>  My apologies to Eric.  We don't want to ruin his vacation
 
Apologies are in order:
 
>One important point here not mentioned is the presence of the pinhole or
>slit in the back focal plane of the objective lens.
>This can exclude up to 99% of all emitted light, requiring use of high
 
 
What is the basis of this number?  Have you measured it?  Have you tried to
do WF/Decon on any specimen in which such a situation could occur?  Please
be careful how you answer.
 
 
>Therefore, the image
>that one acquires with a confocal microscope has these aberations present
>in the image -- yet they are published and accepted as "real" (implying
>"nonartifactual") routinely!  If I were to publish a picture of myself as
>seen in a warped concave/convex funhouse mirror and call it my "true"
>image, people would most likely not be fooled.  Similarly, to publish a
>confocal micrograph of an image and assume it has no artifacts is
>foolhardy.  The artifacts are present -- you just can't point a finger at
>them and say they are artifacts.  Isn't this an artifact of the worst sort?
 
The main effect on confocal of the aberrations noted is to make the signal
disappear.  From the point of view of Science, making things appear is more
of a problem and it is a problem that those who employ deconvolution must
continually struggle to avoid.
 
Now about glass houses....
 
>An empirically derived point spread function
>*does* represent the distortion properties of a given microscope -- with
>*that* coverslip, with *that* mounting medium, with *that* emission filter,
>with *that* objective lens, with *that* alignment, on *that* day.  Under
>these conditions, the point spread function should accurately represent the
>distortion properties of that microscope at that given point.
 
Quite right. "...at that given point."  But what about points at different
points and levels inside the specimen?
 
>We have
>found that, when using a high NA Plan Apo objective lens, this point spread
>function varies little over the field of view that can be acquired with the
>CCD chip.  The point spread function does vary significantly towards the
>edges of the lens (contrary to what the manufacturers say!).  Keep in mind
>that these same distortions, as visualized through a point source of light,
>are present with that same lens in your confocal images as well.  If you
>see distortions of microbeads with your objective lens, you can bet those
>same distortions are showing up in your confocal images --and with your
>confocal microscope, you have no way of correcting for these aberations,
>whereas with true deconvolution techniques, you do.
 
One minute you say that you are using a single PSF and the next you have a
different one for each part of the image.
 
As mentioned before, confocal data can also be "deconvolved" and the
results show a similar improvement over "unprocessed" WF data.  My own bias
is that this process should be used more.
 
Cheers,
 
Jim Pawley
 
 
                   ****************************************
Prof. James B. Pawley,                                       Ph.  608-263-3147
Room 1235, Engineering Research Building,                    FAX  608-265-5315
1500 Johnson Dr., Madison, WI, 53706          [log in to unmask]