What do people expect from their physicians?

[HopkinsMedStudent]

Sigh. I figured it would be something like this. You have glommed onto one Nyquist effect and declared it to be everything you need to know.

And then, you try to "show" me by presenting an image of a 256 by 512 dataset as a 512 by 512 image. And you don't even get the joke! Just the fact of expanding the image introduces artifacts. Since it's a power of two, it only produces one variation in the frequency domain, which is easy to filter out, but it still affects the image.

Nor is that the only source of error. Some of it comes from the rectangular resampling. I can even see features in that image that are likely to be sampling artifacts.

So, how's about you, since you know all you need to know, show me where they are, hmm?

Let me explain this one more time for you. You and I both know that your original argument of "radiologists must know Nyquist in order to properly interpret MRI" is BS. Instead of simply acknowledging this fact, you spin new arguments and shift the goalposts into other areas.

1) Nyquist minimum resolution is NOT the threshold used for determing actual MRI resolution. T2* effects dominate over Nyquist and produce a larger minimum resolution.

2) The computer calculates this minimum resolution (which is a HIGHER threshold than Nyquist) and the radiologist goes by this number.

3) Therefore, since the computer doesnt use Nyquist to compute a resolution limit, and since the computer spits out the number for the radiologist instead of the radiologist computing it himself, KNOWING NYQUIST IS IRRELEVANT TO A CLINICAL RADIOLOGIST.

epepke, if you are going to claim that radiologists are "stupid" because they dont know what Nyquist is, you need to show that knowing Nyquist is relevant in terms of their work (i.e. interpreting images in a CLINICAL context).

If a radiologist's failure to know what Nyquist is, led them to make faulty conclusions as far as image interpretation, then you might have a case. Almost every MR image has some degree of a Gibbs ringing artifact in it. Yet, these artifacts DONT AFFECT CLINICAL IMAGE INTERPRETATION, so to hold radiologists accountable for knowing the mathematics of the Gibbs effect is really superfluous to their work.

But even you admit that you "have no clue" as to how clinical interpretation is done, or how its affected by these things, so I reject your argument based on that alone. Nevermind the fact that your understanding of MRI is superficial at best and that you have confused Nyquist resolution limits as being the final arbiter of MRI resolution WHICH IT IS NOT!

And then, you try to "show" me by presenting an image of a 256 by 512 dataset as a 512 by 512 image. And you don't even get the joke! Just the fact of expanding the image introduces artifacts. Since it's a power of two, it only produces one variation in the frequency domain, which is easy to filter out, but it still affects the image.

Sure it affects the image, but not in a CLINICALLY RELEVANT CONTEXT. You obviously dont understand how MRIs are interpreted in a clinical setting.