Dear Faezeh,
the LPA approach is based on the segmentation of the FLAIR images only, which sometimes results in less accurate results depending on the lesions in the image. This approach is mainly thought if you don't have an additional T1 image and/or it is difficult to find a stable threshold for the LGA approach. The multimodal segmentation of T1+FLAIR using the LGA approach might be the more appropriate approach if LPA fails. Fortunately, this only happens for one data set even if this is disappointing anyway. The LPA approach does not allow to control the segmentation using a threshold...
Best,
Christian
>Hi SPMers
>
>I have used LST toolbox on SPM (LPA algorithm) to do Flair segmentation to hyperintensity regions many times. The issue is that for just one of my subjects which Fair data seems OK, the results of LPA does not make sense. Indeed, Flair map as the result of Flair segementation is so exaggerated by enlarged hyperintensity areas. I do not have any idea why this happened to this data and how I can fix that to a real segmented result. Here, I provided two screen shots of original Flair and Flair map (result of segmentation) superimposed on Flair data, you can see and compare how hyprintensity areas are in different size is these two images. Any help to solve my problem is appreciated and please let me know if you need some more information. Thank you so much.
>
>Regards,
>Faezeh Vedaei
>
