: RE ADVICE PLEASE (14 Dec to 15 DEC)

[Goldsmith, Noel]

Hi,
I make only comments about cameras.
If you want quantitative photometric work, then it can be challenging.
Cooled camers are probably essential, and certainly a very good idea.
I have found that the Diagnostic instruments Spot cameras (SPOT R/T Sliders ) to be good for a wide variety of imaging conditions. They cope with a wide brightness range well.
We look at objects ranging from mirror polished metal (very very bright) all the way down to carbon fibre composites (black, non-reflective).
 The cooling means noise is reduced, and the reproducibility is better, the cooling is usually quoted as 12 degrees below ambient or some such. So if ambient is constant then the camera will be fairly constant.
The use of a chip with a filter in front to provide colour by a 3 exposure path is a good idea if quantitation is what you need. The use of a Bayer mosiac filter in front of the chip for colour is OK for pictures but presents some problems for quantitation of colour images.
I use ImageJ to measure on my images, and let the camera Software do its stuff. The integration of the camera software into (in my case) MAC  OSX is excellent, I  control the cameras with Applescript and automatically bring up new images into ImageJ. Of course, some cameras are able to be driven from ImageJ, and some manufacturers offer a driver kit for the daring, but I now wonder why you would do that, with the level of functionality I achieve very easily. The other Software controls provide with the Spots is very useful for background removal and light levelling.
Best Wishes
Noel Goldsmith


Noel T Goldsmith
DSTO Melbourne
Air Vehicles Division
506 Lorimer Street Port Melbourne
Victoria 3207
AUSTRALIA

phone [phone]
lab phone [phone]
[hidden email]

> Date:    Thu, 15 Dec 2005 09:04:55 +0000
> From:    Tom Smulders <[hidden email]>
> Subject: Advice please
>
> Dear ImageJ users,
>
> I'm a long-time user of NIH Image, but only a recent member of the
> ImageJ list.  I was hoping to solicit some advice on setting up the
> microscopy part of my laboratory, as we have just started a new
> project and I need to get the microscopic quantification going soon.
>
> The project I am working on requires mostly counting of cells and
> measurement of brain structure volumes.  We are looking at seasonal
> patterns in adult neurogenesis, total neuron numbers and hippocampal
> volume in the brains of birds (great tits and willow tits to be
> exact). Our histological strategy will be the following:
>
> We will cut/are cutting 10 micron coronal serial sections from these
> brains (fresh frozen).  We are cutting these on a cryostat (so
> directly onto the slides) and splitting the sections into 10 parallel
> series, so that each series contains sections from throughout the
> brain, spaced by 100 microns.  These sections will be stained with
> fluorescent immunocytochemistry to look for the expression of
> different markers of neuronal differentiation in BrdU-labeled
> cells.  In other words: we need to count single- and double-labeled
> fluorescent cells, as well as count total numbers of cells (probably
> using DAPI in the same sections, or counting from a separate series
> stained with Nissl stain).
>
> My equipment situation is as follows: I have a fluorescent
> microscope, currently without an imaging system on it.  I need to
> purchase a system that can acquire images from the fluorescent as
> well as regular light microscopic sections, and then I need to do the
> counts.  And I would like advice on the following issues:
>
> 1) Would ImageJ be able to do all the quantification I need?
> 2) a) Can ImageJ actually be used as capturing software, or is it
> better to get some software package that comes with whichever camera I buy?
>      b) Any recommendations of a good quality, good value digital >
> camera for low light conditions?
> 3) a) What about stereology?
>      b) Is there a Stereology module/plug-in for ImageJ?
>      c) Can stereological techniques even be applied to the tissue as
> we have sectioned it?
>      d) If I were intent on using stereological methods to count
> total neuron numbers in the hippocampus, am I correct in assuming
> that I would need to cut thicker sections in order to be able to use
> the optical dissector method?
>      e) Can that be done if the sections I take are 100-150microns
> apart throughout the nucleus (doesn't this violate the dictum that
> each particle needs to have the same chance of being counted?)?
>      f) The hippocampus is also rather heterogeneous in its cell
> density (i.e. it is structured, be it not as structured as the
> mammalian hippocampus): are there random sampling schemes that will
> still estimate the total cell numbers accurately?
>
> Additionally, I want the system to be able to measure optical
> densities and count silver grains, but I understand that ImageJ can
> do that without any problems...
>
> Feel free to e-mail me back directly, if it is inappropriate to get
> all these replies to the list.  And I apologize if you get requests
> like this often.
>
> Sincerely,
>
> Tom Smulders
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> Tom Smulders, Ph.D., Lecturer
> School of Biology and Psychology (Division of Psychology)
>
> The Henry Wellcome Building for Neuroecology, University of Newcastle
> Newcastle upon Tyne, NE1 7RU; Tel: ..[phone]; Fax:
> ..44-(0)191-222-5622
>
>