Dear Mary,

        In regards to your query below on pixel dwell
time, there are several trades off to consider.

First, regarding the pixel values themselves, ranging from
between 20 to 40 units, my guess is that if you increased
the laser power and/or gain, (or perhaps aperture, I'm not
familiar with Norans, so I don't know their capabilities)
you should get those numbers more into the middle of the
256 grey scale.  If no manipulation will get you into the
range of higher pixel values, then I would be concerned
that something about the light measuring system (PMT?)
and/or the digitization of the signal is not working
correctly.  I don't know the temporal characteristics of the
PMTs, so I don't know if you are working in pixel dwell
time ranges where you are up against response limits of
the PMT.  [but I'm sure there are people on this list that
who can answer that question!]

       Assuming your PMTs can respond at these dwell
times, and that you can increase your basal fluorescence
levels to 100 units or so, then the issue becomes signal
to noise.  With 100 nsec dwell times, I suspect your values
will be so noisy that you will have to do very substantial
binning (either temporally, spatially or both) to see anything!
Using 2 msec linescans, I would say that with a good
strong fluorescence signal [no, I don't have absolute
number of photons, sorry], that you can approach
2 msec/2 um resolution.  That is, if you binned 2 um's
worth of pixels spatially, you might be able to see a line
to line change in fluorescence, assuming its a substantial
jump in fluorescence/calcium and your indicator is
responding fast enough [most have fast on rates].
[This is not a theoretical notion, just a rough practical
estimate based on imaging calcium in neurons.]

      To provide some "back-of-the-envelope" calculations
I would estimate that I have to bin at least 15 pixels spatially
[using a 40x lens] x 5 usec dwell time (on my linescans), so
I have to bin 75 usec of pixel dwell time to see a good signal change,
and presumably more pixels to see a smaller change in fluorescence.
75 usec of pixel dwell time would amount to 750 of your 100
nsec pixels, assuming that the aggregate noise is the same
in the two situations [and I'm not sure that it is; I don't know
if scanning faster increases the aggregate noise for two
equal total dwell times or not....].  So you will have to do some
substantial binning, I expect, to see calcium changes--  e.g. you
could select a 75 pixel ROI and then bin over 10 frames to
get a signal that might be stable enough to see "bin to bin"
changes.

     James Pawley's, "Handbook of Confocal Microscopy,
2nd Edition", 1995, provides excellent background for anyone
trying to understand these sorts of issues, including
discussions of signal, noise, PMTs and much more.

     Please send me an email if you would like some calcium
imaging references that may be relevant.

Good Luck,
Don


Donald M. O'Malley, Ph.D.
Assistant Professor
Department of Biology
414 Mugar Hall
Northeastern University
Boston, MA 02115
[log in to unmask]
www.omalleylab.neu.edu



---------------------------------------------------original message:

Date:    Fri, 7 Sep 2001 14:13:27 EST
From:    Mary Wagner <[log in to unmask]>
Subject: Pixel dwell time

Hello,

We are doing calcium imaging in isolated cardiac myocytes with a
Noran OZ confocal. The system allows us to take a 2-D image
every 2 msec (image size is 128X120 pixels). The dwell time at
each pixel is 100 nsec. We are loading the cells with Fluo3 via a
patch pipette, usually between 50-100 umol. Our problem seems to
be a very small signal, that we are not using the total range available
(ie. 0-255, our signals usually go from 20 at rest to 40 at peak
values (arbitrary fluorescent units)). So I guess I'm wondering if the
dwell time is too short? Are we exciting the Fluo-3 molecules but
then not waiting long enough at each pixel- thereby decreasing our
sensitivity? I really don't know much about confocal microscopy so
any pointers to literature on these issues would be appreciated.

Thanks very much!

Mary

---------------------------------------------------------
Mary Wagner, Ph.D.             Department of Pediatrics
Assistant Professor            Emory University, Atlanta
[log in to unmask]      (404) 727-1336
----------------------------------------------------------------------------
---------------