CNR Analysis

[marcomeixner]

What would you like to see added in this software?

CNR Analysis is mentioned in this paper:
https://onlinelibrary.wiley.com/doi/full/10.1002/mrm.20900

Accoding to the paper:

"The main advantage is that a functional experiment is not required to quantify the relative performance of different acquisition schemes. Relatively robust and reproducible activation can be induced in some brain regions (e.g., motor, visual, and auditory), but even here the interscan variation can be substantial
...
The method can be used to compare acquisitions with different TEs or flip angles, different image recon- struction methods, different coil selections, different pulse sequences, or any arbitrary combinations of the above. The only requirements are that the underlying contrast mechanism is the same (e.g., T*2 or T2 decay), the head position does not change significantly, and for the measurements to be comparable the shim has to remain constant so as not to affect the decay behavior.
...
The use of differential contrast calculated from simple SNR measurements as a metric for BOLD sensitivity proved to be very useful because it alleviates the need for functional stimulation to assess different acquisition schemes. In principle, the approach should make it possible to predict gains and losses in t-scores more accurately than model approaches that require quantification of the different noise sources (i.e., physiology-, protocol-, and hardware-dependent noise). Regarding the dCNR calculations, it should be noted that this treatment is valid irrespective of the form of H(TE), and should yield optimal results in all brain regions."

I am looking for a good metric to compare different BOLD parameters in their activation performance. I want to evaluate a large pool of data in an automated manner - so without doing what is called above a "functional experiment" for each parameter set and brain region.

Best wishes!

Do you have any interest in helping implement the feature?

Yes

Additional information / screenshots

No response

[oesteban]

@marcomeixner how much of this sensitivity metric overlaps with that suggested in #171?