Recent Advances in Biomolecular and Biomedical Imaging
This is a modal window.
The media could not be loaded, either because the server or network failed or because the format is not supported.
Formal Metadata
| Title |
| |
| Title of Series | ||
| Number of Parts | 340 | |
| Author | ||
| License | CC Attribution - NonCommercial - NoDerivatives 4.0 International: You are free to use, copy, distribute and transmit the work or content in unchanged form for any legal and non-commercial purpose as long as the work is attributed to the author in the manner specified by the author or licensor. | |
| Identifiers | 10.5446/45079 (DOI) | |
| Publisher | ||
| Release Date | ||
| Language |
Content Metadata
| Subject Area | ||
| Genre | ||
| Abstract |
|
Lindau Nobel Laureate Meetings292 / 340
3
5
15
16
18
22
23
27
32
33
34
35
47
54
58
67
69
70
72
73
78
79
83
84
85
86
87
90
92
95
100
102
103
104
105
106
114
115
116
118
119
120
122
126
128
129
130
131
133
137
142
143
145
147
148
149
153
155
156
159
162
163
165
168
169
174
176
178
181
182
189
194
198
201
202
203
206
209
213
214
217
218
219
220
225
227
228
237
240
241
244
245
250
254
257
260
261
266
273
278
284
285
287
291
293
297
302
308
310
317
318
319
321
325
327
328
333
00:00
Particle physicsRoman calendarHot workingBuick CenturyProgressive lensLappingLecture/Conference
00:43
Buick CenturySynthesizerOpticsVermittlungseinrichtungChandrasekhar limitSpeckle imagingSensorWatchProgressive lensOpticsSynthesizerSingle (music)SensorSoundYearSpeckle imagingContinuous trackLecture/Conference
01:17
HailCrystallizationBrightnessQuality (business)YearMedical ultrasonographyFocus (optics)PhotonAngeregter ZustandTransfer functionPhononVisible spectrumCrystallizationIonLightTeilchenrapiditätQuality (business)BrightnessForcePitch (music)Tin canDrehmasseDotierungEnergy levelSizingMatrix (printing)RRS DiscoveryLecture/ConferenceComputer animation
03:23
Magnetic coreAtomhülleClothing sizesDotierungCluster (physics)CogenerationMicroscopeKardierenCrystallizationOpticsSynthesizerMagnetic coreController (control theory)SemiconductorRutschungSingle (music)SizingGentlemanAtomhülleAngeregter ZustandSEEDDotierung
04:33
Superheterodyne receiverScanning electron microscopeMagnetic coreAtomhülleBrightnessRutschungCluster (physics)Electron microscopeMicroscopeBrightnessNoise (electronics)Signal (electrical engineering)Single (music)OpticsConstellationTypesettingVertical integrationSizingKopfstützeSeason
06:19
Halo (optical phenomenon)Field-effect transistorDiffusionHalo (optical phenomenon)Kette <Zugmittel>Acoustic membraneCosmic microwave background radiationDyeingHourWeekBird vocalizationContinuous trackChannelingTelecommunications linkKühlkörperHailPhotographyApochromatCab (locomotive)
08:10
QuantumPeriodic acid-Schiff stainBill of materialsDiffusionAcoustic membraneDiffusionTime clockSingle (music)Die proof (philately)Continuous trackGentlemanComputer animation
08:46
QuantumBill of materialsPeriodic acid-Schiff stainLawrence, Bradley & PardeeQuantumRoots-type superchargerQuantum dotStock (firearms)Transfer functionField-effect transistorAxionComputer animationLecture/Conference
09:27
Speckle imagingMikrofluidikTransfer functionTypesettingFire apparatusField-effect transistorEngineChannelingAxionBird vocalization
10:39
Electric power distributionFoot (unit)Thrust reversalSpare partContinuous trackTypesettingFACTS (newspaper)Volumetric flow rateMachineTropical cycloneDC motorMetreRail profile
12:12
Single (music)Fiat 500 (2007)Airbus A300Engine displacementClothing sizesFrame rateCartridge (firearms)SpantSignal (electrical engineering)Continuous trackQuantumSizingElectric power distributionTemperatureDC motorHot workingEngineAlcohol proofUniverseQuantum dotGroup delay and phase delay
13:37
Data conversionBrightnessAnalytical mechanicsQuantumController (control theory)Hot workingCell (biology)Intensity (physics)TARGET2BrightnessQuantum dotAutomated teller machineFinger protocolComputer animation
14:59
SensorAußerirdische IntelligenzCut (gems)Intensity (physics)DayBallpoint penQuantum
16:32
Local Interconnect NetworkSpeckle imagingSeasonPsyche (psychology)Group delay and phase delayMedical ultrasonographyAdaptive opticsLappingLecture/Conference
17:07
Speckle imagingSpeckle patternHot workingIonRadiationTissue paperShort circuitSound
17:36
Speckle imagingSensorLongitudinal waveDiffractionSpeckle patternSpace probeNoise reductionAudio frequencySuperheterodyne receiverEGPRSWatercraftAlcohol proofCash registerAngle of attackCompound engineEngine knockingReflexionskoeffizientApertureContrast (vision)Interference (wave propagation)Speed of soundSpeckle imagingAmplitudeRoman calendarCartridge (firearms)Angle of attackSensorAudio frequencyShort circuitHull (watercraft)BauxitbergbauWavelengthSoundOrder and disorder (physics)EngineTissue paperLightDiffractionLimiterScatteringArray data structureSpeckle patternPhase (matter)MultiplizitätFood storageEcho <Ballonsatellit>Signal (electrical engineering)Blast furnaceOpticsCylinder headForceComputer animation
21:25
Space probeSuperheterodyne receiverRefractive indexBrechungSpeckle imagingDistortionPressureSurfingEngine displacementSchubvektorsteuerungRefractive indexDistortionRotationSpeed of soundSignal (electrical engineering)DensitySingle (music)SoundSpace probeGlassSubwooferScoutingVideoMaterialWeekDayTissue paperShip classComputer animation
22:41
Superheterodyne receiverSensorWatercraftForgingMetreSpeckle patternSensorBlood vessel
23:10
Standard cellVideoAngle of attackPower (physics)
23:46
Hull (watercraft)Space probeDistortionRotationArray data structureSpeckle imagingRail transport operationsContinuous trackDistortionPhase (matter)TurningWeekProzessleittechnikComputer animation
24:45
Space probeDistortionRotationArray data structureLinear motorSpeckle imagingScanning probe microscopyWriting implementHot workingWeekTypesettingAudio frequencyLecture/Conference
25:13
Weather frontGround stationArray data structureDigital signal processingRapid transitSuperheterodyne receiverAudio frequencyOpticsLinear motorCarl August LinerMeasurementSensorMembrane potentialMagnetic resonance imagingMedical ultrasonographyModulationPulse-width modulationFahrgeschwindigkeitVolumetric flow rateSpeed of soundFrequency mixerSpare partYearClub (weapon)MetreMotion captureOpticsContrast (vision)Audio frequencySoundVolumetric flow rateSimplified Chinese charactersMechanicPhase (matter)KontraktionBubble chamberSensorFilter (optics)Nonlinear opticsVanHot workingDensityWater vaporSemi-trailer truckSpeckle patternMagnetic resonance imagingPhased arrayMedical ultrasonographyComputer animation
28:13
Speckle imagingSynthesizerOpticsSensorLocal Interconnect NetworkContinuous trackSpace probePhotographySensorNyquist stability criterionLastLecture/Conference
28:37
Cell (biology)ClimateArtillery batteryFood storageExtraction of petroleumCell (biology)Climate changeWeatherSingle (music)Differential (mechanical device)YearGenerationContinuous trackKopfstützeIonExtraction of petroleumArtillery batteryHourDay
29:42
WeatherClimateAugustus, Count Palatine of SulzbachSeptember (1987 film)TemperatureSummer (George Winston album)Model buildingRainSpare partWater vaporWavelengthDroughtIonClimate changeStress (mechanics)Climate modelTemperatureCardinal directionClimateDroughtPaperBottleSummer (George Winston album)YearRainWater vaporMonthWind waveOrbital periodFaltenbildungHeatComputer animation
31:06
Water vaporSpare partIonClimateWavelengthDroughtLocal Interconnect NetworkSchwimmbadreaktorHull (watercraft)NanotechnologyCommercial vehicleMaterialVideoLecture/ConferenceComputer animation
Transcript: English(auto-generated)
00:14
All right, today I'm gonna tell you about some work, new work, none of it has been published yet,
00:22
and so don't run out in the lab and do it until we get it out there. But it's, I'm reminded of this work from something the greatest American philosopher of the 20th century said, so played baseball for the New York Yankees,
00:42
and he said, you can see a lot, you can see or observe a lot by watching. And so we know that fantastic progress has been made in optical microscopy, super resolution methods, green fluorescent protein methods,
01:00
all these things, single molecule FRET, and yet there's still more to come. So in the outline of the talk, I'm gonna talk about the synthesis and functionalization of rare earth upconverting nanoparticles. I'm gonna talk about some preliminary experiments we've done in live cell tracking,
01:21
and then I'm gonna turn to something totally different, ultrasound imaging. So let me begin by reminding you of what has been happening in the last eight years in rare earth nanoparticles, and I'm gonna focus on these particles, let's see if this works, sodium, metronium, fluoride,
01:40
and the idea of these particles is if a photon comes in and excites a rare earth ion, there's very rapid energy transfer from ion to ion in this crystal matrix, and finally the energy is transferred to erbium. And so this is the YB over here, and it's transferred over to erbium.
02:03
Of course it can transfer back, but the idea is if there's more YB than erbium, another photon can come in, it can also transfer energy, and there's a second step excitation. And when that happens you have one excitation up to here, this level, eh, it's too fancy,
02:23
one excitation up to here, and then another one up to here, there's a very rapid phonon de-excitation, and light comes out in the green and the red. So in this sense you're exciting with nine 75 nanometer infrared light, and you come out with visible radiation.
02:42
Now you look at this, and this is about eight years old, and you ask yourself the question, why stop at 20% YB? Why not 90% YB? And the answer turns out that if you try to increase the crystal doping beyond 20%, the crystal quality falls apart, there's fluorescence quenching,
03:01
and that limits the brightness. So we were discovering this, working with material science people who are supplying with particles, and then beginning to realize they're not quite as good as advertised, and so in desperation we decided to try synthesizing our own.
03:23
But with a little change, what we decided to do was start with a core of sodium yttrium fluoride, just five nanometers, six nanometer core, where you can grow a very very perfect seed crystal. With that seed crystal we then grow on top of it
03:40
a layer of whatever we wanted, and then we found that we can actually dope it with anything we wanted, and any size we wanted with complete control, because it was an epitaxial growth, just like the semiconductor industry learned that you can start with gallium arsenide, and you can grow on top of that gallium aluminum arsenide,
04:01
gallium indium aluminum arsenide, and so on. So with this trick we start with a single little blue core and grow this yellow stuff what we want, and then we cap it off with an inert shell to keep the excitation confined. And so these three people grew the crystals, did the optical characterization,
04:21
and the biology experiments. The other thing we found was that the nanoparticles tended to cluster together, at least ones that were supplied to us. So we developed methods of making sure they don't stick together, and we would take a slide of these particles shown on the left,
04:40
and in an optical microscope image them. But then we take the same slide and we took it to an electron microscope and said okay in the electron microscope are they really single particles? And so you see on the right hand side the electron microscope, but the beauty of it's the same constellation image as the optical image,
05:02
but the beauty of an electron microscope is you can turn up the magnification. So for example particle eight sees, aha we see a single particle of about 28 nanometers, plus or minus a few percent, but look at this brighter particle, particle number four, that turns out to be a cluster of three particles.
05:21
And so we do this not for a few particles but for hundreds of particles to make sure that most of them are truly singles, and the ones that are not singles we say okay those are clusters. Now the good news is with this epitaxial growth we can go to 92%, 98% YB, and the rest erbium.
05:42
We can make the adjustment any way we want, and we find that at 625 watts per square centimeter, which there had been no published results being able to see a single nano particle of this type, of this size, below a kilowatt per square centimeter,
06:02
but at 625 watts per square centimeter we're seven times brighter, and at eight watts per square centimeter where we actually see a signal to noise of five to one in 100 millisecond integration time, we're about 150 times brighter. Okay so that was very encouraging,
06:21
so we decided to try some live cell tracking, and so here's some preliminary results of this. We started with a membrane protein called channelrhodopsin. This is well known recently especially at Stanford because it's the protein you use to shine blue light in,
06:41
and you can actually trigger neurons to fire in a method called optogenetics. And so being at Stanford you can go down the hole and say can I get some channelrhodopsin linked to a fluorescent protein. We wanted linked to a fluorescent protein
07:01
because to that we added a protein sequence called a halo tag protein over here. The reason we do that is we want to attach our nanoprobes to this linear molecule which will insert itself in this halo tag protein and form a covalent link. So when we see both our nanoparticle
07:21
and the fluorescent nanoparticle we say aha, we've got specific targeting of this membrane protein. And when we see only the nanoparticle we say it must be sticking to some random protein or surface on a cell. So with that we can compare these channelrhodopsin proteins
07:43
diffusing around with very stable organic dyes on the left-hand side, and you see this background of fluorescence. And on the right-hand side you see with the up-converting nanoparticles attracts, instead of photobleaching for one or two seconds
08:02
they don't photobleach for hours. They're completely stable, they don't blink. And so you zoom in on a, oops, you zoom in on a single nanoparticle and you see, hold on just a second,
08:21
this very erratic behavior where it's not pure diffusion of this membrane protein, the channelrhodopsin on the cell surface, but it's punctuated by staying at certain places, zipping to other places, going back. And so you can take time records of this. And this is our first proof that indeed
08:42
we can do single cell tracking. All right, so next we said, let's return back to some experiments I did in collaboration with Manxiao Shui, who's now a tenured professor at Stanford, but at this time she was with me at Lawrence Berkeley Lab and UC Berkeley as a postdoc.
09:03
And we also collaborated with Bill Mobley at Stanford. And we're looking at nerve growth factor with quantum dots, and this is in a dorsal ganglion root neuron, and we see that these particles are tracking back and forth,
09:20
but you see from this that they're blinking, they're going back and forth, and it was not pretty. So we decided we're gonna try the same thing, we're gonna use our nanoparticles, but instead we will put our new rare earth upconverting nanoparticles in the cargo space, and we would see both retrograde transfer,
09:42
that's transfer from the distal axon, if this is the axon, this is the cell body, cargo gets transferred backward, this has been implicated in lots of important things, which include neurodegenerative diseases like Alzheimer's. This retrograde transport is very important to study,
10:03
and that is done by a motor called Dynene on the left hand side, and Kinesin actually transports cargo from the cell body to the distal axon. And so we used the same type of apparatus used when Ben Shao was in my lab,
10:23
she put on axons, sprinkled a little nerve growth factor on this side, and the neurons would actually grow through these little channels over to this side, and in these regions over here, you can have a stream of axons that you can visualize. All right, so this is now with the new particles,
10:45
we can now simultaneously visualize them, the difference is these particles don't blink, and they're decidedly brighter, and so we can actually tell whether they reverse themselves, whether they change tracks,
11:01
things of that nature, and so on the right hand side you see in purple and cyan two tracks as they go over these several microns, we can in fact trace them over a millimeter. We can look at one cargo and it seems to have eight reversals in 25 microns, that's the blue track,
11:23
you look at another cargo on another track with different set of motors, and you find that it only has two reversals, and surprisingly we find that these particles actually, some of them have no reversals, some of them have just a few reversals,
11:41
and some have many reversals. We think the hypothesis is that there are numbers of feet on these cargos, they could be two, four, six, and so on, and so we will now be putting fluorescent proteins on them, we watch how they move, and then you turn on light that actually illuminates the fluorescent proteins,
12:03
and by sequential flow de-bleaching we can find out the number of feet. So this is in part a way of getting into more of the internal machinery. We can also zoom in a little bit, and now we're taking movies at frame rates of two and a half milliseconds per frame,
12:22
and these are the signals, this is the raw signal, and it doesn't look very exciting to you, but if you look at this over many seconds, eight seconds in this case, and you zoom in on this, you actually see single molecular steps of the motors,
12:40
the difference is now they're in a live neuron, at 22 degrees centigrade. And we can look at this indefinitely. And so this is very exciting because if you compare it to an excellent study of in vitro assay done by Ron Vale and his group,
13:02
he finds that if he has a pure dining motor, their step size is eight nanometers, but sometimes you see a little bit of 16, a little bit of 24, but that's in an in vitro study where they had to lower the temperature using quantum dots, they can't track forever because they blink,
13:21
so they use track, which track they're on, and so we have the advantage, we can actually operate at room temperature even higher, we see the same distribution of these dining motors, and again, it's a whole cell, a live cell. All right, so what are some of the other applications? This is not our work, but work done in Fudan University
13:44
where they took rare earth quantum dots, and they targeted these rare earth quantum dots to heal cell tumors injected just underneath the skin of a bald mouse, and they have a control tumor on the right hand side,
14:01
the targeted tumor on the left hand hind side of the mouse, they inject the nanoparticles into the vein tail of the mouse, so it goes through the whole circulatory system, it gets picked up in the liver and the spleen, but finally they find their way to the targeted tumor, they don't hit the control tumor,
14:21
and the other very good news is they clear the spleen and liver, which means that FDA may allow this to happen, and they're non-toxic, and so the good news about our particles at this intensity of 80 to 100 milliwatts
14:41
per square centimeter, we're about 3,000 times brighter. That means we have 3,000 times higher sensitivity of looking at where we target these cells. That has application not only in mouse experiments, but you consider surgery,
15:00
where a surgeon is looking and trying to excise the tumor, and what they do is they kind of cut and they poke their finger, and they say does that look or feel like a tumor? No, it still feels like a little tumor, they cut a little more, they poke, they look, they poke, finally they think I got it all out, I think, they take a slice, they give it to pathology,
15:21
and a day or two later, pathology tells them some bad news. In 30 to 50% of the time, the margins are not clear. That means the surgeon has to call the patient and say you might want to come back for another surgery, because we just sliced into a tumor.
15:42
So the ability to actually sprinkle, if you know the tumor, and you can find antibodies, you can sprinkle these particles during surgery with very low intensity infrared light, the same intensities that the Chinese use, they put rareth quantum particles in their money,
16:01
so that they, it's an anti-counterfeiting scheme. The same thing can be done, these particles light up in green and red, which are adjustable, so the surgeon without any cameras can just look with their eyes, but now with 3,000 times higher sensitivity to find margins. Then they rinse it out, cut a little more,
16:21
and so we're gonna be working with Evan Rosenthal, who is actually a real surgeon who deals with, you know, does surgery with patients, and see if we can use these nanoparticles. All right, so in the remaining time, I'm gonna talk about ultrasound imaging.
16:40
This is done with a postdoc, Yilei Li in my group, and also an undergraduate who did his honors thesis with me on using artificial intelligence to look at adaptive optics, but he is spending some time between graduating as an undergraduate
17:00
and going to graduate school, and so he's spending some time in my lab. To remind you, ultrasound imaging is something that looks like this. If you didn't know better, you wouldn't, we were wondering what this blob is, but this blob is a three-centimeter long little baby in mommy's tummy,
17:22
and so it's good they use ultrasound because it's non-ionizing radiation, and so it's very safe. Now you can ask yourself, what is all this other stuff? All this speckly things, and so first let me tell you how ultrasound works. It's a very short pulse of sound that goes propagating into the tissue,
17:42
and as it propagates into the tissue with this little red dot of a little ping of sound, it creates echoes, and then you actually listen for these echoes, and as the pulse propagates further down into the tissue, it continuously generates this echo signal,
18:01
and so as it continues and continues, you get different reflected pulses that is a continuous reflection. It's not just ping, ping, ping. As the pulse goes on, there's a continuous reflection back. You know where you are in depth because you know the speed of sound in tissue,
18:20
and then the transverse resolution is just given by diffraction expression delta x is equal to lambda, the wavelength of the sound, divided by the numerical aperture of the sound optics, and then it's a phased array so you can scan around, and that's what you get. So where's the speckle come from?
18:40
If there's a sound wave, and this blue oval is containing it, this is an imaging voxel, and in this blue oval you have scattering centers, and so you have sound scattering from one particle and another particle within the imaging voxel. If they happen to be in phase, the amplitude squared that comes back
19:01
is four times the amplitude squared. If they happen to be out of phase, as in this cartoon, what you see is you see no signal, and because these particles within the imaging voxel are randomly distributed, you have this contrast in light and dark. So how do you get around that?
19:21
Well, if you come in with a different frequency, say an octave away, so the round trip time is different, what you can get constructive interference down below, whereas with a different frequency, you get destructive interference. So by taking images at different frequencies, you can get rid of some of the speckle.
19:41
You can also come at different angles, so for this angle you have destructive interference, but for this angle you have constructive interference, and again, by forming images at different angles, you can do this. Now, none of this is new, and what we discovered is what were the limits of what you were doing, number one,
20:03
and if you improve one and use angle for another technique and frequency for another, how does it improve the image? And it turns out it improves it multiplicative, by multiplying, if you get three times in frequency and three times better in angle,
20:21
you get nine times better. So we thought at first, well, maybe you need real precision in order to image the same spot, same region at different angles, and so I was telling this to a former technician of mine who left me and sent me my first graduate student at Stanford.
20:40
He'd rather found Thorlabs and become a billionaire, I just don't understand this. And so anyway, so I told him about what we were doing, and he got excited and said, well, you know, maybe you need a little robot. So he bought us a little robot, and assigned an engineer to program the robot for us,
21:01
and so we took these images, this is an image of this wrist portion by the thumb, and you get it different angles, and what we showed that with no frequency compounding, just the raw image, you get something on the left hand side if you put it in different frequencies, you get something on the upper right hand side, and finally, if you get frequency and angle,
21:22
you get still a better picture, but there's more. Reason there's more is because there is actually distortion due to the ultrasound probe itself as you go like this, and so you're distorting the tissue, the patient could be breathing or have pulsing arteries,
21:41
and most important, as the sound goes through the tissue, the signal is actually generated by differences in the speed of sound and the density of the material, but that's like a change in index refraction, and because you have a change in index refraction, there's weak lensing, it's like looking through slightly wavy glass.
22:03
So we said, well, if we're coming at different angles, and we just want a common denominator, is it possible to do not only translation and rotation of these sub-sections of the image, but do elastic distortion to allow you to make corrections, and it turns out that all of the computer methods
22:23
designed for neural net, convolutional neural net, map beautifully onto this problem, as did Nvidia graphic chips, and so you can actually make real-time corrections with these Nvidia, a single Nvidia graphic chip, and when you do that, you get this image.
22:43
So that dark thing over here is a blood vessel, these are tendons, this is the edge of a bone, the resolution is now 90 microns, and these are all real things. So all of a sudden, the speckle's largely gone,
23:00
and you're getting sub-millimeter, sub-10th millimeter spatial resolution at 15 megahertz. Just to drive home this difference, this is what you get in the upper left-hand side with conventional ultrasound, and this is what you get in the lower right-hand side. You might say, well,
23:21
that takes a lot of computation power, yes. It takes, right now, a few seconds of an Nvidia graphic chip to do all the convolution back calculation, back preparation convolution five or six times to descend on the common picture. And with improvements in the graphic chips,
23:40
we think it can be done in less than one second. So it's really essentially real-time. But there's more. If you want to look at your liver or an abdomen, you have a region of interest over here, how are you gonna get a different angle? This is easy because it was cylindrically symmetric, so you can do this.
24:01
If you try to do this on the abdomen, you go over to the side and push real hard, you've distorted everything, and the patient doesn't like it. Rather, we can now tell the imaging algorithm to say, let's do some real-time tracking. How long does it take? It takes about a tenth of a millisecond
24:22
to actually say, where's the image where I used to be looking at? Use my phase array, point there instead of, so you click on the right-hand side, this is the organ of interest. Then in real-time, the operator can just slide around, and it automatically clicks and gets images.
24:40
And so you can imagine this is your first image, your second image, your third image, your fourth image, and so on. So this, we were in the process of implementing. If I gave my talk two weeks later, you could have seen those images. But we're quite sure they're gonna work because all the other pre-steps. Finally, let me talk a little bit about non-linear ultrasound.
25:03
This is a new type of ultrasound where you have two frequencies at five megahertz and six megahertz, and you look at the difference frequency at one megahertz. So imagine on the right-hand side, you have a phase array sending a single green pulse going down,
25:21
and as the pulse goes down, on the left-hand side, you have another phase array sending a continuous pulse, but the continuous pulse intersects continuously here, here, here, and so on. And so what you have is in an A scan, you have one pulse going down, and the other pulse just streaming along and capturing it.
25:42
And so now you look at the difference frequency. So as physicists, we need quick samples. So what we do is we go to a Chinese grocery market and buy a pig kidney, also chicken hearts, also some other nasty organs that are yummy to eat for some people.
26:02
And anyway, this is the venous part of the kidney. This is the filter part. This is the optical image. This is what you see in non-linear optics, difference frequency mixing. And in linear optics, there is no contrast because it turns out that the density and speed of sound is the same. If you go buy a piece of salmon from Safeway
26:23
with the little things of fat, you see the strands of fat. In non-linear, you don't see. We did this work before we figured out how to get rid of speckle, but all the tricks I showed you before worked with that. So we said, well, this is interesting. We have very, very different contrast mechanisms when you're doing non-linear ultrasound.
26:43
Is there anything else interesting? And so we did a little grave robbing. We went to a colleague with a mouse, and we looked at the brain of the mouse, and in non-linear, you see some light spots standing out. In linear, you see a bunch of speckle, and you do pathology, and you find out those spots
27:03
are geoblastoma tumor. So all of a sudden, you have contrast of tumors that you couldn't have before. Now, ultrasound's pretty good because they can see tendons, cartilage, ligaments, so the normal ultrasound.
27:21
So we also can do Doppler imaging, and with the millimeter or half millimeter resolution, we believe we can see blood flow clots in arteries if you time it right, impaired kidney function, and something we're now beginning to try in collaboration with a neurobiologist
27:41
is to take a mouse brain with optogenetics, and you optogenetically stimulate, let's say, the hippocampus, other parts of the mouse brain, and you, by looking at changes in blood flow, can you have half millimeter spatial resolution with 10 millisecond time resolution, which you can never have with functional MRI.
28:04
So that's another application that we're getting very excited about. This is to show you that we can see pulsing blood. This happens to be water with FDL-Loud bubbles in it. So, in summary, we have some very good photo-stable nanoprobes.
28:23
We see with them live cell tracking with millisecond time resolution and two nanometer spatial resolution, and we have a new way of doing ultrasound. Let me point out that these probes are completely photo-stable, last for at least hours, days,
28:41
and so lots of animal studies are now possible. They're non-toxic and they don't blink. And so future applications also include the tracking of immune cells, cancer cells, looking at the differentiation of stem cells in animals. We also think we can do cell lineage studies seeing maybe a dozen countable generations of cells
29:04
when combined with single cell RNA analysis. We think that would also be very exciting. And so I have one minute left. You might ask, you know, you spent 10 years worrying about climate change and energy. What are you doing this for?
29:21
And the answer is the rest of my copious spare time I do spend on energy and research, new generations of ion batteries, electrochemistry, lithium extraction from seawater, and developing new statistical analysis methods for weather change.
29:41
This in particular is very, very scary. This is a paper we're about to submit, the proceedings in the National Academy. And it has to do with are one in 100 year events getting more frequent? Not using climate models, just statistics. And the answer is they are.
30:00
In a period of time in the United States from 1979 to 2017, they increased 3.6 fold. Unlike climate models, we can also look at rainfall and rainfall in the summer months did not increase as predicted by the climate models, they decreased slightly.
30:20
So if you have these extreme 100 year events where the summer temperatures are increasing 2.7 fold and less rain, this does not look good for agriculture in the US, west of the Mississippi. And so this is just pure statistics, no climate models.
30:42
And this does not bode well, because water aquifers in the US, in California, and in the Midwest, but all around the world are being depleted. And due to climate changes of increased droughts, heat waves, and depleting aquifers, and this is where it's projected to have water stress
31:01
around the world, we could see tremendous stress on water resources and agriculture. Thank you.
Recommendations
Series of 9 media