HTTPS SSH

Analysis code for natural scenes aperture experiment

Preprocessing

Subject-level

Filesystem

  1. Make the subject's directory structure:

    mkdir -p sXXXX/{analysis/,fmap,func/run{01,02,03,04,05,06,07,08,09,10,11,12},loc_analysis,logs,mvpa,reg}
    
  2. Copy the subject's runtime logfiles to the logs directory.

  3. Make symlinks named raw in each functional run directory (exp and loc) that link to the location of its associated raw DICOM directory:

    ln -s /labs/olmanlab/DICOM/YYYYMMDD/sXXXX/MR-such_and_such raw
    
  4. Similarly, make symlinks named mag-raw and ph-raw in each fieldmap directory that link to the locations of the fieldmap acquisition:

    ln -s /labs/olmanlab/DICOM/YYYYMMDD/sXXXX/MR-SEyada mag-raw
    ln -s /labs/olmanlab/DICOM/YYYYMMDD/sXXXX/PH-SEyada ph-raw
    
  5. Make a local copy of the AFNI/SUMA base anatomical that we can use for alignment:

    3dcopy \
       {$SUBJECTS_DIR}/{$SUBJ_ID}/SUMA/{$SUBJ_ID}_SurfVol+orig \
       reg/{$SUBJ_ID}_ns_aperture-anat+orig
    

Conversion

Converts from the raw scanner format to a set of 4D NIFTI files:

ns_aperture_preproc sXXXX convert

After execution, open up each NIFTI file and inspect for image quality and look at the summary image to see how much movement there was.

Fieldmap preparation

Prepares the fieldmap:

ns_aperture_preproc SXXXX fieldmap

Correction

Applies a motion and distortion correction procedure:

ns_aperture_preproc sXXXX mc_unwarp

After execution, open up the summary NIFTI file to check that most of the motion has been removed. To verify that the unwarping has worked correctly:

  • Run fslview.
  • Load the original or corrected image from a given run.
  • Add the magnitude image from the fieldmap as an overlay.
  • Notice the geometric distortions in the functional data.
  • Add the undistorted image as an overlay, and hide the uncorrected image.
  • Toggle the visibility of the undistorted image, and verify that the geometry now aligns well with that of the fieldmap's magnitude image.

Anatomical registration

First, make a copy of the mean functional:

cd reg
3dcopy ../func/sXXXX_ns_aperture-mean.nii sXXXX_ns_aperture-mean+orig

Now, we want to calculate some transformation parameters that will get the two images into rough register. This will give the automated algorithm a good starting point.

  • Start AFNI, from within the reg directory.
  • Set the reference anatomical as the underlay.
  • Position the crosshairs at a landmark on the brain. I like to use the most posterior portion of the occipital lobe, on the right side (in the image). Note down the three position values in the AFNI window ( in mm). Say they are [ 100, 50, 50 ].
  • Then, change the underlay (or overlay, if you prefer) to the mean functional.
  • Position the crosshairs at the same landmark as you used for the anatomical. The position might now be [ 20, 10, -20 ].
  • Calculate ( reference anatomical positions - functional positions ), elementwise. In this example, that would give [ 80, 40, 70 ].
  • Update the subject's configuration structure to include the estimate.

Then run:

ns_aperture_preproc sXXXX sess_reg

Surface projection

Projects the functional images to a standardised cortical surface, averaging between the white and pial surfaces:

ns_aperture_preproc sXXXX vol_to_surf

Smoothing

Applies a small amount of spatial smoothing to the timecourses, calculated along a surface that is the average of the white and pial surfaces:

ns_aperture_preproc sXXXX smooth

Temporal filtering

High-pass filters each run's timecourse, for use in the MVPA procedure only:

ns_aperture_preproc sXXXX filter_tc

Mask creation

Calculates a mask surface that specifies the nodes that are nonzero in each functional run:

ns_aperture_preproc sXXXX surf_mask

Group-level

Filesystem

Run:

cd /labs/olmanlab/Data7T/NatSceneAperture
mkdir -p group_data/{cluster_sim,loc,mvpa,ret}

Average anatomical

Create an anatomical that is the average of all the subjects in the experiment:

make_average_subject \
  -subjects s1000 s1008 s1011 s1021 s1032 \
  -out ns_aperture_avg \
  -sd-out /labs/olmanlab/NatSceneAperture/group_data/

Mask creation

Create a mask that indicates the nodes that are nonzero in all the individual subject masks:

ns_aperture_group_analysis group_mask

Cluster simulation

Runs a Monte-Carlo simulation to determine the FWE cluster threshold for each hemisphere:

ns_aperture_group_analysis clust_sim

After running the above, edit the two scripts to replace Surf_A in the SurfClust call with midway. Then (takes a while):

script_lh.sh
script_rh.sh

Retinotopy

Average each subject's wedge and ring maps, in a phase-sensitive way:

ns_aperture_group_analysis ret_std

Searchlight preparation

Identifies the nodes associated with a searchlight around each node. The single-subject mvpa_prep needs to have occured, for the representative subject, prior to running this. Takes ages:

ns_aperture_group_analysis mvpa_node_prep

Univariate experiment analysis

Subject-level

Design preparation

Computes the experimental design from the logfiles:

ns_aperture_analysis sXXXX design_prep

GLM

Estimate the GLM:

ns_aperture_analysis sXXXX glm

Cluster summary

After the group-level clustering has been done, run:

ns_aperture_analysis sXXXX coh_clust_summ

Group-level

Height threshold

Runs a one-sample t-test on the subject beta weights:

ns_aperture_group_analysis coh_test

Cluster threshold

To apply the cluster threshold:

ns_aperture_group_analysis coh_clust

Cluster summary

To print out a summary of the cluster beta statistics:

ns_aperture_group_analysis coh_effect_size

Univariate localiser analysis

Subject-level

Design preparation

Generate the design info:

ns_aperture_analysis sXXXX loc_design_prep

GLM

Execute the GLM:

ns_aperture_analysis sXXXX loc_glm

Group-level

Height threshold

Runs a one-sample t-test on the ( either > 0 ) regressor:

ns_aperture_analysis loc_test

Cluster threshold

Run:

ns_aperture_group_analysis loc_clust

Multivariate analysis

Subject-level

Design and data preparation

This saves the node info, the condition info, and the z-scored block data for a given subject. Run:

ns_aperture_analysis sXXXX mvpa_prep

Classification

Run:

ns_aperture_analysis sXXXX mvpa

Group-level

Height threshold and cluster

Run:

ns_aperture_group_analysis mvpa_test

Task analysis

Subject-level

Run:

ns_aperture_analysis sXXXX task

Group-level

Run:

ns_aperture_group_analysis task_analysis