- Chato: Then later, through one of the NASA “Bioheat transfer is still a faculty grants, I spent a couple of summers out at relatively small arena. Even in NASA Ames in California and got into the ther- bioengineering, the bioheat mal regulation and thermal control of space suits transfer area is relatively small for astronauts—the cooling of astronauts in the Apollo spacesuits. One thing led to another and I compared to bio-fluid-dynamics did more bioheat transfer work. for instance or biomechanics in general.” * * * * * tor ever built,” and it worked out very well. The cooling tip was 2 mm in diameter and 5 mm in Chato: Cancer treatment with length. That was the evaporator. That’s how I heat—hyperthermia—was particularly a big got started in bioengineering. stimulus of research efforts in bioheat transfer. Researchers were trying to estimate how much * * * * * heat must be applied to the various parts of the tumor and worked on the associated problem of Chato: Probably the most commonly used deciding how best to aim the heat source, such as cryosurgical protocol is in cancer surgery. For an ultrasonic beam. For instance, in using an ul- instance, the liver cannot be cut but can be trasonic generator, the question was how to aim frozen. Therefore when there is cancer in the properly and how to move it so that it would heat liver, if it’s in one area the tumor can be frozen the tissue properly. with some rather clever techniques. The frozen volume is followed using an acoustic radar tech- nique, on-line, and when the frozen ball is big * * * * * enough one can be sure that the cancerous tissue inside has been destroyed. Chato: Another field I worked in that was “oddball,” as I told you, was electrohydrody- * * * * * namics. Nebeker: How did you get into that? Chato: The next question the doctor asked Chato: The Electric Power Research Insti- me was, “If I make the tip 0 °C so that it is as cold tute (EPRI) was formed while we worked with as it can be without freezing the tissue”—be- its predecessor the Edison Electric Institute. The cause he wanted to use rf current to create the le- project on which I worked had to do with under- sion, not freezing—“where is the 27 °C surface ground electric cables using oil as a cooling me- going to be?” My first reaction was to ask my- dium. We were looking at the heat transfer and self, “What is the thermal conductivity of flow characteristics. I took some courses in elec- brain?”—and of course there was no answer to trical fields and magneto- and electro-hydrody- that. That led me to measuring thermal proper- namics at MIT. I had some time for extra ties of biological tissues. courses, and even then I was willing to go up to strangers like electrical engineers and ask ques- * * * * * tions. What happened was that I looked at that project and said, “Here is oil, which is a dielec- tric fluid. Pumping is not possible with mag- Chato: A good example of obtaining impor- neto-hydrodynamics, but maybe we could pump tant information with simple but effective analysis with electro-hydrodynamics. Dielectric fluids occurred when I started looking at the heat transfer would be ideal for that.” effectiveness of various blood vessels. The early bioheat transfer models assumed that the most sig- nificant heat transfer occurred in the capillaries. I * * * * * applied well-known heat exchanger analysis to the various kinds of blood vessels in the human body Chato: Bioheat transfer is still a relatively and found that virtually all the heat transfer should small arena. Even in bioengineering, the bioheat occur in the small arterioles and venules; the capil- transfer area is relatively small compared to laries should be essentially at the surrounding tis- bio-fluid-dynamics for instance or biomech- sue temperature with no significant heat transfer. anics in general. However, I think it is a very im- That was an important discovery. portant field.

* * * * * * * * * * Nebeker: Have you been involved with the Chato: I remember looking at an electrical heat destruction of tissue for surgical purposes? engineering paper that had to do with a question Chato: Not as such. I was involved in burn of the electric field distribution around buried damage, but mostly in terms of legal problems. cables. It was the same problem as the tempera- Nebeker: What was that work? ture distribution around the buried pipe—same Chato: There were things like, for instance, geometry, same controlling equation. For in- someone worrying if a certain piece of equip- stance, in conjunction with this cooling of un- ment that has hot air going through it may actu- derground electric cables, I used the solutions ally cause a burn on a human body in the case of from an electrical engineering paper that was a patient who is exposed to it. from the twenties. As a matter of fact I modified the same solution to estimate the heat transfer * * * * * from a blood vessel near the skin surface.

Shu Chien Chien: [We published] a series of three * * * * * back-to-back papers in Science on [the flow Professor, Whitaker properties of blood] in 1967. … they triggered Chien: I am always interested in merging Institute of Biomedical my 30 years of collaboration with Richard [engineering and biomedicine]. Hemorrhagic Engineering, University Skalak in the Department of Civil Engineering shock is a major medical problem in inten- of California San Diego and Engineering Mechanics at the Columbia sive-care surgery, which in turn is closely re- (interview by Frederik Engineering School. He expressed his great in- lated to engineering. At the same time, we also Nebeker on 11 December terest in them, especially the paper on the flow of moved from studies on the blood as a suspension red blood cells through narrow filter pores, to investigations on individual cells. 2000) which is very similar to capillary flow. Nebeker: Do you mean in the sense of mod- Dick had a student, Tio Chen, who was start- eling of the processes? ing to work on a thesis computing the flow of Chien: Both in experiments and modeling. small particles through a narrow tube, a situation We isolated the individual cells, determined analogous to blood cells flowing through a capil- their viscoelastic properties, and modeled those lary. Dick put me on Tio’s thesis committee, and properties. Dick Skalak, two of his students through that we began to collaborate. We contin- (Aydin Tözeren and Richard Zarda), and myself ued to collaborate for 30 years and published to- published a paper about the mechanical proper- gether many full-length papers. That’s where I ties of the red blood cell (RBC) membrane. received a lot of my engineering training. So I was able to go from physiology to the biophysics * * * * * of flow properties of blood and to engineering simulation and modeling. Then I began to appre- ciate the power of simulation and modeling. That Chien: I think the problem with engineering really put my math interests into practice. modeling in the early days [the 1960s and 1970s] is that many of the modeling studies were * * * * * not done in relation to the biological reality. Nebeker: More abstract mathematical mod- Chien: Later on I realized that using the net- els? work theory and related areas one can begin to Chien: More abstract and very elegant solu- understand how the physiological systems tions to a problem, but not necessarily applica- work. There were several papers out during ble to biological systems. Therefore, it is very those years [the early 1970s] about how to simu- important that biology and engineering be late the physiological systems using the engi- brought together. neering control theory. Therefore, I took a course in control theory. * * * * * Nebeker: Was this kind of engineering sys- tems analysis something that a number of people Chien: Shelly Weinbaum and Bob Pfeffer were beginning to explore at that time? (Professor and chairman of chemical engi- Chien: That’s right. Dr. Arthur Guyton of neering at the City College at that time) each the University of Mississippi was one of those spent a sabbatical year with Colin Caro [at Impe- people, but there were not too many. rial College] working on atherosclerosis prob- I became interested in this and gave a course lems in the early seventies, and they talked about on “The Application of Control Theory in Physi- continuing that work after Shelly returned to ological Systems” in the physiology department New York. Shelly told me that Colin said, “The at Columbia. person in New York you should collaborate with is Shu Chien.” I already knew Shelly and Bob ogy, but it’s a very specialized area of bioengi- from other contacts in New York. In 1969, Dr. neering. However, when one talks about Y.C. Fung traveled from San Diego to give a biomechanics and biorheologists, it is very hard number of special lectures at CCNY, and he to distinguish these two. For example, Dr. made sure that Shelly met Dick Skalak and me Fung’s pioneering work in biomechanics is also during his visit. pacesetting in the field of biorheology. Nebeker: Can you describe their interests? Chien: By training Weinbaum was a mechani- cal engineer and Pfeffer was a chemical engineer, “I think the problem with so they were primarily interested in engineering engineering modeling in the early modeling. Both of them were interested in trans- port phenomena; how molecules move. days [the 1960s and 1970s] is They had not had much contact with biology that many of the modeling studies until they took the summer physiology course at were not done in relation to the Columbia and then went to work in Caro’s lab. biological reality.” As mentioned above, they wanted to continue in this line of research after returning to New York, so they came to me and said they wanted me to * * * * * collaborate on this project. At that time I was working very heavily with Dick Skalak, so I got Chien: In the most recent edition of this ex- Skalak involved in it too. This is a problem re- cellent book Molecular Biology of the Cell by lated to my interests because of the blood vessels Alberts et al., there is a chapter on the mechan- and fluid mechanics. I was focusing more on the ics of cells, which is a new chapter not present blood cells for quite a while, and here was an op- in previous editions. This is a clear indication portunity to look at the blood vessels and the that cell and molecular biologists are now see- cell-vessel interface as well. I was interested in ing the importance of engineering in the study initiating new approaches, and we began to work of biology. There is still not sufficient quantita- together. tive treatment in biological research, but I think Nebeker: When was this? the next generation of biologists will be edu- Chien: This was in 1974. We later submitted cated with the quantitative capability. Quanti- a proposal titled “Studies on Endothelium in tative treatments are already being used in Atherogenesis” to the NIH, and this grant appli- some of the biological fields. For example, sig- cation was funded in 1976; it is still active now nal processing is being used in the study of ion after 24 years. channels in excitable cell membranes. The opening and closing of these channels in a par- * * * * * ticular pattern of frequency distribution modu- late the function of cells such as the neurons. The analysis of such data requires the applica- Chien: Rheology is the study of flow and de- tion of quantitative methodology. I believe formation of matter. There is a Society of Rheol- such quantitative analysis and engineering ogy, which is part of the American Institute of modeling will be increasingly applied to bio- Physics. logical research. In my current research, we Nebeker: Do you know the history of the study how cells modulate their signal physiology interests of rheologists and when it transduction and gene expression in response to developed as a field? flow or deformation of the cell. The time course Chien: Biorheology probably developed in and the extent of the responses need to be ana- the forties and fifties. The late Dr. Alfred Copley lyzed by treating these molecules as a circuit or was a strong proponent of the field. He formed network, because of their interrelations and in- the International Society of Biorheology and or- teractions. ganized many congresses. He also started the journal Biorheology, which has Dr. Harry Gold- smith of McGill University and Montreal Gen- * * * * * eral Hospital as the Editor-in-Chief. The Congresses were usually attended by several hundred people and they provided a forum for Chien: One of the emerging areas in the exchange of scientific information on a lot of biomechanics is molecular biomechanics. We interesting work in the field. have begun to have the instrumentation and ca- Nebeker: It strikes me as a bridging field. pability to study the biomechanical properties Chien: It’s a bridging field. That’s right. It and behaviors of individual molecules. For ex- overlaps a lot with bioengineering and physiol- ample, the cytoskeleton network is composed of structural proteins including actin. What are the Chien: For example, there are now several mechanical properties of the actin filaments and people working in my group on bioinformatics the network during cell movement and in other because we are doing DNA microarray studies functional states? The answers to this are impor- to systematically assess gene expression in re- tant in understanding the mechanisms control- sponse to mechanical stimuli. With this novel ling the motion, mechanical stability, and shape technology, we can simultaneously search for of the cell. How do the cytoskeletal elements the expression of 10,000 or more genes, change to deform the cell during adaptation to whereas the previous techniques only allowed the environment? For instance, in the endothe- us to study one gene at a time. Bioinformatics is lial cells subjected to sustained flow, the cells needed in order to sort out the meaning of the and actin fibers become oriented and aligned response pattern of a large number of genes. with flow. What are the mechanics involved in We need people from all of these different dis- such reorganization? ciplines, and we also work with colleagues in our department and in other departments and * * * * * institutions.

Clark Colton Colton: [For my Ph.D. thesis] I chose to fo- had a debate within myself. “Well what am I go- cus on understanding the process of hemo- ing to do? I want to do biomedical work, but is it Professor of Chemical dialysis. That involved a multiplicity of things: really legit and will it be viewed as legit?” I Engineering, diffusion in blood, diffusion through dialysis knew that Professor Merrill was not universally Massachusetts Institute membranes (particularly cellulosic mem- viewed as doing solid work because of his work of Technology branes), convective transport in well-defined with blood. There seemed to be an attitude of (interview by geometries (I picked the flat plate, which was the “What kind of nonsense is this?” I wanted to most common at that time), and overall analysis continue in this direction but I also had to do Michael Geselowitz of performance. In the course of doing that I be- good engineering. I wanted to pick problems on 20 February 2001) gan measuring membrane properties in a batch where I could see an engineering component. dialysis device I designed. I did the drawings, Geselowitz: Did you want it to have an engi- the whole works. It was modeled after some- neering component so that you could explain to thing Ed Leonard had published, and he was other chemical engineers who had no interest in willing to sell me one of his devices, but Prof. Ed biomedical work the problem in terms they Merrill felt that I should design and build it my- would understand and think of as relevant? self, so I did. Colton: Exactly. We would make use of the Building something mechanical with pipe skills we learned as chemical engineers and ap- threads and screw threads was something new ply them to a new system. That characterized for me. In any event, I got into how the fluid me- some of my early work, although I rapidly be- chanics and mass transfer worked in the device. came a lot more biological or physiological than That was a kind of side study. I wound up with I had anticipated in some ways. And at that time publications on that. We did a numerical analy- I was in touch with a small number of students sis finite difference solution of the convective around the country who had obtained their transport equation. It’s a very complicated sys- Ph.D.s working on biomedical problems. When tem that had never been looked at intellectually. I graduated I knew everybody in the country That turned out to be a lot of fun. Then I made who was working on a biomedical problem. measurements with it and had an undergraduate That was only about three or four people, al- doing measurements of diffusion in blood. I then though it started to grow rapidly. At meetings or put it all together in a dialyzer in a very well-de- talks we gravitated toward one another because fined geometry. we were not part of the mainstream.

* * * * * * * * * *

Colton: At that time MIT had a grant from Colton: I continued to be interested in mem- the Ford Foundation. They hired a lot of their brane processes. My interests bifurcated and own students and made them assistant profes- trifurcated over the next couple years. Right af- sors and Ford Foundation Fellows. I got that and ter my thesis I had a summer job at Amacon Cor- I had it for one year, but when the program ended poration. That was an incredibly stimulating I was allowed to stay. Thing were very tense at environment. I was out there when fiber that time, I’ve got to tell you, because the per- ultrafiltration membranes were first developed. centages were poor. When I joined the faculty I They were looking for applications and trying to understand how they worked, and I became in- * * * * * volved in some basic work in an application that led to a new process called hemofiltration. That Colton: Right now I am focused on diabetes was really very exciting. and the area of artificial organs that I got started When I was a student, I was involved in work in about 1973. I’ve expanded that work. A lot is on membrane oxygenation through some con- going on with islets of Langerhans per se. It sulting. I had sat next to a student, Richard might be called cell biology or biochemistry or Buckles, who had worked on oxygen transport biochemical engineering, but the obvious appli- in blood. There was a request for a consultant cation is biomedical. It’s for transplantation. It’s that came while I was a doctoral student working really a blending of everything. It does have a on a device to take oxygen out of blood to power transport flavor. A lot of what I’m doing deals a fuel cell for an artificial heart. I made use of the with oxygen consumption and uptake. Someone work in his thesis in my consulting. That got me in the lab introduced me to NMR, and now we’re interested in oxygen transport in blood. using NMR to study islet properties in vitro and possibly in vivo. I’m still getting into new things— that doesn’t end—but it’s all now more “If you give students just a bit of focused in the diabetes area. engineering and get into the * * * * * applications too early, you have stolen from them the opportunity Colton: The point is, this stuff [engineering] to develop some competence. For is hard to learn. If you give students just a bit of that reason bioengineering is a engineering and get into the applications too early, you have stolen from them the opportunity really tough thing.” to develop some competence. For that reason bioengineering is a really tough thing. I think * * * * * that in most cases it’s been a mistake to develop large numbers of engineering programs. Gradu- Colton: Then there was yet another area in ate programs? Yes. Those are appropriate. And which I became involved, and that was through there you focus in certain areas and develop con- our department chairman at MIT, who was push- fidence in certain areas. The students can choose ing enzyme engineering. It was part of a larger when they decide where they want to go in what program aimed at using enzymes to carry out a subareas a department should be active, because chemical process to synthesize chemicals. We every subarea is a bit different. That’s dangerous were looking at antibiotics with enzymes using at the undergraduate level, because you run the ATP as an energy source to push the synthesis. risk—which I think has played out—of training That was very exciting, and I had a very big lab people who are only modestly competent, if at on that. all, in fundamental areas.

Eden: My uncle was not much older than 12 corset, and we had little brass buttons to put Murray Eden or 13 or 14 when he came to the U.S. He must through the buttonhole, then you could put it have been a remarkable man because he was ul- anywhere you wanted on the chest. Professor Emeritus of timately admitted to Harvard and received his Nebeker: Just to have controlled positions? Electrical Engineering, undergraduate degree, and he then obtained a Eden: That’s right. I don’t know whether he Massachusetts Institute Harvard M.D. about 1924 or 1925. He was a ever used it on patients or not. of Technology very important influence on my life and the lives Nebeker: So that was biomedical engineer- (interview by of some of my cousins. We used to pal around; ing at an early age. he would take us with him and try to teach us Frederik Nebeker something. He was a pediatric cardiologist and * * * * * on 10 November 1999) was director of a hospital for rheumatic fever kids in Roslyn, Long Island [New York]. He had Eden: I worked on two important pieces of one of the earliest of the electrocardiogram ma- work at that time [the mid 1950s]. We developed chines. He was interested in observing what hap- a special-purpose analog computer to analyze pens if you put the old-fashioned three-position the kinds of spectra that you might get in electro- connections (electrodes) anywhere other than phoresis, or anywhere the curves of the experi- the two arms and the left leg. He and I designed a mental data had peaks, under circumstances corset. It had buttons on the side so that you where you know that these peaks and valleys could open it and put it on, and it was full of but- represent concentrations of some entity, mole- tonholes. There were buttonholes all over this cules or ions usually, but with a lot of overlap of the peaks. This is the kind of problem that is Eden: By that time or shortly thereafter I left amenable to mathematical description, usually, Rosenblith’s group and set up a group with Wil- and it’s surprising what you can do. We could liam Schreiber and Samuel Mason. Mason’s take the circumstance where you had a peak and claim to fame was that he invented flow graphs. just a little teeny bit of shoulder on one side and He was an ingenious engineer with a wry sense you could estimate the concentrations of the two of humor. His research was on sensory aids for components that are overlapping with remark- the blind. One such, with Schreiber and Don able sensitivity, so long as you knew the distri- Troxel, was a machine that read printed books. bution function that the peaks followed. So it Bill Schreiber came to MIT from Technicolor. could be used in a variety of applications, in- He has worked mostly on picture processing. In cluding optical spectra where there is pressure recent years he has been one of the leading ex- broadening. We built this device, and it worked perts on high-definition TV. fairly well. It was ultimately manufactured by So we set up this lab. We had a number of grad- Dupont for a period of ten or 15 or 20 years. uate students who later became faculty including Troxel, Tom Huang, Oleh Tretiak, Jon Allen, Ted * * * * * Young, and Barry Blesser. After a while we were joined by Paul Kolers, a psychologist who had Eden: Another instrumental development I been at Harvard. We couldn’t get him an appoint- worked on [in the mid 1950s] had to do with sep- ment at the MIT Psychology Department, but he arating out groups of red blood cells based on stayed with us a while and he chose our name. We their density. Human red blood cells reach a became the Cognitive Information Processing maximal age of about 60 days. The young are Group, and it was that from about the early 1960s lighter in terms of density than the old ones, so until long after I left in 1976. We had hired some you want to separate them out by tagging the strange people. We had a psychiatrist for a while newly formed cells with some radioactive mate- who tried to model the very different world per- rial to identify a particular fraction. I also contin- ceived by very disturbed children. ued to work on what I had done at Princeton, Nebeker: Can you give some examples of basically a growth model. As far as I can tell … it things you and the others in this group worked is the first algorithmic, computer-worked-out on in those years? model of two-dimensional growth. The original Eden: Yes. Most of the things that I worked model is very easy to describe. on had to do with biology. That’s how I came Nebeker: Microbial growth, for example? back to biology; I had sort of drifted away. Most Eden: That’s precisely what I was trying to of the pattern recognition problems that we pro- model. A cell divides in two. They remain sitting posed for students to work on came from medi- next to each other. Then one or another or both cal applications, diagnostic applications, and we of them divide. There are all kinds of algorithms really worked on a whole range of problems. We you can invent. My algorithm was very straight- look at patterns. forward. I said, “One of these two cells shaped Chromosome karyotyping is a very good ex- as squares will divide, and as they’re still con- ample. When we started, it was very difficult. nected each has only three sides that are not cov- We knew very little other than each chromo- ered. It can divide right, left, or down. You build some’s size and shape. Then a biologist by the more and more.” The work on this model and, name of Caspersson, if I remember correctly, later, on others reflects the development of com- discovered that you could differentially stain the puters, because the first use of this model back in genes in the chromosome so you’ve got a bunch Princeton was on what was a computer called of transverse stripes of different widths and gray the Johnniac. There were three names: John von levels. Suddenly the problem changed because Neumann, Herman Goldstine, and Julian we now had additional information related to Bigelow associated with this computer. Bigelow structure. We looked at debris in urine. We was the engineer. We made pictures. The way looked at cancer cells in Papanicolau stains. We we made pictures prevented us from adding looked at X rays to tell the difference between many cells. The most we could do was on the or- the presence of tubercular lesions or not. der of 20 or 30 cells—the computer was too slow and the only output was a Hollerith punch card. * * * * * We would produce a picture on a punch card. The holes on the punch card are rectangular, not Eden: We independently invented comput- square, in columns. But if you look imagina- erized tomography. I think it’s fair to say we did tively you can see a picture. it independently, although Hounsfield at EMI Nebeker: An early graphical output. may have begun a year or two earlier. We pub- lished the first section of a human tissue in the 1969 International Conference on Engineering * * * * * in Medicine and Biology. * * * * * They looked like a patchwork of different col- ored squares. We could only digitize in a matrix Eden: We went over to the Brigham and of 32 × 32 or 16 × 16—very small numbers. We talked to Herb Abrams, chief of radiology, and could have done better, but it would have re- his staff. They were mostly interested in quired more computer power than we had angiography. They wanted to look at skinny lit- readily available to us. They looked at our pic- tle blood vessels in the hand or the heart or tures, which were full of quantizing noise, and maybe elsewhere. Our pictures were very crude. said, “Well, we have no interest.”

Fung: Dr. Sechler, my mentor [at Cal Tech], Fung: I became interested in biomedical prob- Yuan-Cheng “Bert” Fung asked, “What would you like to work on?” lems for a personal reason. In 1957 I took a sabbat- Professor Emeritus of “Well,” I said, “I would like to work on ical leave with a Guggenheim Fellowship and Bioengineering, aeroelasticity.” “Ah,” he said, “just right. In went to Germany. My mother had acute glaucoma, University of California, 1942 the Tacoma Bridge on Puget Sound in a very painful disease, and I could do nothing for San Diego Washington State was blown down by wind. her. In frustration, I read American literature on Von Kármán said that the oscillation was caused glaucoma and sent her a weekly translation or (interview by by vortex shedding from induced aerodynamic summary. I told her, “If you cannot use it, give it to Frederik Nebeker forces. The State of Washington Public Works your surgeon.” Many years later, in 1973, I finally on 12 December 2000) sponsored a research project here. When von went back to China. My mother’s operation was a Kármán retired, that job was given to Dr. Louis success. Her surgeon thanked me. Dunn. When Dr. Dunn became the director of In Germany, right across the street from the the Jet Propulsion Laboratory a few months Aerodynamics Research Institute was the Physi- later, he handed the job to Maurice Biot. Biot ology Institute of the University. soon left. They left a filing cabinet here which Nebeker: This is in Göttingen? you may take a look at.” Fung: Yes. It was the Göttingen Physiology I inherited a wind tunnel designed particu- Institute. I enjoyed their library and their facili- larly for this project. I used it to study the bridge. ties. That’s the beginning of my interest in that I found bridge aerodynamics very difficult, very field. Gradually one thing became quite clear to awkward. Airfoils are simpler and cleaner. So I me: the biologists do not think about what we began working on aeroelasticity of the wings. I engineers always think about—namely, the formulated the general aeroelastic problem and force, motion, and transport phenomena. Fur- then designed ways to systematically study it thermore, biology is full of interesting nonlinear step by step. problems. The field seems to ignore the kind of Nebeker: You were interested in a general things which you can do. I found the field very theory of aeroelasticity rather than in how a par- attractive. ticular structure responds? After returning to Caltech I began to work on Fung: Yes. I wrote one of the first books on physiology. A few years later my feeling was aeroelasticity. that I wanted to work on it full time; I didn’t want to dilute it with other work. That’s how I ended * * * * * up at UCSD.

Nebeker: Can you tell me how you first be- * * * * * came interested in biomedical problems? Fung: Now, the mechanical properties of bio- logical tissues were virtually unknown in the 1940s. The question was how to describe the me- chanical properties and what to measure. Today you would consider successively the hierarchies of organelles, cells, interstitial materials, tissues, organs, and the individual. Historically, the study was done in downward order of hierarchies. Every step is difficult. For example, it was very difficult to quantify the mechanical prop- erty of tissues, but without that you cannot be- gin. With testing, theorizing, and hypothesizing, you hope that you’ll eventually understand the Bert Fung working on tissue remodeling with materials that you are dealing with. Wei Huang, Ghasan Kassab, and a student at the University of California, San Diego. * * * * * Fung: Capillaries have a diameter of about blood flow does not apply to the lung. It showed ten microns (about one-third the thickness of our that it is simpler to think of a continuous sheet hair) and a wall thickness of one or two microns. with a pattern of obstructions. Dr. Zweifach and his associates could not detect the change of the capillary diameter when the blood pressure was changed by as much as 100 * * * * * millimeters of mercury. Since they used an opti- cal microscope for the measurement, this means that the change of diameter is much less than one Fung: [At UCSD] we had tremendous rela- wavelength of light, or 0.5 micrometer. tionships with the surgeons. Nebeker: Isn’t it surprising, because such a Nebeker: More so than with other doctors? thin structure should respond to that change in Fung: I think so. Surgeons really see imme- pressure. diately that engineering is what they need. We Fung: Yes. Very surprising if we think of the had very good collaborations with the surgeons. capillaries as small cylindrical tubes. Collagen The barrier was broken down almost without ef- or steel tubes can do it, but not a single layer of fort. I think we were extremely lucky on that. endothelial cells with a thickness of one to two microns. Where did this rigidity come from? I looked at the problem and my suggestion was that the ri- gidity came from the surrounding media. In * * * * * other words, the capillary blood vessels we have are not a tube, but a tunnel lining. To verify that the rigidity comes from the tunnel material outside, I measured the mate- rial properties of the stuff outside the capillary, Fung: Microcirculation is a big field, but then computed the expected deformation of the really the organs they study are not many: the capillary. I found agreement. That kind of thing mesentery, the cheek pouch, the skin flap, and makes engineering thinking suddenly relevant. a few others. Everybody is interested in mus- cle, but there aren’t many muscles tested. First the frog. * * * * * Nebeker: That goes way back. Fung: When A.V. Hill worked on muscle mechanics he used British frogs and got poor re- sults. One day he visited Italy and got the Italian Fung: The next thing I wanted was to deter- frog; then [his experiments] worked. mine the elasticity of the capillaries of the lung. Nebeker: So that was the advantage Galvani Then … work out the implications of the new had. understanding on the physiology, pathology, Fung: There are really not many models that diseases, and injury of the lung. In this endeavor, people make observations on. I was fortunate to have the collaboration of a fa- mous physiologist, Dr. Sidney Sobin, and my former student and colleague, Michael Yen, and many graduate students of ours. This was where the payoff from an engineer- * * * * * ing point of view of physiology came in. If you look at the lung, you see that it is an organ of cap- illaries. These capillaries, however, are not sin- gle long tubes; they are organized into Fung: I think biomechanics will be there al- two-dimensional sheets. The geometry is ways. No understanding can be reached if force unique. We had to measure the dispensability of and motion are ignored. I don’t ever think that the sheet, analyze the flow, and understand the application can be minimized. Applications al- physiology. ways have surprises of some real good important Nebeker: Even the microanatomy wasn’t discovery. fully known at that time? Nebeker: From the purely scientific view- Fung: The Sobin-Fung sheet-flow model of point. pulmonary capillary was novel. It changed the Fung: Even from the scientific point of view. language of pulmonary microanatomy. It Biomechanics is the middle name between showed that the classical Poiseuille formula of structure and function. Geddes: My first paper described the RC for a six-week summer course with stipends for Leslie Geddes AMC electromyograph. It was published in tuition and expenses. We had people from all April 1945. I was working on it in 1944. It was over the world as well as the United States. This Showalter Distinguished quite an elegant device with a differential ampli- National Heart Institute training program was Professor, Emeritus, fier, a loudspeaker, and oscilloscope. There the longest single managed training grant that Department of were a lot of soldiers returning home with nerve NIH had ever funded up to that time. It ran from Biomedical Engineering, injuries. You could diagnose nerve injury with 1956 to 1974—18 years. We trained about 1,000 Purdue University the electromyograph—just place some elec- people. (interview by Frederik trodes in a muscle and one could detect a charac- teristic electrical activity. That was what I was * * * * * Nebeker on doing during the last year of the war at McGill. 13 October 2000) I then joined the Montreal Neurological In- Geddes: Let me tell you something about the stitute in 1945 as a research assistant and stayed early days of ventricular defibrillation. The first there until 1952. I was an instructor in the Elec- human transchest defibrillation was in 1947. By trical Engineering Department at McGill from about 1954 our students had all done ventricular 1945 to 1946. I was one of the early biomedical defibrillation with a physiograph. We had sev- engineers. They didn’t know what that was at the eral calls in 1957 and 1958 from the local hospi- time. It was all medical physics at that time. I tals where we had former students. They would was responsible for the operating room equip- call and say, “Send us over the physiograph ment and for the electroencephalography labo- defibrillator and the paddles. We have a patient ratory. in fibrillation. We’re doing cardiac compres- sion. Hurry up. Send us the defibrillator.” So we * * * * * sent the defibrillator over and they put the pad- dles on the heart and defibrillated. Geddes: At Baylor I built the physiograph. Nebeker: There were few companies out The teaching of physiology was in the dark ages there producing any kind of defibrillator at that then. Everybody used a smoked drum (the Lud- time. PHOTO BY VINCENT WALTER wig kymograph) to record only mechanical Geddes: As a matter of fact, it didn’t really events. So, I built the physiograph as a modular start until 1952 with transchest defibrillation electronic system that permitted recording many with a 700-volt, 60-cycle alternating current different types of physiological events. This is defibrillator. Then Lown did it in 1962, ten years where my Northern Electric experience and later, with a damped sine wave. And so, human ham-radio experience paid off. The Bell Tele- defibrillation didn’t start until the early 1960s. phone System has a philosophy that any new de- vice that is made must fit with the old system—it * * * * * must be compatible. So a modular concept was the keynote for the physiograph, which had a Nebeker: What was the response of the med- transducer, an amplifier, and a display device. ical community [to your publication of the One could connect a variety of different compat- defibrillation dose concept]? ible transducers to the amplifier, which ener- Geddes: You see, cardiologists only saw gized the transducer and amplified its signal. adults. They knew they had difficulty One could then have the event written out on a defibrillating 200-pound subjects. But they graphic recorder, or displayed on an oscillo- could defibrillate most of the subjects with the scope, or heard by a loud speaker. We displaced defibrillators they had. Turn it up to the top and the smoke drum kymograph around 1954 or so. we defibrillate all. So, who is interested in the dose concept? That fight went on from 1974 for * * * * * about the next eight or so years. It was only when they decided to make defibrillation efficient be- Geddes: This system created so much atten- cause they had to implant them that they began tion that the NIH funded us for a training pro- to know that big hearts are difficult to gram which originated in the following way. defibrillate. Small hearts are easy to defibrillate. My mentor (Hoff) and I were sitting in the Washington National airport about 1955 or so. * * * * * We had three hours to kill. We sketched our ideas for a summer course on physiology on the Geddes: I was a consultant for the VSAF at restaurant menu. Brooks AFB in San Antonio, Texas, for many On arriving home, we applied to NIH for sup- years. In the early 1960s, when I was at Baylor port to teach modern physiology. The program Medical College in Houston, NASA contacted was titled, Classical Physiology with Modern me and invited me to be a consultant on physio- Instrumentation. The concept was like “Shake- logical monitoring of astronauts before there speare in Modern Dress.” We received the funds was a suborbital flight. I, along with other scien- tists (Pat Meehan and Col. Jim Henry), met with Nebeker: Would you tell me a little about a few NASA people in a garage rented from how that book came to be? Homco Oil Co. located at the intersection of Old Geddes: That book came as a result of teach- Spansh Trail and Wayside streets in Houston. ing medical students, residents, and interns, and Then there was no Manned Spacecraft Center. the need to make devices for research. It also We all discussed the important physiological came as a result of the summer course “Classical events that we should measure and telemeter to Physiology with Modern Instrumentation.” earth from the astronauts. There was just no way you could buy devices to From the outset it was obvious that it was de- make the measurements. Lee Baker was my sirable to telemeter the four vital signs: tempera- graduate student. We decided that we would put ture, pulse rate, respiration rate, and blood this all down so people could learn how to make pressure (T, P, R, and BP). It turned out that the their own devices from a book. It’s really a mea- chimps on the sleds at Holloman AFB would not surement handbook. tolerate a rectal thermometer; they unceremoni- Nebeker: It’s become a real classic with ously ejected it. The heart rate could be detected three editions. from the ECG. Respiration was detected by the cooling of a heated thermistor on the micro- * * * * * phone in the helmet when the astronaut exhaled. In the first flights, indirect blood pressure was Nebeker: You were saying that it is very im- not obtained due to the cramped space. Accord- portant as a teacher to provide historical back- ingly, only ECG heart rate and respiratory rate ground. were telemetered to the earth-bound monitoring Geddes: Yes. I think it is important because stations. But there was a problem because the by knowing that, down the road the student will respiration signal was lost when the astronaut perhaps be able to make the next leap from the turned his head away from the microphone. At knowledge that he or she has now. It gives them this point NASA contacted me and asked if I the thought that discovery is an evolutionary could devise a more reliable method of detecting process. For every discovery, there is a little respiration. On a Saturday afternoon at Baylor, I background and there is going to be a future. We went into the lab, placed two electrodes across are not at the end of the discovery road now. my chest at the xiphoid level, and measured the Nebeker: Occasionally, an earlier approach 20-kHz impedance changes that occurred with that didn’t result in anything useful might be- respiration. The method was described later in come feasible when new technology arrives. Aerospace Medicine in 1962, and we used it in Geddes: That’s exactly the point. New tech- the medical student physiology laboratory; the nology is around, and one can do things now that method became known as impedance were very difficult to do previously. The knowl- pneumography. edge was around, but the execution was very dif- ficult. * * * * * So, I always start my lectures with the history and what was done and why it was done. To just Geddes: During my time at Baylor, I had al- say what was done is not enough. One has to get ways been interested in the measurement of di- into the mind of the person who did it to find out rect and indirect blood pressure. I published two why he did it. books on the direct and indirect measurement of blood pressure. We actually developed the * * * * * oscillometric method, which is the standard method used now. We did the study that showed that when cuff pressure is decreased, the ampli- Geddes: I don’t know that biological engi- tude of cuff-pressure oscillations increased, neering it is getting fragmented. It has changed. reached a maximum, and then decreased. The The transactions on biomedical engineering are maximum oscillation point signals mean pres- becoming a little more theoretical and more con- sure. cerned with modeling than they used to be. The trouble is that doing experiments is expensive. It * * * * * is easier to sit at the computer and put data in and model. It is expensive to take those data and sub- Nebeker: Was it in 1968 that the first edition ject them to the real-world animal or human test. of your book with L.E. Baker on medical instru- So that is a change that I am not particularly mentation was published? happy to see because I like to see theory and Geddes: Here is the first edition, 1968. practice joined. Nebeker: I understand that while you were that box is always there. They roll it over at night Wilson Greatbatch at Cornell you started working with some instru- and they can never forget that it’s there. mentation for animals for the Psychology De- President and CEO of partment. * * * * * Greatbach Enterprises Greatbatch: Yes. At that time Cornell’s (interview by Psychology Department was one of the coun- Greatbatch: Five years later when transis- Frederik Nebeker try’s foremost centers of Pavlovian psychology, tors had been invented and had become readily on 4 April 2000) which is a physiological psychology. I got a job available I thought, “Now I can make a pace- building amplifiers for a hundred sheep and maker.” goats, measuring their heart rate and blood pres- Nebeker: Can you recall dates? sure and so on. I also participated in conditioned Greatbatch: It was somewhere between reflex experiments. My experience there turned 1949 and 1951 when I learned about the disease. out to be quite useful. Because I knew what con- We built our first pacemaker with transistors for ditioned reflex was, a few years later I had the implantation into an animal in 1958. Much to opportunity to build all the amplifiers for one of my wife’s consternation, when I realized I could the first monkeys shot up into space. That was actually make pacemakers I quit all my jobs. I what really got me into the field of medical elec- had two thousand dollars in cash, which was tronics. enough to keep my family going for two years [1958 to 1960]. I gave the family money to my wife and went up in the barn behind my house. In “One of the first pacemakers that two years I made 50 pacemakers. ran on a lithium battery was Nebeker: You made them by yourself? implanted in a patient in Greatbatch: Yes. Nebeker: Did you make any attempts to in- Australia. That patient went out terest a medical equipment manufacturer? I into the outback and became sort would think a young engineer might go that of lost from civilization. They route with a good idea like that. finally caught up with him and Greatbatch: I considered that and I did talk to some people, but no one was really interested discovered that his pacemaker in it. Of course we hadn’t implanted devices in had been running for 22 years.” our first patients yet. Once we had our first pa- tients with the device successfully implanted, there was a great deal of interest. We worked * * * * * fairly closely with Earl Bakken and his Medtronic Company at that time. Nebeker: I’d like to hear more about your instrumentation for these animals. Were you * * * * * first building amplifiers for the EKG? Greatbatch: Yes. The heart generates a sig- Nebeker: How did the connections with nal which is about a millivolt in amplitude, so Doctors Chardack and Gage come about? this has to be amplified by a factor of something Greatbatch: Our local chapter of the PGME like a thousand before it can get up to the point [Professional Group on Medical Electronics of where it will record. At that time we didn’t have the Institute of Radio Engineers] in Buffalo had transistors yet, so all our work was done with a meeting every month. Ours was the first chap- vacuum tubes—which is not the right way to do ter in the country. We tried very hard to get an this sort of thing. equal number of doctors and engineers to attend the meetings, and sometimes we had as many as * * * * * 50 people at a meeting. The engineers offered to send a team of engineers to help any doctor that Greatbatch: A friend of mine in Minneapo- had a research problem—for free. Quite a few lis, Earl Bakken, was building battery-operated doctors took advantage of that offer. In fact, pacemakers that could be worn on a belt. They about five different medical doctor-engineer had wires going right through the skin. That was teams that resulted from that stayed together for only marginally satisfactory, though some peo- quite some time. I went up with one of those ple lived a long time with them. We engineers groups to see Dr. Chardack. have not to this day learned how to run a wire through the skin and have it seal. It’s always an * * * * * open wound. Therefore the people have to put antibiotic jelly around it every morning and ev- Nebeker: It sounds like you became very in- ery night. They can’t bathe, go swimming or terested in battery design. take showers. They can take a sponge bath, but Greatbatch: Yes. Once we realized that the battery was the limiting factor, we started look- to treat complete heart block. That was almost ing at all sorts of power supplies. We even unheard of in the industry. looked at rechargeable batteries. However, we found out that the life of a rechargeable battery * * * * * with recharging was not as long as the life of our primary cells without recharging. That knocked Nebeker: Was Wilson Greatbatch Ltd. es- out that idea. tablished in 1970? Nebeker: Would this have been a kind of re- Greatbatch: Yes. It has expanded ever since charging that could be done through the chest? with new battery models and more people until Greatbatch: Yes, inductively through the today. We either make or license over 90% of all skin. That would have been no problem. the batteries in the world that go into pacemakers and implantable defibrillators. * * * * * Nebeker: Was that always your market? Were you always exclusively aimed at these Greatbatch: We came up with the lithium implantable devices? battery in 1970. The lithium battery had a much Greatbatch: We have diversified some- longer life than mercury. And it didn’t generate what, but it took just about all of our abilities to gas, so it could be hermetically sealed. The fact satisfy this one market. We also make batteries that it didn’t have gas was a big deciding factor. for use in outer space. Any time that the astro- The fact that it could be hermetically sealed was nauts go outside of their bird they carry with an even bigger factor, because that meant we them a television camera, a life support system, were no longer running under water. One of the and communication equipment—all of which first pacemakers that ran on a lithium battery run off of our batteries. We also make a battery was implanted in a patient in Australia. That pa- that will run at 150 °C. They are taken down to tient went out into the outback and became sort the bottom of oil well holes to log up the results. of lost from civilization. Last year they finally We have specialty markets like that, but our big- caught up with him and discovered that his pace- gest business is in implantable medical batteries, maker had been running for 22 years. implantable power. Now we are into implant- able power for artificial hearts. * * * * * * * * * * Nebeker: Did the paper published in Sur- gery and other publications create an immediate Nebeker: What was the rationale for freely demand? licensing rather than selling your own patents? Greatbatch: Yes. Of course we also went to Greatbatch: Building confidence in the all of the shows. Medtronic had a booth at every field was our main objective. We didn’t want to show. I published in the IEEE/PGME journal, have an image like that of Microsoft. We wanted and I published any new battery developments. our customers to be good friends. We are very Dr. William Chardack and I published jointly in careful in quality control and look for and find the medical journals. I used to accompany a the one bad battery in 10,000 before it gets out of Medtronic salesman and visited a lot of hospi- the factory. People like that, and the FDA is tals, and I gave talks at their research seminars. agreeable to it too. We are very careful. If there We gave a very strong educational push in those is a problem, our customer finds out from us first five years. It is interesting to me that it took rather than from their customer. We’ve built a only five years for the pacemaker to become uni- good relationship with our customers, and part versally “indicated”; that is, accepted as the way of that goodwill is from freely licensing.

Dov Jaron Jaron: While I was an undergraduate stu- to measure psychological responses. Dr. dent I worked as an electronics technician at the Offner built it for him. Calhoun Distinguished Medical Center of the University of Colorado Nebeker: Do you know what sort of psycho- Professor of Engineering in Denver. I worked for a psychiatrist named logical responses? in Medicine, School of Sidney Margolin, who actually was trained un- Jaron: Yes. ECG, EKG, galvanic re- Biomedical Engineering, der Freud and did what was called sponse—any electrical response that is mea- Drexel University psychophysiology. He lived in New York for sured today, he had the first unit that did that. I many years and treated some of the most fa- maintained this unit for him. It was built with (interview by mous movie stars as a psychiatrist, then he de- tubes and all done by hand. Frederik Nebeker cided he had enough of the city life and moved on 6 October 1999) to Denver to be on the faculty. He was an elec- tronics nut. He had the first electronics system * * * * * Jaron: If you have a current generator inside at the same time you are reducing the workload the body, it generates potentials on the surface. on the heart, making it easier for the heart to re- The electrical activity of the heart is the equiva- cover from a heart attack. lent of such a generator. You can measure the potential distribution on the surface of the body Nebeker: You worked with Kantrowitz in ac- because you have access to the surface. The tually building this device? question is, from the surface of the body, can Jaron: That is right. We actually produced you deduce what happens to this generator and the balloons. We made the drive units, tested the in particular, when there is a problem with it. system on animals extensively, and then used it That is done to a certain extent with electrocardi- on patients. ography—that is what electrocardiography is all Nebeker: What year was this that you were about. But electrocardiography has serious limi- first using the device? tations. [David Geselowitz], one of my Ph.D. Jaron: It was 1968. We had done maybe five advisors, was trying to figure out whether there or six or seven patients, and the word began to is a better way of getting more information about spread around the country and people started the generator. We realized that an implanted calling us and asking for our team to come and pacemaker can be used as a model to study the treat a number of very famous individuals. relationship between an internal generator and the potential distribution on the body surface that results from its action. With a pacemaker, * * * * * we knew exactly its properties as a generator. We knew exactly what the current of the pace- maker is, what the voltage is, so we tried to mea- Jaron: Ever since I was president of the sure surface potentials and relate those surface EMBS I have pointed out that the EMBS focuses potentials to the current generated by the pace- only on those professionals whose roots are in maker. electrical engineering. Biomedical engineering has become much more than electrical-engi- neering oriented. As a matter of fact, many bio- * * * * * medical engineers now may not even know what an electrical circuit is. There are many more me- chanical-, chemical-, information-, computer-, Jaron: I then [in 1967] went to work with a and materials-based biomedical engineers. I heart surgeon by the name of Adrian Kantrowitz predicted it ever since I became president of the at Maimonides Medical Center in Brooklyn, Society, and I have urged the Society to move New York. He had just received a very large more and more in that direction. grant from the NIH to develop and test heart-as- sist devices. Together we developed the intraaortic balloon pump. The intraaortic bal- * * * * * loon pump is a plastic balloon, maybe 20 centi- meters long, mounted at the end of a hollow catheter. It is introduced through the femoral ar- Jaron: We are entering a new era. When we tery all the way into the descending thoracic finish sequencing the genome, we will begin to aorta. The balloon is inflated and deflated in syn- understand how the genes relate to proteins and chrony and out of phase with the heart’s action. how they relate to cell function. How the cells When the heart begins to eject blood (systole), behave, how tissues behave, and finally how the balloon is deflated, thereby creating a lower whole organs and the organism function. I think resistance in the circulation, making it easier for the only people that will be able to put this pic- the heart to pump the blood. When the ejection ture together are biomedical engineers. The way period of the heart is finished and the aortic scientists approach a problem is by the valve closes (beginning of diastole), the balloon reductionist approach. They break the system is inflated. This displaces the blood to the pe- down into its smallest components and concen- riphery, but more importantly, also towards the trate on the very small elements on the system. heart. The coronary vessels get all their blood Engineers try to integrate. Biomedical engineers supply during diastole, when the heart is re- who understand the living system should be able laxed, which is the time when the balloon is in- to integrate facts and knowledge about the sys- flated. As a result, you are increasing the blood tem’s components and begin to understand how supply to the heart and also to the periphery, and the whole system functions. Richard J. Johns Johns: The action potentials [in neurons] with strain gauges on it that measured the force, that you evoke from this are in the millivolt re- and then there was a sinusoidal rotational drive. University Distinguished gion, so you had to have very good differential The bottom line was that the stiffness of rheuma- Service Professor, amplifiers. Otherwise, if you charged up the in- toid arthritis is a visco-elastic stiffness and Johns Hopkins put capacitance of your amplifier, all you would nonfrictional—there is no excess friction. It was Medical School be looking at is the exponential decay from the this visco-elastic stiffness that increased in the (interview by stimulus artifact. You had to have very good dif- morning. It was an actual swelling in the joint Frederik Nebeker ferential amplifiers with low input capacitance capsule that produced it. to be able to do that. Those were some of the on 26 April 2000) technical parts of it. * * * * * Our research interest was in finding out the nature of the disorder of neuromuscular trans- mission in myasthenia gravis. There was a big “One of the great unexplained controversy at that time as to whether the prob- lem was in the nerve ending with not enough things was the nature of joint acetylcholine being released. Or, whether the stiffness in rheumatoid arthritis. disease was of the musclemotor end-plate and The thought was that because the that a normal amount of acetylcholine was being joint cartilage was destroyed, it released, but it was insensitive to it. Nebeker: Were you able to answer that ques- was friction that gave them the tion? stiffness.” Johns: Yes. The way that we were able to an- swer it was by comparing the effect of externally Johns: I was the national president of the injected acetylcholine in normal subjects and in IEEE Group, too. It was then called the Group on patients with myasthenia gravis, when you put Engineering in Medicine and Biology, and in the same amount in. Since acetylcholine is 1973 and 1974 I was the chairman. quickly destroyed in the blood, the trick was that The IEEE Engineering in Medicine and Biol- you had to put it into the artery so that it would go ogy Society was primarily the electrical engi- directly to the muscle very quickly before it was neers in biomedical engineering. They were a destroyed. That showed very clearly that there fairly diverse group. They weren’t all circuit de- was a diminished response to injected acetyl- sign for biomedicine people, and there were choline. So we knew that the problem was not in many of the Dave Robinson types (the guy the nerve but in the motor end-plate, and that studying control of eye movement who was an subsequently was shown by other methods as electrical engineer). They didn’t think much of well. chemical engineers and the kind of stuff that they did. I was okay, coming from medicine. * * * * * Nebeker: So an M.D. had no trouble? Johns: Yes. But a chemical engineer, they did- Johns: One of the great unexplained things n’t quite see what they were doing in that. Then the was the nature of joint stiffness in rheumatoid ar- Biomedical Engineering Society was primarily thritis. The thought was that because the joint people in the systems physiology business. cartilage was destroyed, it was friction that gave them the stiffness. Nobody could understand * * * * * one of the real characteristics of rheumatoid ar- thritis, which is called morning stiffness. People Johns: [At Hopkins] we were talking about would wake up and their hands would be very myocardial contractility and how to measure it. stiff, and then as the day wore on their stiffness Some of the myocardial contractility people and would decrease. Well, there is no reason that the the MRI people and the image processing people friction should change, and that did not make came together. What you can do now is, with sense. magnetic resonance saturation, in essence, put a We developed a method of measuring joint whole bunch of planes through the heart and stiffness with a hand-holder, which is something then watch those lines, those planes, move as the that mechanically moves the finger so as to have heart beats. They said that we could see if there passive movement of the finger sinusoidally. are some areas of the myocardium that are not We would impose a sinusoidal rotation, and it contracting. Then the imaging folks figured out would sense the force that was required to do how to create a pseudo-color display, such that that. the less it contracts the bluer it is, and so forth. Nebeker: A device to move the finger and Then they get the cardiologist in on it doing thal- sense the force that is encountered? lium scans to see if that is where the thallium up- Johns: Yes. The other thing which is impor- take is wrong. It does make a good environment tant is the axis of rotation. It was a cantilever bar for doing collaborative stuff. * * * * * big-ticket items but often the very routine items. At that time, the most expensive thing was CT scans, and the cheapest was X ray. If you could Johns: Then I was personally interested in reduce your chest X rays by 10% you would the clinical information system business and save more money than if you eliminated CT was involved with those. In one piece of that, scans entirely because of the tremendous vol- Don Simborg and I were interested in seeing ume. We tested to see how hard would it be to about using information to save costs. It turns save 10% in that. It turns out, I think, 3% of all out that the biggest wastes of money in terms of the X rays were taken of the same person on the laboratory and ancillary services are not the same day.

Katz: For instance, how did the motion of at- talks on hemoglobin and myoglobin for which J. Lawrence Katz oms or molecules relate to elastic properties, a they won the first Nobel Prize for the structure of field called lattice dynamics. We worked on zinc proteins. Professor of Biomedical from liquid helium temperature all the way up to I had kind of an epiphany attending that Engineering, just below the melt, and measured the properties meeting with Perutz and Kendrew in England Case Western Reserve both in plane—how the zinc atoms vibrated— where I said, “Gee, fantastic, you can use these University and perpendicular, and that gives you a relation- techniques and really work on living things.” (interview by ship to the internal forces. That started me in Nebeker: And you felt real breakthroughs Frederik Nebeker good stead. The way I got into bioengineering is were possible in biology. I saw this announcement to apply for a science Katz: Yes, that is right. I think these other on 15 October 1999) faculty fellowship. I had only been at Rensselaer old-timers had the same kind of feeling. Sud- a couple of years. It was not so much research, denly they say, “Hey, with the kind of things that but you went some place to learn what to do for I am doing I can make a contribution to under- teaching. The English system for teaching crys- stand how it may be possible to improve the tallography was very good. I wanted to go to quality of life.” I really think that has been the University College, London, to work with driving force, even with the young ones now Dame Kathleen Lonsdale, the first woman who take it up as a career. member of the Royal Society and a Dame of the British Empire. * * * * * While I was there we received an announce- ment of a minisymposium at the Cavendish Lab Katz: I did have a state-of-the-art laboratory. where Perutz and Kendrew were giving their I found out later I had one of the only labs in the world at that time which had a computer-con- trolled system to measure the crystal properties. This is 1963 and early 1964. I had one of the early DEC-PDP8s, and I had some brilliant freshman and sophomores because we had an honors program where undergraduates could pick a lab to work with. Since I had one of the best computers on the whole campus of Rensselaer at the time, my lab was a focus. The PDP8 was driving this device, and it was dump- ing the information into another early computer, an IBM 1620, with a tape drive. When we re- ceived the money from the Institute of Health I could upgrade it to a card-driven computer to speed up the way the Fourier calculations were done. I started this working with very bright stu- dents. The beauty of it was that I did not have to write a research proposal. If I wanted to do something, I could take on a student as a trainee. The idea was to train them in how to do research.

PHOTO BY JIM THOMAS We began by trying to understand how bone and J. Lawrence Katz’ major interest is study- mineral form. Very early on that it became clear ing how and why bone structure forms and also that you had to understand something about how it relates to properties. how to measure these things nondestructively, and that is how I got into the ultrasound work. Katz: Dentistry was the leading [area for im- RPI was a good place to be for that. Dr. Hunting- planted materials] because they already knew ton during the war years worked at the MIT lab about amalgams and they already had the poly- where they developed sonar for anti-submarine mers to make impressions, and so they knew warfare, and he had used those techniques to these things could be used inside of biological measure the properties of single crystals. This is tissues. Already, some orthopedic implants were the way you measure the elastic properties, the being made out of steel and cobalt chrome al- relationship between the sound wave and elas- loys, even produced commercially, because ticity. You learn about the wave equation in any some of the post World War II people began to kind of instrumentation electrical engineering try implants when people had accidents and so course, then the diffraction work to look at struc- on in order to replace damaged tissues. There ture. It turns out there was important work con- was also the neat external orthotics materials sidering that there was probably an electrical and prosthetic materials for all of the war inju- driving force in the development of bone. So this ries which happened after World War I, but even thing began to integrate looking at mechanical more after World War II. properties, electrical properties, and ultrasound measurements. * * * * *

* * * * * Katz: That time period in the 1970s was when the space R&D went down, and they be- Katz: Our publications in the mid to late gan to fire a lot of the solid-state and material 1960s on some of these ultrasound properties scientists. The National Academy of Science were the first saying, “these are the properties, (NAS) had a special meeting in Washington to but this is how they relate to structure.” Prior to see what could be done in applications and mate- that was, “here are the numbers, this is the value rial science. I was invited to speak about of the elastic modulus of a bone or the elastic biomaterials. They had a whole session on bone, modulus for a tooth,” because they wanted this and they had some other people there. Starting in general information to see how it related to the the late 1960s, when all the societies had their amalgam. Coming from the solid state, crystal annual meetings they had a symposium on physics, my interest was very basic, as what is biomaterials—chemistry, physics, and the old the relationship between structure and proper- Institute of Materials and Mining Engineers. It ties? I essentially first developed the conception was almost the same working group giving the that for bone you could not just simply consider talks at each symposium. it as a mixture of the inorganic apatite and the or- Nebeker: The idea was that some of these ganic collagen. Because there were so many dif- materials people were no longer employed in the ferent levels of structures involved and each one space race. contributed to it, you had to consider it as a hier- Katz: So they wanted to find out where to go. archy. That way you had to understand how the Even before the National Academy of Engi- apatite crystallites and the collagen organize and neering (NAE) existed they were concerned, be- how they form their joint structure. cause a lot of them were scientists, chemists, and physicists. Biomaterials at that time was begin- * * * * * ning to blossom.

Katz: When I started I called my laboratory * * * * * the Laboratory for Crystallographic Biophysics. When I began to go to some of these meetings Katz: For instance, I got into acoustic mi- like the ASME meetings and other engineering croscopy because an electrical engineering col- meetings, they said you are doing biomedical league at Rensselaer, who worked on surface engineering, and I had to say, “What is that?” In acoustic waves, was developing a transmission the 1960s, being a physicist, I said I am working system in the early days and he thought it might on “bio” things so it is biophysics. If you are in- be useful for us in bone. I said, “You bet it is use- terested in bone as a structure and its properties, ful!” Before we were only measuring average then that is engineering. Then eventually the properties. We would take transducers and put it school changed my title from Professor of Phys- across the material to get its property on the av- ics to Professor of Biophysics and Biomedical erage. In that case the resolution went down to Engineering. So early on I had a bioengineering 100 micrometers as opposed to millimeters. This title, probably even before there were any de- was very exciting, and that led us into acoustic partments of biomedical engineering. microscopy full time.

* * * * * * * * * * Katz: My major interest is still to understand tory of publication and being at the cutting edge how and why the bone structure forms, and how in this, my colleagues who themselves are not it relates to the properties. Any tool or device biomedical engineers but who have something that will do that, I will either work with those like this will say, “I have this thing and it looks people who use it or work with it myself. Each like it should be related to your work.” I say you case is like a bug zapper that attracts bugs to bet it does, and we start working on it, then we it—because of my interest and of having a his- merge our expertise.

Mates: By then [in the early 1970s], the Mates: I spent a year in the physiology de- Robert Mates space program thing was starting to wind down a partment and I learned how to operate on dogs, little bit, at least the basis research part of it, and and do surgery, implant flow meters around cor- Professor of Engineering so money was starting to dry up. I was kind of onary arteries, and all those kinds of things. I be- Emeritus, looking for new things to do, and, again fortu- came much more independent then. Using that State University of itously, I became very good friends with David background I was able to do a lot more detailed New York Buffalo Greene, a cardiologist at the university. We studies of coronary flow distributions. We (interview by served together on the University Personnel would take dogs with flow meters on the coro- Promotions Committee for three years. He was nary arteries and insert catheters to measure Frederik Nebeker quite a bit older than I and was an intensely curi- pressure, and we obtained much more detailed on 13 December 2000) ous person, a very bright guy. But he had never information on the way in which the coronary studied anything about mathematics or physics. flow distributed, particularly in disease states. He would read articles in the cardiology journals that would have some equations. People were * * * * * just starting to do some modeling and he would bring these to me and say, “Could you explain Mates: One of the practical problems with this to me?” I said, “Well, this is really kind of coronary artery disease is that most patients are intriguing stuff.” It was getting harder and asymptomatic until the artery is just about to- harder to get funding in the space areas, so I tally closed because the circulation compen- thought, “Well, let’s give it a try.” sates. As your artery starts to constrict, the downstream circulation starts to relax to reduce * * * * * the resistance. So you can occlude, in most cases, over 90% of the area of an artery without Mates: I became more intrigued with the reducing flow. All of a sudden though, you get to physiology of the heart, and what we could do that critical point and then flow starts dropping with patients was very limited because you off. could only make measurements of the overall Nebeker: Were you able to model that? behavior. During the time I was in the Depart- Mates: Yes. We were able to do a lot of mod- ment of Medicine, I became good friends with eling of that. another cardiologist who was about my age, Fran Klocke, who was more of an experimental * * * * * cardiologist. His interest was in the fundamen- tals of cardiac contractility and using animal Mates: Very complex models can simulate models, where you could make much more de- anything you want, but the problem is how to se- tailed measurements. So I started working with lect the parameters. My idea was that unless you him, and got into sort of a more fundamental had a simple enough model so that you could level, not only in the muscle mechanics, but at identify the parameters in a particular animal or that point we also got into modeling the behavior in a particular patient, the model wasn’t terribly of coronary arteries, particularly coronary arter- useful from a practical point of view. Although ies that were partly constricted. we never quite made it, I thought that eventually Nebeker: Modeling the flow behavior? we would come to the point where you could Mates: Yes. We actually built a large coro- take measurements in the catheterization lab, nary artery of plastic that we could instrument, model the behavior of the coronary circulation and we had a pulsing pump that would simulate and say, “Okay, this patient has an 80% occlu- the flow as a function of pressure upstream. We sion in this artery,” just from the wave forms. did some very detailed measurements, and that Now we weren’t there, but eventually somebody turned out to be pretty interesting. will be able to do that, for sure.

* * * * * * * * * * Mates: One of the things that complicates the quency response of the circulation under differ- coronary circulation is the fact that the heart is ent conditions, beating and nonbeating. So that’s beating. So every time the heart contracts, the basic mechanics, but also with the electrical pressure goes up. But simultaneously, those ves- technology to make it work. sels downstream get squeezed by the heart mus- cle. They’re going through the heart. So it’s not like a simple hydraulic system, and that makes it “The other thing that has much harder to understand the phasic relation- happened in physiology, which is ships between pressure and flow because the re- sistance is varying as the pressure varies. We even more dramatic, is the shift in were able to uncouple those with this device we emphasis from the sort of had because we could stop the heart from beat- systems physiology, like large ing for a moment, and then still perfuse it with scale stuff, down to cellular and the same pressure and look at the difference in flow patterns. molecular.” So we were able to get some information about the relative importance of the squeezing effect and the pressure. That device, called the * * * * * hydraulic servovalve, was a feedback control system where you would prescribe a waveform, and then it would follow electrically whatever Mates: A lot of physiologists worry that waveform that you prescribed using electrical since there’s so little systems physiology re- engineering principles. We could simulate all search being done, probably because they can’t kinds of waveforms. We actually did a fre- get funding for it any more, who’s going to teach quency response in the coronary circulation by the medical students how the heart works? All applying sinusoidal input pressure of varying the lectures will be about the cellular and molec- frequencies and measuring the amplitude of the ular physiology of the heart, but nobody will ac- flow response, and we could determine the fre- tually ever study the intact organ.

Edward Merrill Merrill: Then in about 1959 or 1960, kind of chemical engineers might be able to do in connec- Professor Emeritus of by accident, two doctors at the Peter Bent tion with medicine, since chemical engineering Chemical Engineering, Brigham Hospital in Boston by the names of Dr. involves, as one of its major thrusts, transport of Massachusetts Institute Roe Wells and Dr. Robert Denton called up the molecules back and forth across membranes be- of Technology Department of Chemical Engineering to find out tween fluids, as well as chemical reactions. (interview by who knew something about rheology, and they It then became of interest to think about what were put in touch with me. They came over and we would do with machines through which blood Michael Geselowitz said, “Can you help us measure the viscosity of would be circulating, and there are two principle on 22 February 2001) lung mucus?” It turned out to be an intractable machines. One would be the so-called artificial job, a terrible job, because it’s so widely variable kidney, or the hemodialyzer, in which blood is and the mucus itself is so complex a mixture of purified by being taken from a patient, run thin and thick liquids. As a result, they turned their interest and my interest to the flow properties of human blood. At first blush you might think that human blood is a straightforward fluid, but it isn’t. We found out it’s complex, and it has what we would call a “yield stress.” That is to say, like yogurt or may- onnaise, as you slow down its stirring, suddenly it stops flowing so that you can get to the point where it won’t flow at all, even though there is a residual pressure.

* * * * * Edward Merrill looks at an experimental biomaterial made with polyethylene glycol and Merrill: In conjunction with the silicone with students (from left to right) Fulton-Berman film on the flow of blood through Cynthia Sung, Jennifer Raeder and Rovena the hamster’s cheek pouch and the transport prop- Sobarzo in the chemical engineering lab at erties going on, it then led me to think about what MIT. through the machine, and circulated back to the level with the head of NIH at the time, Jim Shan- patient. This would be presumably on a chronic non, and the idea was that NIH was trying to get basis for patients who have lost kidney function. MIT to set up a medical school because the idea The other one, back around 1960, would have of bringing engineering into medicine was being been the so-called bubble oxygenator, because at widely promoted, and Jerry Wiesner and Walter that point, blood oxygenation during open-heart Rosenblith were particularly interested in this. surgery was brought about by making a trough I might also point out that by the same token and simply bubbling oxygen through the blood. my story would not be complete if I didn’t men- As we recognized at the time, this doesn’t do the tion my dear friend, John Trump of the Depart- blood any good. It denatures proteins in the blood ment of Electrical Engineering. John Trump was and breaks up red cells. the one who was sitting on what we call the * * * * * Rosenblith Committee and coined the phrase “Engineering and Living Systems,” and that is Merrill: So, what I am talking about is how what we worked on and what the title of it was. [biomedical engineering] started, and so we go back again to the decade from 1960-1970. In * * * * * 1963, I gave the first graduate course titled Merrill: Another part of my story has in- “Chemical Engineering in Medicine” and the volved the use of electron irradiation to carry out MIT course titled 10.56, and I offered it in 1963, polymer chemical reactions and make new and 1965, 1967, and 1969. Then it was taken over by interesting polymers, thanks to the input from one of my doctoral students who is presently a John Trump and from his enthusiastic collabora- professor here, Professor Clark Colton, who did tion with me through his [high-voltage research] his doctoral thesis on the artificial kidney. laboratory. Many of my graduate students Colton’s contribution was to completely eluci- worked in the high-voltage research laboratory date the mass transport aspects of this, so you carrying out their work. Briefly what happens is could design the artificial kidney with the opti- that electron beams can cross link polymers, and mum membrane, the optimum flow rates of the they can also graft new materials onto old mate- dialyzing fluids, and of blood. rials. So you can make, for example, a hydro- * * * * * phobic (water repelling) polyethylene sheet Merrill: I guess it was natural that a major hydrophilic (water loving) by grafting a hydro- component of my research in biomedical engi- philic layer onto a hydrophobic support. neering … somehow involved the issue of what * * * * * kind of polymers can you use to achieve the re- sults you want, especially that of preventing or Merrill: It now turns out that biomedical en- mitigating clotting. A major part of my output gineering is far more entrepreneurial than it used over the years has been addressing that issue, to be. People are moving away from the kind of which is an issue of polymer chemistry and basic research that I was doing in blood viscos- polymer modification. I am credited … [with] ity, the results of which are unpatentable, to pro- the perception of polyethylene oxide or polyeth- ducing fruits of research that are patentable, that ylene glycol, one and the same thing, as being an are products. almost magic biomaterial in that it is remarkably * * * * * inert. In 1974 or 1975, I wrote a paper for the Transactions of the America Society for Artifi- Merrill: The other thing that I observed back cial Internal Organs (TASIO) and the title of it in the days when I was deeply involved with was “Polyethylene Oxide as a Biomaterial.” Boston hospitals, it turned out that in some cases That turned out to be enormously important be- some of the physicians and surgeons treated en- cause now, as the fellow cited, all kinds of peo- gineers as if they were light-bulb changers. In ple are involved in proprietary secret research in other words, there was a kind of pecking order in putting polyethylene oxide and polyethylene which the surgeon or the chief was the chief on glycol onto proteins and onto surfaces and so on. top of everything else, and everybody else re- ported to him. There was a certain, how shall I * * * * * say, arrogance that I think is now gone, because I Merrill: Although on the one hand Dean think the medical profession realizes how much Gordon Brown thought that getting into this engineering has changed what they’re doing. blood research was not in my best interest, on the * * * * * other hand Jerry Wiesner (then president of MIT) and Walter Rosenblith (then provost of MIT) Merrill: If there are two areas where the were very much interested in the application of chemical engineers are going to be highly in- engineering to medicine and surgery. In fact, volved or are already, one is tissue engineering Wiesner called me into a meeting at the highest and the other one is drug delivery, for sure. Nicholas Peppas Peppas: We find ourselves very well pre- added to make the system more stable, as well as Showalter Distinguished pared right now to contribute to a major area of a variety of other chemical compounds that were Professor of Chemical bioengineering called metabolic engineering, needed in order to “solidify” the material. So in and Biomedical because chemical engineers have an exceptional the very early days I became interested in com- Engineering, background in computers and in understanding ing up with alternate ways of preparing materi- School of Chemical cells. The result is that the chemical engineers als without using toxic compounds. That led to Engineering, Purdue have “jumped” into the area of metabolic engi- the very early studies of benign manufacturing neering trying to understand how metabolic pro- of biomaterials. University cesses can be optimized. This is only one of By serendipity I discovered in 1974 that if I (interview by many areas that I could suggest where the chem- take a solution of a particular biomaterial and Frederik Nebeker on ical engineer has contributed because of his or freeze it and thaw it several times, I can create a 12 October 2000) her background—the mathematical background solid structure by a process of solidification that and the physicochemical background requires entanglement of the macromolecular chains and at the same time formation of crys- * * * * * tals. It doesn’t work for any material, but it does work for a rather wide range of biomedical mate- Peppas: We [that is, Peppas as a graduate rials. Some of the early studies were done in student and his advisor Ed Merrill] worked at 1975, the year of the first paper, and the patents MIT in the very early days of biomedical engi- were filed. Over the last 25 years such materials neering. We started developing a series of mate- have been used for a wide range of applications. rials that could get in contact with blood and would not cause blood clotting. Therefore, they * * * * * could be used for artificial organs. These materi- als are known as nonthrombogenic biomaterials. Peppas: I think another major milestone in Nebeker: Surfaces? our development as a group here [Purdue] in bio- Peppas: Surfaces, but these surfaces are medical engineering was in 1978. Steve Ash had modified with the appropriate utilization of just arrived in the town. He arrived from the Uni- anti-clotting agents, which in this particular case versity of Utah where he had studied with one of were heparin. So we were really the originators the fathers of biomedical engineering, Wilhelm of the earliest heparinized materials that were Kolff, in the artificial kidney program. When he used in catheters, in artificial kidneys, and to came here he established a kidney program. some extent a little bit later in artificial hearts, al- What was happening with artificial kidneys in though we were not directly contributing to this. the 1960s and 1970s is that a patient would have to go to a hospital, lie on a bed, be connected to a * * * * * dialysis unit. This unit required all this water in order to take away the urea and uric acid. That Peppas: It’s important to appreciate that the process would take four hours every second day. biomedical field is such an interdisciplinary and So, Steve’s idea was, “Can we come up with cross-disciplinary field that good training is very a portable artificial kidney?” He received money important. I’m delighted that I did postdoctoral from various sources including some compa- work at that time [1974]. But it was only one nies, and for a period of about five years we had a year. It is not unusual to have biomedical engi- tremendous development in that particular area. neers who spend two or three or four years, espe- I was developing the biomaterials that were used cially learning new cellular techniques or in this artificial kidney. getting into genetics or into gene therapy. So At the same time I got much more into the ar- postdoctoral work has become a “must” for the tificial kidney as a biomedical engineer identify- field. In classical electrical engineering or chem- ing flow behavior and transport phenomena. I ical engineering you can have students who be- can tell you the kidney worked very well and it come professors immediately after their Ph.D. was sold to a small company that was created. degree. But in biomedical this additional work is Now they are in the second generation of sys- needed. tems with artificial livers.

* * * * * * * * * *

Peppas: In the very early days of our contri- Peppas: In 1982-1983 we started working butions to biomaterial science, one thing that be- with smart gels and smart materials. There is came very obvious was that the biomaterial nothing “smart” about them except that they itself was not toxic. What created the toxicity have certain functional groups that interact with was a series of unreactive monomers, adjuvants, a surrounding environment in a particular way. stabilizers, and other compounds that were For example, we know that if we have certain It was important to go back to the principles and try to understand the molecular structure that creates a particular property, and whether the material can release a particular protein very fast, or the ability of the material to adhere to a particular surface. We were able to come up with a series of papers in which we defined how mo- lecular design could be used in these new biomaterials and drug delivery systems. Nebeker: This field is called molecular de- sign? Peppas: Molecular design of biomaterials. Nicholas Peppas (right), the Showalter Distin- There are several groups working in this area. guished Professor of Biomedical Engineering at Purdue University, and graduate student * * * * * Aaron Foss discuss the latest oral delivery sys- tem for insulin developed last year at Purdue. Peppas: Recently, we have also embarked upon new areas of biomedical engineering. functional groups such as carboxyl groups, these We are working on biochips and bionanotech- groups in a low-pH environment will not be ion- nology. ized while in a high-pH environment they would Nebeker: Are these chips silicon chips? be ionized. What this means is that the gel in a low-pH environment would stay collapsed Peppas: They could be silicon based. The while in a high-pH environment the gel will ex- ones we have are methacrylate based. In fact, we pand. That can be used to our advantage to create use the same type of micro-lithographic tech- switches—biomedical on-off devices to release niques that could be used for integrated circuits. drugs, to release proteins, to push valves, and to The work is performed in electrical, biomedical, push molecular pistons. We started with these and chemical engineering. This shows how smart materials around 1982, which have be- these areas merge together. come a major part of our work. It is interesting I think that in the next 20 years nanotechnol- that smart materials were developed predomi- ogy and bioengineering will merge. I really nately in Japan. Since these early days we had think we are leading towards more miniaturized significant interaction with Japan, so it was only devices. I tell my students that in 1950 an artifi- natural to be involved in those developments. cial kidney was a big hospital unit, to which one had to be connected for four hours. In 1980 or * * * * * 1985 the kidney was a unit about 30 centimeters long and often portable. Don’t you think there Peppas: There was no way for us to develop will be a progression and that perhaps by 2020 it better drug delivery systems or improved thera- will be replaced by a small box with a few peutic devices without a true molecular design. “chips”?

Nebeker: What other applications were gamma radiation, and the cobalt 60 reactor to Buddy Ratner thought of at the time for hydrogels? graft or covalently attach one polymer to another. Director, University of Ratner: A lot of people were exploring them Allan had the idea of taking the hydrophobic en- for medical implants, the kind they use on the sur- gineering plastics (silicone rubbers and Washington face of the eyes, a contact lens. But there were all polyethylenes and such) and seeing if he could Engineered Biomaterials sorts of medical-implant applications. In fact, graft this hydrogel (the same material I worked on (interview by what I was looking at was a membrane for an arti- in my thesis) to the surface to make a hydrophilic Frederik Nebeker ficial kidney. That was the specific area that I was water-like surface. The basic philosophical idea on 5 December 2000) working. So, I was working in membrane trans- is, what could be more biocompatible than water? port and also on the physical chemistry side, the What would the body like better than water? The fundamental interactions of molecules. body would obviously tolerate water very well. This was an extremely stimulating idea, and it * * * * * used a very interesting polymer technology.

Ratner: I came up here [the University of * * * * * Washington] and started doing the post-doc. One of Allan Hoffman’s specialty areas at MIT was Ratner: This led us into an exploration of radiation grafting. It was using radiation, like how blood clots interact with synthetic materials and how we might design interfaces that look neat medical materials here, there’s no way a liv- better. There were good engineering plastics out ing cell could ever recognize a piece of Teflon or there. There’s polyurethane, there are silicone one of these hydrogels or a piece of titanium or rubbers that are strong elastomers, and people gold that is used in medicine. Evolution had know how to manufacture them. We wondered never provided the recognition mechanisms or if we could just alter the surface structure to receptors for biology. make them more compatible with biology. Nebeker: Wasn’t the idea that you just This led me into what has been one of the ma- needed a benign substance that wouldn’t cause jor focuses of my career, which is how we ana- a problem? lyze surface structure and how we relate surface Ratner: Yes, that was the original idea. We structure to biology. It turned out that although wondered how benign we could make it. It turns there were some things going on, the best work out that the body’s response to these materials on surfaces was being done in those days on was to look at it and say that this is benign. semiconductor surfaces. The microelectronics There’s nothing benign in the body. Conse- community and the petrochemical catalysis quently, it must be foreign, and the body’s re- community were looking at the surfaces of semi- sponse to it was to put a wall around it. conductors and catalysts. These two communi- We started asking whether biomaterials, in- ties had very powerful tools, particularly the stead of being inert, should be recognized by the semiconductor people. body. Maybe we as the engineers should be able to control the healing reaction, the biological re- * * * * * sponse. So we started looking at things like normal Ratner: There were three components. One wound healing and the sorts of molecules, pro- component was making the materials. One was teins, or signaling molecules that are involved learning how we could analyze and characterize there. Then, taking some of our surface knowledge them. This was very different from characteriz- and skills, we ask if we can take existing medical ing semiconductors, so we had a lot to learn devices—they’re basically manufactured in the there. The third area was the biological interac- right shapes, FDA approved, and the surgeons tions of these. All these three things were going know how to handle them—we’ll just take the sur- on pretty much simultaneously here. faces of these things, use our surface skills to put on the right receptors, and instead of making them * * * * * inert, we turn on and control specific biological Ratner: At the same time I had applied to the processes to engineer biological reactions. National Institutes of Health. I had become I had this idea to start another center, namely an known as one of the proponents of using modern engineering research center through the National surface tools for biology and biomaterials. A lot Science Foundation, to see if we could take the field of people were gaining interest in this and I of biomaterials to another level or new paradigm, thought everybody could benefit from it. I made instead of inert, to bioactive and engineered sur- an application to the NIH to start a national cen- faces. We called these engineered biomaterials. The ter to make this type of technology available for center is UWEB, University of Washington Engi- biomaterial studies. This national center was neered Biomaterials. This center is now working in funded and we gave it the name NESAC/BIO, collaboration with some 22 investigators, including which is National ESCA and Surface Analysis material scientists (the roots I came from), funda- Center for Biomedical Problems. mental biologists that don’t really know much about We brought in new types of instrumentation, materials (it was never their area) but they know a such as a machine called SIMS, which stands for Secondary Ion Mask Spectrometry. It allows you to do a mass spectrum of the surface zone. We had a few very powerful pieces of equip- ment. The scanning-tunneling or atomic-force microscope was one. This added another surface tool that the Center uses, and we started building on all these things. On one hand I was looking at surfaces and trying to relate them to biology. The other side of it was looking at biology. We were doing biological experiments. We still had grants from the NIH on making materials and bi- ologically evaluating them for medicine. * * * * * Buddy D. Ratner is using an electron spectros- Ratner: I came to realize that although we’re copy for chemical analysis (ESCA) instrument making better materials and we’re getting very in 1980 at the University of Utah. lot about wound healing and proteins in biology, * * * * * and physicians and people involved in healing. * * * * * Ratner: But what has happened is the real- Ratner: For example, we did a very inter- ization that there’s a real profession here. esting study that was published in Nature last year [1999]. We took protein molecules and There’s a certain way we have to educate our stamped them into a polymer surface and students. There’s a core knowledge base that made pits or imprints in the shape of the pro- gives us commonality. A person that does elec- teins. Within those pits or imprints there were trical engineering, a person that does the me- receptor chemical groups that kind of gave a chanics of a hip joint, the person that designs a “lock-and-key” interaction with the proteins. pacemaker that goes into the body and designs So, instead of trying to sell a medical device the electrodes for the pacemaker, the biomateri- with proteins on it, we might sell a medical als person, the person doing a chemical sensor or device with pits in it, and those pits would at- biosensor—there’s a commonality within what tract or interact with proteins in your own all these people do. They borrow from other en- body to turn on the signals that we want at the gineering disciplines, but there’s a central core surface. that is the biomedical engineer.

Nebeker: Can you tell me about this first ul- turned it 90° so we were going crossways, and John M. Reid trasound system that you built in John Wild’s the screen was suddenly filled with echoes. It basement? What was the design, and what was wonderful. He thought that was a wonderful Professor of equipment you were able to use? discovery, the anisotropy of reflections from Bioengineering Emeritus Reid: The biggest help in the design was the skeletal muscles. He quickly ground it up, and and Research Professor, Radiation Laboratory series of books on radars, we obtained an intermediate picture. Drexel University because the circuits were very close. I had al- (interview by ready gone through the sonar classes and had * * * * * Frederik Nebeker kept my Navy notebooks and knew that was an on 6 October 1999) entirely different kettle of fish. I had some guid- Reid: Once I was back at the linear accelera- ance there. I bought a big power supply, figuring tor lab where I had a bunch of friends working, that I would have to power Lord-knows-what by one of them said, “Why don’t you make images? the time we got it running. I had 300 volts at 1-2 Why don’t you scan the pictures?” I said, “We amps, I think. I built a calibrated attenuator by need some position data transmission system. using dc measurements in the electrical engi- We need sweep resolvers or sine/cosine potenti- neering building, filing the carbon resistors to ometers.” My friend at the lab said, “No you value. The signal generator I already owned be- don’t. If you keep your angles small the sine is cause of the radio repair business, so I brought approximately the angle and the cosine is about that along, also my tube tester. I decided to see one. You can use linear potentiometers and what I could use in surplus. I found a 60 mega- make an image.” I thought about it and went hertz IF strip complete, which looked like a good back and told Wild, “Let’s make images.” We place to start. I would have to run an oscillator at had the signals and the video coming out of the 45 to beat the 15 up to 60, which was sort of machine since I had detected the RF, and he said, backwards for most equipment in those days. “Oh Jack, that is getting so complicated. Things They very seldom went to a higher frequency. I in medicine have to be as simple as a paperclip or was a little unsure about designing an oscillator a hairpin. If it gets more complicated than that, because we did not go into that, but I found the nobody will use it.” I said, “John, it is already instruction books for the Hewlett Packard pretty complicated. I don’t think it will take a lot equipment that we had in the linear accelerator more.” He wanted to know what more it would project. I copied down an oscillator using a tri- take, and that is what hooked him. I said, “We ode, and that’s what I built. need a motor to drive this thing back and forth—a variable speed.” He said, “I have this * * * * * wonderful variable-speed drive, based on overdriven clutches.” It turns out he had been Reid: One thing that impressed John [Wild] looking for a project to use this thing. The idea of was when he brought in a cube of beef and said, hooking up a dc motor to it and running it really “Let’s see if we can get echoes.” He sat it on the intrigued him. We had to make a mechanical transducer and there were a few little echoes at linkage to run some dual potentiometers only for the start, but not much. I looked at it and said, the sine, because the cosine was constant. The “All the muscle bundles are going this way [in- vertical coordinate was just the range sweep and dicating longitudinally with his hands].” I the horizontal coordinate was the fraction of the range sweep picked off by a potentiometer. I had body else was imaging soft tissue, much less try- to build a little 12-volt dc power supply for one ing to differentiate cancer in the image from of the surplus motors we had. noncancer.” It did not seem to be possible. “How were we doing it?” Using sound waves. Sound * * * * * was something you used to talk with people on the telephone. “You guys are some kind of nuts. Nebeker: Was that 1952 Science article What kind of crazy idea is this?” After meeting much noticed? Wild, they became firm in that conviction. One Reid: Yes. We had reprint requests from all of the things that led me to leave was that people over the world, and it is the basic reference that were starting to look at me a little funny, too. people use if they want to go back to the begin- “How can you stand to work for that guy?” ning. Nebeker: At the time there was considerable interest? Reid: Yes, up to a point. Radiologists con- sidered that radiology was X rays. It was not un- til they lost all the business in nuclear medicine that they realized that maybe they should broaden the field. It was just in time for the CAT scanners and MRIs. Nebeker: In the medical profession there was not a great deal of interest immediately? Reid: [Nodding “yes”] Wild was an odd character. He tended to get people’s complete John Reid (left) with Dr. John Wild using the support or complete animosity; there was very view camera that was mounted for specimen little middle ground. People used to say that “no- and scope photography.

Murray Sachs Sachs: I became tired of doing radar kinds of * * * * * Massey Professor and things very early on—in my first semester of graduate school. I asked Moise Goldstein what I Director, Whitaker Sachs: For my doctoral thesis I recorded could do in bioengineering, though it wasn’t from single auditory nerve fibers primary affer- Biomedical Engineering called that at that time. The Communications ent neurons in the auditory nerve and studied a Institute, Johns Hopkins Biophysics Laboratory at MIT had been started phenomenon we called two-tone inhibition. It is University years before by Walter Rosenbluth and was do- very much like center surround inhibition in the (interview by ing what we would now call bioengineering. It retina. I remember my very first experiment. I Frederik Nebeker focused largely on processes of hearing, apply- guess I was doomed—or privileged—to con- on 25 April 2000) ing engineering and communication theory and tinue to find myself in situations where I had to computers—such as existed at that time—to be the first to do things. This was the first time solving biological problems. Moise, who had the LINC computer was ever used in the Pea- worked on the first computer for averaging bio- body Laboratory. logical transients, wrote a paper with Larry I remember my excitement when I recorded Frishkoff and Bob Capranica, both of whom from my first single neuron. I called Nelson were at Bell Labs. Capranica was also a graduate Kiang at 1:00 in the morning and told him, “It student with me at MIT. I think their paper, pub- works, it works, it works.” lished in the Proceedings of the IEEE, was titled Nebeker: For what purpose were you using “What the Frog’s Ear Tells the Frog’s Brain,” the LINC? something like that. It was a very influential pa- Sachs: We used Schmitt triggers. We re- per because it was the first paper to demonstrate corded spike trains and used a threshold device to that some animals have communication sys- trigger the computer to record the times of occur- tems—both receivers and transmitters—that are rences of the spikes, and we then used those to attuned to each other. The frog’s auditory sys- construct interval and poststimulus histograms. It tem is closely attuned to its vocal communica- was the signal processing frontier at the time. Ev- tion system. Bullfrogs from neighboring erything was stored on LINC tapes. Those be- counties have different croaks and different in- came DEC tapes when DEC took over the ner ears that filter the croaks. I went to work with industry. I remember transferring the LINC tapes Moise studying that phenomenon in the frog and to IBM mag tapes and schlepping them across the did that for my master’s thesis. That is how I got river to the TX0 computer in Building 20. Then it into this business. took all night to process the data from one experi- ment—which would take about 20 minutes to- rons? We tried to record from as many as 300 or day. We had to read in all the programs with 400 neurons in the same animal, which is why paper tape and control the computer with toggle the experiments took three or four days. By do- switches. And things would go wrong. It was like ing this we were able to build up a picture of the boot camp, and there was such a spirit of camara- coding of the sound across this whole popula- derie that we were working on the edge of some- tion of 30,000 neurons. We sampled 400 in one thing that was going to really blossom. animal and extrapolated from that—reason- ably, we think. We now know how sound is * * * * * coded in those discharge patterns. Nebeker: Is it not coded in a simple-minded Sachs: Another major influence on my work Fourier way? came from Bill Peake who was my official thesis Sachs: Yes and no. It is clearly coded as a advisor. About three-fourths of the way through spectrum. When we started out there were a lot my thesis, I had done what could have qualified of questions about how that could possibly as a neurophysiological thesis. But Bill said, work. There were limitations on the sys- “This is not an electrical engineering thesis un- tems—largely dynamic range and nonlinear less you do some modeling.” That had a pro- properties—that made it questionable that the found effect on my life. I did models that I have code could work. We discovered that by apply- pursued ever since and still use today. In many ing mathematical analysis of the discharge pat- ways it was Bill’s push to modeling that turned terns, signals could be pulled out even with a me from neurophysiology to bioengineering. highly limited dynamic range. You can hear and * * * * * understand my voice if I whisper, and you can hear and understand my voice even if it hurts you when I shout at 120 decibels. “The question was: How is the sound of the voice—which is very * * * * * complicated—encoded in those Sachs: One thing is that the nonlinear phe- patterns of those 30,000 nomenon seems to disappear. In a hearing-im- neurons? We tried to record from paired person, the ear gets linear. The as many as 300 or 400 neurons nonlinearity that enables the bandwidth to remain fairly narrow is gone. It is as though the ear is try- in the same animal, which is why ing to listen through filters that have a bandwidth the experiments took three or too wide to filter individual frequency compo- four days.” nents. Therefore our current approach is to try to understand the coding in the pathological ear. We Sachs: At that time no one believed biology use cats that are deaf in some ways. could be nonlinear. I went way out on a limb with Nebeker: Is this specifically with the inten- this model because it tried to relate the auditory tion of being able to design better hearing aids? nerve data to nonlinear cochlear mechanics. It was Sachs: Ultimately, yes. The current goal is to the only time in my life that one of my papers was develop a model precise enough to predict the rejected by a journal. I fought back and finally had patterns of discharges in the auditory nerve in the paper published. It is the most important paper pathological ears so we can use those models to I ever published. Experiments and modeling have test hearing-aid designs. been done many times in the subsequent 24 years since 1976, and much more elegantly than I did * * * * * them, but the answer is still the same. Sachs: We do not have enough graduate stu- * * * * * dents because of our limited space resources, but there are 500 students in our undergraduate pro- Sachs: The idea was to try to understand gram. Of the 1,100 or 1,200 students in the engi- how speech is encoded in the auditory nerve. neering school at Hopkins 500 are bioengineers. The auditory nerve consists of 30,000 neurons This is also happening in other places. If an un- carrying electrical signals from the ear to the dergraduate bioengineering program is begun, it brain. Any information that the brain gets about rapidly becomes the largest and best in the uni- the sound has to be carried in those electrical versity. That is because there are a lot of kids signals. The question was: How is the sound of who want to be engineers and want to be techno- the voice—which is very complicated—en- logically involved but also want to do things for coded in those patterns of those 30,000 neu- humanity that involve medicine and biology. Herman Schwan Schwan: [Shortly after coming to the Uni- diovascular session, an ultraviolet one; a more versity of Pennsylvania in 1950] I became a general one on microwave radiation biological A.F. Moore Professor of member of both groups [concerned with medical and health effects, one on ultrasonic radiation; Biomedical Engineering electronics, one in the Institute of Radio Engi- and one on infrared radiation. There were several Emeritus, University of neers (IRE) and the other in the American Insti- general sessions, a cardiovascular session, and Pennsylvania tute of Electrical Engineers (AIEE)]. At that one on various types of nonionizing radiation. (interview by time, I became aware of the Joint Executive You see in the program that we have now a large Frederik Nebeker Committee in Medicine and Biology (JECMB), number of papers. The attendance at that meeting which was a small group of about six people, two was even larger than in Minneapolis. Otto on 6 October 1999) appointed from each of three societies: two from Schmitt had attracted around 400 people. The the medical electronics group committee of the previous conference had something like 50 peo- IRE, two from the AIEE committee on electrical ple. In Philadelphia it was again significantly techniques in medicine and biology, and two more; we had around 550 people attending. from the Instrument Society of America. The Joint Executive Committee was primarily re- * * * * * sponsible for organizing the so-called annual meetings. Nebeker: When did IRE or AIEE start pub- In 1952, here is the program of the fourth lishing in this area? conference on electronic instrumentation and Schwan: The first regular publication was nucleonic medicine, reflecting how the field was from the IRE. It was first Transactions for Medi- conceived at that time. There was a strong inter- cal Electronics, but then they changed the name. est in the atom bomb and all related topics, nu- Particularly, it reflected Otto Schmitt’s and my cleonics and medicine and electronic constant hammering on what was important in instrumentation. At that meeting there was a to- the field, and that slowly but steadily changed tal of just 18 papers. the name of the Transactions. The earliest Transactions came out somewhere between * * * * * 1952 and 1955. There was a special issue in 1955 on the biological and health effects of micro- Schwan: There was a quantum jump. At the waves. ninth conference there was the usual small num- ber of papers. Attendance was declining. That * * * * * meeting had only five papers each day, 15 pa- pers in all. Then came a meeting that Otto Schwan: The Biophysical Society went its Schmitt organized. He was not a member of the own way and became highly successful. It was Joint Executive Committee. I had suggested to not strongly engineering oriented at all. The bio- the Joint Executive Committee that we must engineering community split a bit. The ultra- broaden our perspective and had suggested that sonic group became primarily active outside the Schmitt be asked to organize the next meeting. IRE and IEEE when the American Institute of Otto organized an annual conference [in 1958] Ultrasound in Medicine was formed. That was dedicated to computers in medicine and biology. primarily due to the late W.J. Fry at the Univer- It was attended by about 400 people and was a sity of Illinois at Urbana. huge success. There was a reason for the development. In Nebeker: It is still called “conference on the ultrasonic case there was strong participation electrical techniques.” by medical people. By definition they could not Schwan: Yes, but the heading changed be counted part of the IEEE, because to be a slowly but steadily. We … moved away from the member of IEEE you must have a certificate or a nucleonics image which had been used before degree in engineering, which medical people of without success. The annual meeting still used course do not have. That led to a severe crisis in “electrical techniques” and was still sponsored 1968, which I have discussed in several of my by three societies, but this was now a big meet- publications. In 1968 I served two consecutive ing. I organized the next meeting in Philadel- years as chairman of what was then called the phia, and that was an even larger meeting as you Group of Engineering in Medicine and Biology. may see from this conference attendance record At that time I was striving to find ways to here. It was the 12th annual conference on elec- achieve recognition of medical doctors who had trical techniques in medicine and biology. significantly contributed to this field as full Nebeker: Was part of that jump due to in- members. That, of course, shook up the IEEE cluding computers in medicine? very badly since that would have meant a major Schwan: No, it was not computers alone. The change in its total policies. It was debated in topics at our Philadelphia meeting had nothing to many sessions over a period of two years of the do with computers, but it had to do with radiation, IEEE, but it was to no avail. An associate mem- but not with ionizing radiation. There was a car- bership was established, but the associate mem- amount of money. This has helped us to attract so many excellent people in the field, of course.

* * * * *

Schwan: The negative attitude of the molecu- lar biologists for biomedical engineering has deep roots, in my opinion. In 1955 I worked dur- ing the summer at the famous Cold Spring Harbor laboratories. I moved up electronic equipment to measure the electrical properties of E. coli, to- gether with the old and famous Hugo Fricke. At that meeting I usually had lunch in the cafeteria. I remember my conversations with a famous man Dr. Herman Schwan adjusting a doll, filled whose name was Max Delbrück. with appropriately chosen saline solution to Nebeker: Yes, I know of the physi- mimic a human body. It was exposed to a cali- brated microwave beam in an unechoic cham- cist-turned-biologist. ber. This type of research helped to establish Schwan: Yes. Max Delbrück was curious, and standards of safe microwave exposure during he wanted to know what I was doing there with the 1960s. Fricke, so I told him. Then he challenged me in stating that there were all sorts of people such as bership never came off very well since it was Fricke and Cole, “What’s all that stuff good for, perceived by the medical profession as a sec- what Cole did?” That was at a time when Cole’s ond-rate membership. work had led to the Nobel Prize for [Alan] Hodg- kin and [Andrew Fielding] Huxley. I was amazed * * * * * at his negative comment. So there was already un- der the molecular biologists a preconceived nega- Schwan: I had developed a training program tive attitude about electrical measurements of already at Pennsylvania. We offered a number of biological material and related things. specialty courses in the field. My early students received their degree in electrical engineering, * * * * * not yet in bioengineering. With NIH help I estab- lished in 1961 our graduate bioengineering pro- gram, separate and independent from electrical Schwan: Clearly, the IEEE has played a engineering. As a result of our recommendations, dominant role in the creation of the new disci- an NIH committee, a study section was estab- pline of biomedical engineering. Its annual con- lished, making funds available. A key role was ferences have grown far above anything played by another pioneer, Jack Brown, associate anticipated originally. I regretted the member- director of the National Institute of General Med- ship issue. Another basic regret that I have about ical Sciences, one of the five major institutes the field is that it is split too much. We have which compose the National Institutes of Health. many biomedical engineering organizations He pushed bioengineering. He obtained special now, including the International Federation, the funds allocated for training purposes. The first American Institute of Medical and Biological NIH supported training programs were set up at Engineering, the Biomedical Engineering Soci- Rochester, Pennsylvania, and Johns Hopkins. I ety, and the IEEE Engineering in Medicine and obtained significant NIH support for our training Biology Society. But the cooperation between program for 25 years. It was a very substantial the various societies is not quite clear.

Tirrell: I actually started in three main areas. neers). The idea that we were pursuing was Matthew Tirrell One was an attempted outgrowth of my Ph.D. related to his own work. It was on shear flow thesis, in which I tried to get an early start in bio- damage to blood cells. Professor and Dean, engineering. I had a very significant colleague at Nebeker: That must happen in open-heart College of Engineering, Minnesota named Ken Keller who, at the time surgery when the heart-lung machines pump the University of California, that I went there, was the department chair of blood. Santa Barbara chemical engineering and material science. Tirrell: What are the mechanisms of cell break- (interview by Later on he became president of the University down? We did some things. We had a joint student of Minnesota. He’s a fellow of AIMBE (Ameri- together, Becky Bergman. In fact, she’s very suc- Frederik Nebeker can Institute of Medical and Biological Engi- cessful now. She’s a vice president of Medtronic. on 7 December 2000) have to get stuff from other people and try it out, * * * * * the time scale is slow, the creative cycling is slow. We have done a lot of things on studying, Tirrell: In the course of this, I also became optimizing, and just performing polymerization interested in two other areas of technology that reactions better that have led to an ability to ultimately come back to some implications for make our own materials. biology. One was adhesion. What is it that makes things stick? Why do polymers stick to surfaces and how can they augment and enhance “I see the interface between the adhesion between surfaces? We started to biology and engineering as a lot tackle that problem to develop some methods to broader than medicine. There’s measure and actually predict adhesion a little all this material science that might better. Nebeker: Predicted on the basis of the ac- eventually have something to do tual molecular configurations? with medicine, but there’s Tirrell: That’s right. If you look at some of agriculture and the environment.” the publications early on my list we did some of these things that we called, “Calculations of the Healing Process.” What we were talking about there is kind of a leap from what I was just say- * * * * * ing, but it’s another problem in this spirit. If you take two chunks of plastic and you stick them to- Tirrell: The other was that in 1987 or 1988 gether and supply enough thermal energy, how we were awarded an engineering research center long does it take them to weld just because the from the National Science Foundation at the polymer molecules diffuse across the interface University of Minnesota. It was about a to the point where you can’t tell where the inter- three-million-dollar-a-year grant that had sev- face was anymore? eral programs within it. I was the original direc- tor of the polymer program. That was very * * * * * successful in the sense that we trained a lot of students and we attracted a lot of industrial sup- Tirrell: To keep the connection to our bio port. For the first four years we had a polymer topic, I’ll say that cells have a form of polymer program, a coatings program, a surfactant pro- layer like this on them, too. They’re covered gram, and an inorganic thin films program. with polysaccharide molecules, and the purpose As we went out to industry, we started to at- of the polysaccharide molecule is to create ge- tract, somewhat to our surprise, a huge amount neric repulsion between the cells and a kind of a of interest from companies making biomedical cushion into which the cells want to adhere to devices. It shouldn’t have been much of a sur- form a tissue. Specific receptors stick out and prise in Minneapolis, to tell you the truth. they form ligand receptor interactions. But be- Medtronic, Saint Jude, 3M all have a huge num- fore you can have specific binding, you have to ber of interests in this area. After a couple of block all the nonspecific binding, and to some years, probably 1991 or 1992, within the Center extent, that’s what we were doing with the col- for Interfacial Engineering we decided to start a loidal particles: blocking the nonspecific bind- bio-interfaces program. That was a key moment ing. for me. I just basically volunteered to start it and Nebeker: That’s how nature works. It first to develop the relationships with companies, and blocks the nonspecific and then provides for the I saw that as a way of rekindling my long dor- specific binding. mant interests in this area. Tirrell: Because if anything stuck to any- thing, we’d look a little different. We’d be kind * * * * * of blobs of cells or something. So in the back of my mind, there was always a learning process Tirrell: I started thinking about our surface and a fermentation of some of the things that we forces apparatus, about our block co-polymer were learning and how to apply it to biology. work, and how we could display peptides and or- ganize a layer of peptides on the surface in a * * * * * spontaneous way, and that’s when we developed molecules like this. Now, this is a triad of mole- Tirrell: As a parallel thing this has fed into cules for a particular reason, but what we have my bioengineering interest in one way, and that done and are still doing is making synthetic mol- is that we do new material synthesis in my lab. ecules that we call peptide amphiphiles, where The ability to make your own materials is a tre- we take a small piece of protein and put a syn- mendous asset in the biomaterials area. If you thetic lipid tail on it. If you take a surface that’s hydrophobic, like a piece of polyethylene, and Tirrell: One of the types of cells that we dip it in a solution of these molecules, these mol- have worked with is endothelial cells, cells on ecules spontaneously organize on the surface the inside of blood vessels. Through a collabora- with the lipid tails packing like they would in a tion we are, in fact, looking at whether one cell membrane, and the hydrophilic head groups could, in synthetic materials for small diameter going out. So these are like little block co-poly- vascular grafts or stints that are used to hold mers. These are the stickers and these are the open blood vessels, put our coatings on to help things that stick out, but now they’re there for develop a viable layer of real cells—anchor a the purposes of mediating the interaction that the real endothelium on a synthetic surface. It’s a surfaces have with their surroundings in a differ- challenge. ent way, not just to keep everything off. How- ever, this keeping everything off comes right * * * * * back into this problem, too. Lots of things could stick to that. What you also want to do is mix this with inert molecules that kind of shield the sur- Tirrell: I see the interface between biology face from cells and other things that are sticking and engineering as a lot broader than medicine. for the wrong reason, and you can then really try There’s all this material science that might even- to highlight the specific interaction that you get. tually have something to do with medicine, but Nebeker: So, you’re trying to do something there’s agriculture and the environment. It all in- like nature—having specific ligands there. volves having people with the analytical quanti- Tirrell: Exactly, and spaced out and inter- tative computational skills that one normally spersed in a background of protective, associates with engineers, but with a sophistica- nonadhesive stuff. That’s what we’re doing tion in biology that you don’t normally associate now. with engineers. We’re in the process of trying to design what I’ll call “bio-nonmedical engineer- * * * * * ing” here at UCSB.

Valentinuzzi: Another technological Nebeker: You actually recreated those ex- Max Valentinuzzi concept was measuring maximum and mini- periments! mum systolic mean pressure. Due to friction Valentinuzzi: Yes. Professor of and the blood column (inertia), the response Nebeker: Did they work? Were you able to Bioengineering, time was slow. Actually, it was average pres- do them? Universidad de Tucumán, sure that [Stephen] Hales measured [in the Valentinuzzi: Oh yes, it was beautiful. Argentina early 18th century]. You can read the original (interview by version in his classic book, Haemastatics,a * * * * * nice description, indeed. It is probably one of Frederik Nebeker the first scientific papers ever written in Eng- on 13 October 1999) lish. In those days they still wrote scientific Valentinuzzi: Early in 1962, I received a lit- papers in Latin. He described also two oscil- tle brochure about a symposium that was going lations, a slower one, respiration, and a to be held in Houston on information theory ap- quicker one, which corresponded to the heart plied to neurophysiology. I thought, “Informa- beats. tion theory is what I like. This is for me,” and sent in my name. I didn’t present anything, be- cause I didn’t have anything to present. I had a baby a few months old, and we drove all the way from Atlanta to Houston to attend the three-day symposium. It was there that I met Dr. Hoff and Dr. Geddes. That turned out to be a major turn- ing point in my life. On a Friday night, I went to a cocktail party in the Doctors’ Club at the Texas Medical Center. I told Dr. Hoff that I was an electronics engineer at Emory University in At- lanta and explained my unhappy situation to him. He was always very kind and happy. He said, “Why don’t you postpone your leaving by Max Valentinuzzi in his lab in 1997 controlling another day and come to my lab tomorrow movements and actions during an experiment morning?” and I agreed to do that. He said, with “subject” Julio C. Politti, a graduate stu- “Come at 9:00 tomorrow morning to the lab. dent. I’m going to show you something, and I’ll intro- duce you to Dr. Geddes.” When I went there, There was some activity in bioengineering in there was another person visiting the laboratory Argentina before, but it was a very small seed. at the same time as me, Dr. Mary Brazier, a top Now we have a relatively solid laboratory with neurophysiologist. the first biomedical engineering master’s degree When I first visited the laboratory I had no and Ph.D. programs in Argentina. idea that I would be spending many years there. After that Dr. Hoff said, “When you get back to * * * * * Atlanta, send me your curriculum vita, and I might have something for you.” Five minutes af- Valentinuzzi: In the Electrical Engineering ter getting back to Atlanta I sat down to write out Department—there were a number of details in my curriculum vita and sent it to him. A few days between not worth going into here—almost im- later I received a letter saying, “We have a spot mediately I became the head of the Laboratory for you here. We have a contract with NASA, of Bioengineering. It was a small laboratory, and there are some respiratory physiology sub- with only a couple of students and myself. It was jects on which we would like you to work.” One really nothing. Essentially, we built everything month later, in March or April of 1962, I went to up from zero to what we have now. I consider the Houston. I was within that contract with NASA, birth of our laboratory to be in 1974, so we now and they were developing the first impedance have been in existence for 25 years. pneumograph, which was used in the first subor- Nebeker: Did you change its name? bital flight with John Glenn. Valentinuzzi: It went from Laboratory of Bioengineering to Institute of Bioengineering to * * * * * what it is now, which is the Department of Bio- engineering. This is because of the faculty orga- Valentinuzzi: [Baylor] had a biomedical en- nizations in the university. There is, for instance, gineering program with a Ph.D. in physiology. I the Faculty of Social Sciences, Faculty of Engi- took a major in physiology and a minor in bio- neering, Faculty of Medicine. It is the European physics. The program was based strongly on ex- scheme. Within the Faculty of Engineering we perimental aspects, and we did a lot of laboratory now have the departmental organization, so we exercises. I will never regret that, because physi- say Department of Electrical Engineering. We ology is intrinsically an experimental science. report directly to the Dean of the Faculty. This is a very important point in light of the new tendencies toward replacing experiments with * * * * * computer models. I am not against the computer models and they are a complement, but please do Valentinuzzi: With electrical impedance not forget that in the end we want to understand you can detect the presence of microbes in dif- our body. We want to understand the animal. ferent kinds of materials, especially liquid sam- ples. It started what is now called “impedance * * * * * microbiology,” with its origins in a paper by Stewart back in 1899. For the next 50 to 60 years Valentinuzzi: I have been rather successful probably no more than five or six more papers in Argentina in spite of many, many difficulties. were published on the idea. When I returned to Argentina, the country was in One day, 20 years ago or so, a biochemist and a very dark period. Some of the obstacles were colleague of mine, Dr. Richard (Ricardo) Farías, really huge. called me up and said, “There is a paper in the Nebeker: Did that adversely affect you or Journal of Applied Microbiology. Apparently, people you knew personally? they are doing something with that thing that you Valentinuzzi: Yes, it affected us. We lost mentioned so many times. I do not know what it is many people and many opportunities. For sev- about.” I told him to send it over. It was a short pa- eral years I was torn with the question of whether per, two or three pages in length. I read it and said, to stay or leave. I feel fine in the [United] States. “Easy, we can do this in two or three months.” In the States I feel at home, have many friends, Never say that! Well, it has taken us 20 years to and there are no real problems. The years I spent fully understand what it is, how it works, and how in the United States were very productive. I pro- it should be developed. Now we are at the stage in duced a high concentration of papers during my which we can really do many things. It is full of years in the States. I recognize how much I re- possibilities, especially in the food industry …. ceived there and maintain strong contacts with the U.S. On the other hand, I felt I needed to do * * * * * something for my country. So much needs to be done. Ten days ago we celebrated the 25th anni- Nebeker: Are you developing instrumenta- versary of our laboratory. I review everything tion for this purpose [detecting contamination in for the laboratory, and it has been productive. dairy products]? Valentinuzzi: Yes. Other people have done since we started with the project of the Regional it, but we think that ours is better. We have ob- Council. tained two patents for that. In fact, we did some Nebeker: Do they have a meeting every calibration with one of the big dairy plants in Ar- year? gentina. It’s a beauty, it works perfectly. In less Valentinuzzi: Yes. Right now there are nine than six hours you get the results. Not only that, or ten Latin American countries that are official you can detect which farm’s milk was too con- members of the Regional Council. Those coun- taminated. Probably, their milking techniques tries are Argentina, Brazil, Mexico, Cuba, Chile, are not hygienic enough. Then you can say, Venezuela, Peru, Uruguay, and Colombia. “You have X number of days to correct it. Other- Nebeker: What has the attendance been like wise, we will no longer buy your milk.” It has an at these meetings? important feedback effect. Valentinuzzi: Good. Always they try to have it during a national conference. Its influ- * * * * * ence has been very positive. It has made it possi- ble for us to stimulate the organization of Nebeker: You have been president of the laboratories, programs, and societies in each Latin American Regional Council of Biomedical place, both the national societies and regional Engineering. groups within the IEEE structure. It’s been Valentinuzzi: Yes. I was one of the organiz- highly beneficial. For example, in Peru we ers of the Council. started in 1994. They now have a laboratory of Nebeker: Has there been good cooperation bioengineering, a master’s degree program in between Latin American countries in this field? biomedical engineering, and one of the newest Valentinuzzi: Yes. It has been ten years societies.

Welkowitz: [At the University of Illinois] I things because I had a good friend who was Walter Welkowitz had originally been a teaching assistant in ma- working in the original computer laboratory at Professor Emeritus, chinery laboratories, but that was not a research Illinois. Illinois was then building the ORDVAC activity, and I was interested in doing research. I and the ILLIAC I. I had a friend in that group, so Department of was offered an assistantship with Bill Fry, head I did two projects there. I designed an ultrasonic Biomedical Engineering, of the bioacoustics laboratory. The laboratory transducer using the ORDVAC. At that time Rutgers University was in electrical engineering, but he clearly was there were only about a half a dozen electronic (interview by oriented toward bioacoustics. computers in the world. I did something more in- Frederik Nebeker Nebeker: That was not because of any previ- teresting on the ILLIAC I. On the bioacoustics ous work of yours in that kind of engineering, but laboratory nerve project, when they were de- on 18 October 1999) because it was a research lab. stroying neuron sections, they had people count Welkowitz: The work they were doing up how many cells were destroyed. To do this, looked interesting. I spent the rest of my time at one looks in a microscope and counts destroyed Illinois in that laboratory. cells in a region. I became interested in whether Nebeker: What projects was Bill Fry’s one could do this automatically. At that time group doing? there were some experimental flying-spot scan- Welkowitz: The major ones were ultrasonic ner microscopes, where the image was projected projects. The strongest area was probably in ul- onto a pickup tube and scanned. I did not get in- trasonic effects on nerves, and the group ulti- volved in counting and sizing scanning equip- mately developed equipment for radiating ment, but I was interested in whether one could various brain regions to correct disorders. While program the ILLIAC I computer for use with everyone was somewhat involved in that pro- such equipment. I actually fed in simulated im- gram, I did a Ph.D. thesis on ultrasonic effects on ages from such a scanner, with a goal of being muscle. able to count and size cells for any shape of cell. Nebeker: Like a diathermy treatment? If you had cells of different shapes, the program Welkowitz: We were able to prove that our would tell you that there were five in this size effect was not diathermy but was something range and 23 in that size range. I published this else. work in the Review of Scientific Instruments in 1954. As far as I know it was the first computer * * * * * vision work published.

Nebeker: All this time, after the first year, * * * * * you worked in Bill Fry’s group and on your dis- sertation? Nebeker: Was Bill Fry’s group doing imag- Welkowitz: Yes. I also did some other ing also? Welkowitz: Not at that time. He did set up a electrical engineering professorship. It occurred series of meetings at Illinois where he invited all at the same time that Rutgers started the medical the people working in medical ultrasonics. Peo- school that is now the Robert Wood Johnson ple came in and presented what they were doing, Medical School. At that time it was the Rutgers so we were fairly current in the field. That is Medical School and was part of Rutgers Univer- where I met Jack Reid, since he was doing ultra- sity. I met with the Dean of Engineering and with sonic imaging. He was working with Dr. Wild, a the Dean of the Medical School, and they both physician at the University of Minnesota. All the felt that my interest in biomedical engineering different people who were working in would be desirable for them. (It was being called ultrasonics in medicine met together. Bill Fry that at a number of schools.) I started a program was considered to be an outstanding scientist, in the electrical engineering department in bio- and people came from many places to meet with medical engineering. the group. For example, while I was there, Nebeker: You were the first one in the EE Herman Schwan came because he was inter- department doing that sort of work? ested in mechanisms of electronic and acoustic Welkowitz: Yes. I actually started with a interaction with tissue. The group with master’s program here, which later developed McCullough and Pitts came out from MIT, be- into a Ph.D. program. cause Bill Fry was doing brain irradiation and they were very interested in a more rigorous ap- * * * * * proach to studying the brain. McCullough and Pitts did the original work on neural networks. Welkowitz: When I first came to Rutgers Nebeker: I remember Schwan telling me that there were other such programs in the country, he was looking more at electromagnetic fields and there were programs that clearly preceded and impedance of tissue, and that his main idea ours. Originally, there was a question of whether was to try and understand the physics of these bi- this was really an academic field, and how one ological materials. It sounds like at least your could describe it. It was a strange field. Com- thesis was related to this. panies were not very receptive to hiring such Welkowitz: Yes, but it was based more on graduates. I remember talking to the Hewlett- the mechanical properties of the tissue, rather Packard people, because at that time they went than the electromagnetic properties. into and are still in monitoring, and they did a lot of patient monitoring. They told me that they * * * * * would not hire somebody with a degree in bio- medical engineering—they concluded that such Nebeker: What did you work on initially [at a person could not do electronic or mechanical Gulton Industries in the mid 1950s]? design. They were very skeptical that this was Welkowitz: I started by designing some really an academic field. In fact, they tended to medical instruments. I first worked on some discourage schools from starting it as a separate catheter instruments. field. They said that they would rather hire an Nebeker: Intracardiac catheter microphone? electrical engineer and have that person meet Welkowitz: Yes, I designed some of those. with physicians to pick up medicine and biology One could insert them in a typical cardiac cathe- while working. It took a long time before people ter, and they were good for locating holes in the recognized biomedical engineering as an engi- heart wall. The sound was localized, and if you neering discipline. moved the microphone around and watched it under a fluoroscope you could see where the * * * * * sounds were greatest. That led us into designing intracardiac pressure gauges. There was always Welkowitz: It turned out that some of the more a big interest in intracardiac pressure gauges. accepted concepts in blood flow physiology were People had made some very complex ones using probably not correct—physiologists rarely ana- electrodynamic approaches; but that required lyzed the system as an engineering system. There winding hundreds of turns of tiny wires. They have been other engineers who analyzed it that way were very difficult to construct. We used and came to the same conclusions that we did. For piezoresistive silicon to make a unit. It was one example, the aorta tapers both geometrically and, of the very early ones using that approach. Peo- interestingly, in its elasticity. The amount of elastin ple are still using similar designs. This was the in the walls varies. It turns out most people said that first one built by a company, and naturally the many of the waveforms observed in the aorta are company patented it. just due to reflections at the end of the aorta. We showed that you could get the same waveforms * * * * * with no reflections. If you analyze the taper care- fully, you find that tapered systems can give you the Welkowitz: At Rutgers I was offered an same waveforms without reflecting. Why is the sys- tem tapered, and why isn’t it a nice uniform tube? It * * * * * isn’t. I suspect the reason it tapers this way is to match varying terminal impedances, so you get the Nebeker: Since you have written about appropriate pulsatile flow most easily. medical instrumentation generally, could This is the sort of information you get only if you comment on what developments over the you do an engineering analysis. If you were a last decades have really had major conse- traditional physiologist and you made wave- quences? form measurements (which they do, and make Welkowitz: There is no question that far some very good ones) you might look at them above everything else was the development of and say, “Why does the input look one way and imaging systems. Thirty years ago the only way the output look different?” You might say, of looking inside the body was with an X-ray “Somebody told me that there is such a thing as system. That produced a crude picture because wave reflection, so maybe that is what is caus- it was a summation of all the absorptions ing the change.” That is a typical traditional through the body. Bone is more absorbing than physiological approach to analyzing the prob- other tissue, so one mostly saw bone. Many tis- lem, after doing very good pressure measure- sues are not very absorbing and you do not see ments. It is not an engineering analysis, and it them at all. will come up with an explanation that may or may not make sense. * * * * * We engineers try to approach the problem differently. Suppose we say that their ap- proach is not what occurs. What would happen Welkowitz: I think medical people now if this were a nicely matched system. Would fully accept modern imaging to provide them this difference in waveforms occur? What I necessary information. Now they say, “Maybe show in my book is that the input and output engineering has something to offer.” waveforms match up very well with this taper- Nebeker: Do you think if there had not been ing concept, rather than with a concept that this visual output, if we were talking about im- does not match the anatomy. A reflection ap- pedance measurements or something like that, it proach might match well if these were uniform would be a harder sell? systems. One realizes that there are many sys- Welkowitz: A much harder sell. With imag- tems in the body that are well designed, but ing they get an internal picture that they recog- much thought is not often given as to why they nize as something they have seen in surgery or in are designed the way they are. medical texts.