Skip to content

Automated pipeline to quantify bidsified PET dataset using Logan plot model

Notifications You must be signed in to change notification settings

CyclotronResearchCentre/PETpro

Repository files navigation

Parent function for metabolite

The correction of metabolite is performed by scaling imput function by a parent function. This parent function is choosen from aviable ones in pet.json configuration file, in section file, together with it's parameters.

UCB-h

For mean over 4 participants the best fit produce following parameters:

Sigmodal:
A0:  1 (fixed)
e:   76.0390292 +/- 25.2213235 (33.17%) (init = 0)
a:   0.92514893 +/- 0.02308516 (2.50%) (init = 1)
b:   230.814420 +/- 38.6266001 (16.73%) (init = 683.7752)

To insert into pet.json:

    "metabolite":{
      "method": "Sigmoidal",
      "parameters": {
        "A0": 1,
        "e": 76,
        "a": 0.925,
        "b": 231
      }
    }

DoubleExp

General model Exp2:
fitresult(x) = a*exp(b*x) + c*exp(d*x)
Coefficients (with 95% confidence bounds):
a =      0.9538  (0.3682, 1.539)
b =   -0.001434  (-0.00326, 0.0003923)
c =      0.1069  (-0.4965, 0.7104)
d =   3.562e-05  (-0.001228, 0.0013)

To insert into pet.json:

    "metabolite":{
      "method": "DoubleExp",
      "parameters": {
        "a": 0.9538,
        "b": -0.001434,
        "c": 0.1069,
        "d": 3.562e-05
      }
    }

Parent fraction:

x = [0, 180, 300, 900, 2100, 3600, 5400];
y = [1, 0.92, 0.76, 0.29, 0.19, 0.14, 0.12];

Partial volume correction

For PVC you will need an extrnal tool petpvc installed and added to path.

The point spread function FWMH

The FWMH was estimated by gaussian fit of point source image using tool pet_fwhm. Results where averaged between centrally placed and z = 10cm:

"FWHM": [6.48, 6.58, 4.67]

Modelling

For the Logan plot, you will need a magia toolbox

About

Automated pipeline to quantify bidsified PET dataset using Logan plot model

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published