FMRI_DATA_ANALYSIS_REPORT_AND_PREPROCESSING_STEPS .pptx

PREPROCESSING STEPS FOR AN FMRI

1. Learning outcomes: Introduction Steps of preprocessing : Slice timing Realignment Co-registration Segmentation Normalization Smoothing

2. NEURP/ UMAR ASIF NEUROIMAGING DATA ANALYSIS UMER ASIF

3. Statistical parametric mapping (SPM ) is an established statistical data analysis framework through which regionally specific effects in structural and functional neuroimaging data can be characterized. . 

4. SLICE TIMING: Fmri is often collected in slices and slices are acquired with corresponding time points . Fmri data slices can be acquired in ascending descending countinuous or interleave form. Slice time is an interpolation method and modifies an original data . There is risk of artifacts from one volume being propagated to other volume in time series. Slice time proceeds in two steps: Fourier transform of measured signal Phase shifting back to slice and transformation back into signal space

5. Realignment : Realignment refers to correcting the functional data for movement that has occurred during scanning .Realignment occurs in two steps: Estimation Reslicing

6. Resampling In this step we apply estimated transformation for each image

7. Estimation Acquiring rigid body transformation from each image to scan .This consist of six parametres consist of translation and rotation parametres . Translation parametres elaborates the measurement around each axis and rotation parametres describe the rotation around each axis.Siz parametres also called head motion parametres are calibrated to minimize difference volume which shpuld be realigned and the reference volume

8. Co-registration The next stage is co-registration.. The realigned functional images need to be linked to the structural – the structural has superior anatomical localisation , the functional has the BOLD signal – need the two overlaid together

9. We need to know how well one set of images matches the other – this is achieved by fitting the source image with a fixed reference image ( ie fitting the functional to the fixed, anatomically superior structural) and then looking at the intensity of the voxels in the functional w.r.t the structural. This works by estimating the intensity of groups of voxels in functional image and relating to the structural – interpolation. Principle is affine registration and maximisation of mutual data Co-registration

10. Segmentation: Segmentation means a single image is differentiated into different tissues that exist in human brain. Usually differentiation can improve mapping of image into standard space. Classifies voxels with in an image into different anatomical division: Grey matter White matter Cerebro Spinal fluid

11. Normalization: A form of coregistration : between subjects To warp images from individuals into the same standard space(a template) to allow averaging across subject. There are inter individual differences in brain anatomy individual brain differ in shape ,size and in folding .In order to correct all the differences we perform normalization. The goal is then to have same anatomical localization from different brains in some voxel. Normalization occur in two steps: Affine transformation Non linear transformation

12. Affine transformation: Determines the optimum 12-parameter affine transformation to match the size and position of the images 12 parameters = 3 translation 3 rotation 3 scaling/zooming 3 for shearing or skewing Fits the overall position, size and shape

13. Non-linear Registration (warping) Warp images, by constructing a deformation map (a linear combination of low-frequency periodic basis functions)

14. Smoothing: Smoothing is one of the last steps and aims at blurring the functional images to correct for any remaining functional and anatomical differences between subjects. Nevertheless the more smooth the image is less resolution we can get. .