@C Elegans: I was just going to send you a PM on this subject, but, since you bumped it … Sorry for the lengthy delay, but I just finished my last final exam of the semester, so it has been a busy few weeks.
Originally posted by C Elegans You are of course correct that we will probably never be able to get rid of the PET scans, they are always necessary to investigate a specific brain.
Yes; I think I got carried away with that suggestion.
Originally posted by C Elegans However, as I mentioned earlier, each and every radioligand we used must first be modelled in order to obtain informative data from a specific brain. When PET scans are made in order to model a radioligand, we are not looking for characteristics of an individual brain, instead every brain (for instance the brains of 6 healthy control subjects) is used to analyse how the radioligand behaves in the tissue.
*Nods head* Yep, yep.
Originally posted by C Elegans When it is concluded from autoradiography of brain slices that a radioligand might be suitable for PET, it has to be modelled so we know how to interpret the data obtained by the PET system. This is always done on monkeys first, then at a group of healthy control subjects. Data from time-radioactivity curves, plasma and full blood are then used to create a model for that specific radioligand. This model is then used to evalute all following PET scans, ie those scans are for really investigating characteristics of brain functioning or diseases. Now, every single PET scan costs about $3000, plus a lot of time and work for the scientists, so if some or even all of these scans could be simulated instead, it would be very valuable. The simulations that are currently available are far from sufficient, they can't yet add anything to modelling in vivo. So this might still be a possible future use for FEA or other mathematical modelling methods. If an eyeball can be simulated with FEA, why not a brain, based on data from n healthy subjects, that can be used over and over again for modelling of different radioligands?
Yes. As I have mentioned in previous posts, any time you wish to make a model of any kind of movement of a substance (or energy) through another substance, FEA can be of great help. Here, you wish to track a radioligand through tissue, and, given the proper data, I have to assume an FEA specialist could help you out. My knowledge of the human brain is – uh, lacking, to say the least. The only possible problem I can see cropping up is this: are individual brains different enough to make a generalized model inapplicable? I suppose you are looking at
tissues, which, I have to guess are the same for all people … it just seems almost beyond belief that everybody’s brain is, in essentials, the same! Frankly, I find that somehow disturbing. But that’s just me.
Originally posted by C Elegans Now, this is what I learned at the conference: The first and most acute problem the PET-world needs to solve, is the generalisation from dead to living brain tissue. A lot of hypothesis testing made with PET, is based on the assumption that the receptors in brain stay the same in life as after death. Post mortem studies of the brain allows obviously allows for a lot more detailed and sophisticated studies of cells than PET does, but there is one problem - degradation. Receptors are made of protein configuration, and proteins are highly sensitive to temperature. Heat will cause the proteins to degrade, and the receptors to dissolve. Temperature changes is our worst enemy when we try to extrapole post mortem data to testable hypothesis in the living brain. What is worrying, is that recent in vivo PET data from schizophrenic and depressed patients, do not coincide with the large body of post mortem studies. To take a simplified example: post mortem studes report abnormally low density of receptor X1 in brains of dead schizophrenic patients, suggesting they had a too low X-transmission. So, PET and fMRI labs around the world start PET studies or X-transmission - demonstrating no difference between living schizophrenic patients and living healthy controls. So what does this mean? Hyopthesis are many, but if we had a model for how receptors degrade in a dead brain, we could exclude errors caused by unsecurity in the fate of receptors after death.
Now I have to get morbid, I hope you excuse me. Two senior researchers at my lab are trying to address this problem. They have obtained lots of temperature measurements of dead brains in different parts of the brain, at different "depth" and over time. The brain consists of different kind of tissue volumes and surfaces with different conductive or insulating properties. They are looking for a model that can address the heat loss over time in all those different "compartments". Is FEA such a model? One of them believes it is, but he knows virtually nothing of the method, so if we knew what we were looking for, we would be able to perhaps consult a suitable person for what models to use and get software to use it, or even hire a mathematician or engineer to do the calculations.
Well, now you are getting back into things more my speed! I understand perfectly how going from recently dead (or not-so-recently dead) subjects to live ones could create some serious experimental difficulties. Frankly, I am a little surprised dead tissue would ever be looked at – I would think it vastly different from live tissue - ? But what do I know!
Anyway – yes, heat transfer is a strong suit of FEA. Given the proper data on the tissue (such as its thermal conductivity, density, etc) and an idea of initial temperatures, an FEA specialist could set up a model to track heat loss through a brain. A couple of difficulties could arise:
1) Very specific data would be needed for those initial conditions and tissue properties. Again, you would run into problems getting this information from LIVE subjects. I would not know how to go about getting a thermocouple (ie thermometer) into somebody’s (living) brain; nor how to measure heat transfer through a living brain. I suppose such things are possible …
2) Again, I have to question how different individual brains are from one another - ? Slightly different geographies within the brain could make for rather different heat transfer paths.
Originally posted by C Elegans The second issue is much more speculative, and is somewhat related to my question above whether our emission could be viewed as an energy flow in the same way as heat.
The MRI-people use FEA to calculate how heat behaves in the brain - one of the major problems with MRI is that it heats the brain tissue, creating disturbances and of course a potential risk to subjects. The idea here is that maybe our gamma emission could be analysed in the same way as heat emission in MRI?
If I understand this abstract correctly, they here use FEA (or finite-support extrapolation, whatever that is) in a context that resembles ours.
This one is beyond me, too. The abstract you have here is just plain unintelligible to me.
But, if you ask whether positron (or gamma) emissions could be modeled by FEA, I would have to doubt it. The basis of FEA is looking at
matter, and observing how finite elements of that matter interact with one another when subjected to some stress or energy or such. As far as I know, positrons do not require any actual matter to travel through – they could move through empty space as easily as through a brain. Also, I would doubt that the matter they travel through has any impact on how they travel. So there would be no
starting point for FEA to begin it’s analysis, and no way for an FEA model to take that starting point and track positrons as they move through a substance.
Hmmm. That is a hard concept to explain. Does that make sense?
