Hello everyone. I have one question about one particular mounting media, Immuno Mount Epredia Immu-Mount:Histology and Cytology:Slide Mounting Media | Fisher Scientific, our user used to mount the sample. In the description, it mentioned "Refractive index: 1.495 ±0.002 (liquid); 1.586 ± 0.002 (dry)
". The RI after dry is higher than common mounting media we use such as Prolong Gold, 1.46. We are using 20x/0.75 objective to image a brain section of 15um thickness on a point scanner confocal. We see blurry nuclear stained with DAPI (also other channels as well) compared to previous sample mounted in 70% glycerol (RI of ~1.44). I suspect the higher RI contributed to the blurry image. Unfortunately, the air objective doesn’t have a correction collar for me to test. I’m wondering if this makes sense and if I miss something else that also contributes to the blurry image in this scenario. Thanks for your suggestion.
This sounds reasonable to me, but “blurry” is pretty subjective. Could you provide some numbers, such as the resolution in Z is “X” lower? I could see this difference in RI reducing the resolution by some percentage, like 20%, but if it is 2x or 3x worse than the issue may be somewhere else. I also think that the RI would make the largest difference in Z due to increased spherical aberration and X/Y should not be as affected.
Thanks @Hazen_Babcock for your help. I measured the ratio of standard deviation over mean intensity on those two images. The one with ImmunoMount (RI 1.59) has 0.787 while the image with 70% glycerol has 1.101. The imaged region are slightly different but both on hippocampus. Can I say it has a ~25% decrease in lateral focus? Thanks.
I think you usually do this be measuring the size of something with a known size, like a fluorescent bead. I’m not sure that your calculation provides an accurate estimate of the resolution.
Nikon has some documents that might be helpful, in particular this one.
Thank you for sharing the information. That would be great if I could make test beads sample to measure. Since I don’t have the beads sample yet, I used the CellProfiler’s ImageQuality module, to check the normalized variance across the image. I can also use edge filter on my image, then calculate the normalized variance of detected edge. I understand this is not the same as measuring the bead size, would this information be an indicator of focusing quality or increased spherical aberration?
If you have something that you know is an edge, like a stained cell nucleus then I think you can get some idea of what the resolution is by looking at how sharp the edge appears.
I’m not really sure whether the normalized variance would tell you what you want to know. Is this something that you are measuring in 3D? Spherical aberration will primarily affect the Z resolution and not so much the XY resolution.
I’m not measuring any structure in 3D for this study. However, I will make some z measurements just to get some idea of how increased RI mismatch affects the z resolution. Thanks again for your help.
Hi @szh141 ,
I do not have the solution for you, but I commonly obtain a measurement of resolution by performing Fourier Ring Correlation in FIJI, on subsequently acquired frames in a time course. I average across a couple measurements from the time series, and I take various FOVs of the sample to see the distribution.
- Obtain a small time course on the microscope, of the same FOV.
- Run FRC in FIJI and designate two consequtive frames of the time course as the input.
- Obtain the calibrated FRC value as a measure of resolution.
Thanks @JelleP for sharing. That’s very helpful. For SMLM, splitting data into two halves and reconstructing separately into two images feeding into the FRC plugin. In my case, I think I need to take two images of the same FOV for estimating 2D resolution.
Usually if someone comes in with a mounting media that I haven’t heard of, they’re also likely using coverslips which aren’t #1.5 or #1.5H.
But with a 20x/0.75 objective, I’d only expect this to matter if the coverslips were #0 or #3+.
After confirming that the coverslip is sufficiently clean, ask them to do a quick test by using different mounting media and using the same mounting media (but this time sealing with nail polish or something to prevent curing). If the problem is completely due to the high 1.586 RI, then the uncured 1.5 RI should look fine.
Thanks @WillGiang . They got the coverslip from us, so it’s No 1.5. I didn’t catch their mounting media RI in the first place.
The user actually did some test, and sealed the uncured mounting media. The edges of the nuclear did look sharp again.