Hello everyone,
I'm new for MRIcroN. I am conducting a research about a certain drug using influencing the volume of rat hippocampus and indeed changing cognitive function. I used "pen tool" drawing a region of hippocampus and used "fill tool" to cover the whole hippocampus. Then, the descriptive statistics showed as follows:

VOI notes
Center of mass XYZ 87.41x76.83x13.00
VOI nvox[cc]=min/mean/max=SD: 669 0.02 = 12.1168 19.4356 23.3717 = 1.6913
VOI <>0 nvox[cc]=min/mean/max=SD: 669 0.02 = 12.1168 19.4356 23.3717 = 1.6913
VOI >0 nvox[cc]=min/mean/max=SD: 669 0.02 = 12.1168 19.4356 23.3717 = 1.6913
Mode: 18.7517070770264

My question is that is the value "VOI mean=19.4356" can be directly for volume compare? That is I obtain every VOI mean of my rats, and use it for statistic analyse. 

Thanks for your time, and expecting your reply.
Ying Tian, PUMCH

The size of your drawing covers 669 voxels, or 0.02 cubic centimeters.

The range of brightness of the background image voxels has a range of
min...max  12.1168...23.3717 with a mean of 19.4356 and a standard deviation of 1.6913. Since image intensity in MRI scans is relative, these values are typically not of huge interest, though they are informative if you are drawing on a calibrated image (e.g. CT scan, CBF, etc).

Thanks for your reply. It was so timely.
If my research is only to compare the volume change of hippocampus and there is no other lesions, can I just take the coverd voxels of one same slice into statistics analyse? That is, I choose the max hippocampus volume slice, and use pen tool and draw tool to get the voxels of this slice. If the voxels in the same slice changes significantly, I can say the significant difference is exist between groups. I'm not sure whether this way is OK.
Thanks again for your professional support and wish you a good day.

Taking data from a single 2D slice will always be inherently noisier than sampling the entire region. From first principles, SNR follows the square root of the number of averaged samples:
  https://en.wikipedia.org/wiki/Signal-to-noise_ratio
A further concern is that if there are differences in angulation and position between your images, you will likely have large variability caused by how oblique the acquired slice is relative to the region of interest. This is why chefs cut carrots on a bias, to increase the surface area. Likewise, the volume of a non-spherical structure on a 2D slice will vary tremendously if the imaging plane is perpendicular to versus oblique with the structure.