Welcome to Brant!

To facilitate data processing and deal with above listed issues, we’ve written an extendable MATLAB based toolbox BRANT (BRAinNetome Toolkit), which integrates fMRI data preprocessing, voxel-wise spontaneous activity analysis, functional connectivity analysis, complex network analysis, statistical analysis, data visualization as well as several useful utilities. We designed the toolbox using dynamically generated GUIs, with which other developers can generate their own GUIs by adding a few lines of MATLAB code. Also, to simplify the input process during using BRANT, most functions are initialized with default settings, users will only need to specify several necessary parameters, with free access to all.

Functions of BRANT are arranged into 7 modules, which are preprocessing, functional connectivity (FC), spontaneous activity (SPON), complex network analysis (NET), statistics (STAT), visualization (View) and utilities. More details on proper module can be found in its own part.

Please cite this work if you use the Brant.

Xu, K., Liu, Y., Zhan, Y., Ren, J., Jiang, T. (2018) BRANT: A Versatile and Extendable Resting-State fMRI Toolkit. Front Neuroinform, 12:52.

_images/main_gui.png _images/arrangement.png

Preprocess

Raw data collected from MRI scanners are formatted as DICOM (Digital Imaging and Communications in Medicine) files, which are firstly converted to a single 4D NIfTI (Neuroimaging Informatics Technology Initiative) image for efficiently processing. For converted data, visual inspection is recommended to censor data with low quality (artifacts and distortions). Qualified data can be further processed in the preprocessing pipeline.

System Configuration

_images/preprocessing_1.png
  • Output to wk dir: Set to output results to wk (working) directory defined below. BRANT will create new directory for each subject and copy necessary files to the new directory, then start processing.
  • Check Board: Open/Close CheckBoard.
  • Sync: Synchronize parameters of TR in slice timing and denoise.
  • Parallel Workers: The number of workers used during processing. e.g. when set to 2, BRANT will run 2 subjects in parallel. The processing speed depends on both CPU and Hard Drive speed, if there are a lot data IO with less computation task, set to more workers will slow down the entire process.

Directories

_images/preprocessing_2.png
  • wk dir: Working directory to save intermidiate files. By default is set to the current directory.
  • data dirs: directories of each subject, can be input from an SPM input dialog of directories or from a *.txt file filled with one directory at a line.
  • filetype: Initial filetype for processing, normally wildcard after DICOM conversion. The item can update itself after each process.
  • data in 4D: Checked means input data is in 4D format, which is highly suggested. If 3D file format is used, each subjects directory will have up to thousands of files after process.

Preprocess Modules

For parameters, press help in each input dialog.

_images/preprocessing_3.png

Slice Timing

Correct for timing information of each slice during one TR.

Realign

Correct and estimate spatially the head motion.

Coregister (optional)

Coregister structural image to mean functional images.

Normalize

Normalize functional images to standard space (both SPM12 and SPM8 methods is valid).

Denoise

Multi-variable regression and filter.

Smooth

3D spatial smooth with Gaussian kernal.
  • Buttons:
    • R: Refresh (only checkboxes, parameters will remain untouched). Uncheck all selected items and recover the Run button when an error occurs.
    • S: Save parameters of the current panel to a *.mat file. The *.mat can be further loaded for the panel or be used in a script processing.
    • L: Load parameters from *.mat for the current panel.
    • ?: Help information.

FC

Functional connectivity is calculated as the temporal correlation between pairs of time series extracted from ROIs or voxels.

In BRANT, three methods of preparing ROIs are provided, including drawing spheres/cubes from coordinates, extracting ROIs from an atlas and merging separate ROI files into one number-tagged template.

Draw ROI

Draw ROIs is implemented as automatically drawing spheres or cubes with ROI coordinates and a header reference 3D image. The ROI coordinages and labels are sorted in a *.csv table for output indexing purpose, while the header reference image is used to define the output image properties such as bounding box, originator, orientation, inclusive mask, and voxel size.

_images/draw_roi.png

  • type: draw roi as sphere or cube

  • unit: unit of input radius

  • radius: radius of sphere or 1/2 edge length for cubic

  • input type:
    • manual: input coordinates seperated by ‘;’

    e.g:

    25,5,10;10,15,10
    • file: input a csv file with 3 columns for x,y,z (the first line should be ‘x’, ‘y’ and ‘z’)

  • mask roi afterwards: mask generated roi using input mask

  • output to one roi file: instead of output one file for each roi, BRANT will generate one roi file with each roi labeled by number

  • ref&mask: reference and mask file. for extracting information of origin, voxel size, bounding box, etc..

  • out dir: directory for output

  • Buttons:
    • S: Save parameters of the current panel to a *.mat file. The *.mat can be further loaded for the panel or be used in a script processing.
    • L: Load parameters from *.mat for the current panel.
    • ?: Help information.

Merge/Extract ROIs

Given a number-tagged atlas, a subset of ROIs indexed by integers can be extracted and exported to one 3D image. At the opposite, given seperated ROI files, the current function can also merge them into one combined atlas-like ROI file, with ROI labels stored in a *.csv table.

_images/merge_roi.png
  • Operation: select to merge 3D rois into one, or extract ROIs from an atlas.
  • Merge:
  • filetype: files in the filetype will be searched in input directories.
  • data dir: directory in which stores 3D rois.
  • out fn: output filename
  • out dir: output directory
  • Extract:

  • roi file: ROIs in one nifti file

  • roi info*: (* means optional) labels of tagged ROIs in a *.csv file. For example:

    1,SFG
2,MFG
3,IFG

  • roi index: a vector of integers. used for selecting wanted ROIs.

  • output to single file: choose to output to only one file.

  • out dir: output directory

_images/extract_roi.png
  • Buttons:
    • S: Save parameters of the current panel to a *.mat file. The *.mat can be further loaded for the panel or be used in a script processing.
    • L: Load parameters from *.mat for the current panel.
    • ?: Help information.

ROI Calculation

With a predefined atlas-like ROI file and a descriptive number-label table, the current function can extract mean time series from ROIs and voxels, and calculate Pearson’s correlation as well as its Fisher-z transform. An option is provided to calculate partial correlation between each pair of ROIs, with mean signals of other ROIs as covariates.

_images/roi_calculation.png

  • roi file: ROIs in one nifti file

  • roi index(*): optional. labels of tagged ROIs in a *.csv file. For example:

    1,SFG
2,MFG
3,IFG

  • clustersize thr: threshold of cluster size.

  • mask: could be whole brain mask or gray matter mask.

  • id index: identifier to find unique string for each subject

  • filetype: files in the filetype will be searched in input directories.

  • 4D nifti files: if the input data is 4D, check this item. Otherwise uncheck.

  • input dirs: directories can be input either using a *.txt file or spm select window.

  • extract mean: extract mean time series for each ROI

  • roi to roi correlation: calculate correlation between pairs of ROI

  • roi to whole brain correlation: calculate correlation between each ROI’s mean time series and voxels in the mask.

  • Partial correlation: (check to use Partial correlation, uncheck to use Pearson’s correlation) when calculating correlation, between one roi mean time series and voxels/other time series, the rest of roi mean time serieses will be regressed out from the calculation.

  • out dir: output directory for saving results.

  • Buttons:
    • S: Save parameters of the current panel to a *.mat file. The *.mat can be further loaded for the panel or be used in a script processing.
    • L: Load parameters from *.mat for the current panel.
    • ?: Help information.

SPON

Voxel-wise metrics of time series implemented in the current module include amplitude of time series (AM), (fractional) amplitude of low frequency fluctuation (ALFF/fALFF), regional homogeneity (ReHo), functional connectivity density (FCD) and functional connectivity strength (FCS).

AM

AM is calculated as the average amplitude and the standard deviation of the mean-subtracted time series. The AM represents the strength of time series’ temporal fluctuation, which is similar to ALFF/fALFF.

_images/am.png
  • mask: could be whole brain mask or gray matter mask.
  • id index: identifier to find unique string for each subject
  • filetype: files in the filetype will be searched in input directories.
  • 4D nifti files: if the input data is 4D, check this item. Otherwise uncheck.
  • input dirs: directories can be input either using a *.txt file or spm select window.
  • time series: choose the time series to calculate, seperated by ‘,’.
  • mean temporal ampilitude: calculate absolute value of detrended and demeaned time series.
  • standard deviation: calculate standard deviation of time series
  • variation: calculate variation of time series
  • normalize transform: in output file, a suffix of _m means the output is divided by mean intensity in the mask; a suffix of _z means the output is subtracted by mean intensity and divided by standard deviation.
  • out dir: output directory for saving results.
  • Buttons:
    • S: Save parameters of the current panel to a *.mat file. The *.mat can be further loaded for the panel or be used in a script processing.
    • L: Load parameters from *.mat for the current panel.
    • ?: Help information.
  • Reference:
  1. Liu Y, Yu C, Zhang X, Liu J, Duan Y, Alexander-Bloch AF, et al. Impaired long distance functional connectivity and weighted network architecture in Alzheimer’s disease. Cereb Cortex 2014; 24(6): 1422-35.

ALFF/fALFF

ALFF is calculated as the amplitude of the time series in a certain frequency band, which is the averaged square root of the power spectral density of the filtered time series. To increase the stability of ALFF across subjects, fALFF was proposed as calculating the fraction of a certain frequency band against the whole available frequency band.

_images/alff_falff.png
  • mask: could be whole brain mask or gray matter mask.
  • id index: identifier to find unique string for each subject
  • filetype: files in the filetype will be searched in input directories.
  • 4D nifti files: if the input data is 4D, check this item. Otherwise uncheck.
  • input dirs: directories can be input either using a *.txt file or spm select window.
  • time series: choose the time series to calculate, seperated by ‘,’.
  • TR: repetition time, used as sample frequency 1/TR to estimate width of frequency band.
  • lower cutoff (Hz): lower cutoff for band pass filter.
  • upper cutoff (Hz): upper cutoff for band pass filter.
  • normalize transform: in output file, a suffix of _m means the output is divided by mean intensity in the mask; a suffix of _z means the output is subtracted by mean intensity and divided by standard deviation.
  • out dir: output directory for saving results.
  • Buttons:
    • S: Save parameters of the current panel to a *.mat file. The *.mat can be further loaded for the panel or be used in a script processing.
    • L: Load parameters from *.mat for the current panel.
    • ?: Help information.
  • References:
  1. Zang YF, He Y, Zhu CZ, Cao QJ, Sui MQ, Liang M, et al. Altered baseline brain activity in children with ADHD revealed by resting-state functional MRI. Brain & development 2007; 29(2): 83-91.
  2. Zou QH, Zhu CZ, Yang Y, Zuo XN, Long XY, Cao QJ, et al. An improved approach to detection of amplitude of low-frequency fluctuation (ALFF) for resting-state fMRI: fractional ALFF. J Neurosci Methods 2008; 172(1): 137-41.

ReHo

ReHo is calculated as the Kendall’s coefficient of concordance (KCC) among a seed voxel and its neighbor voxels, which indicates the degree of spontaneous activity in the seed voxel’s vicinity. Voxels of higher intensity in ReHo maps indicate greater similarity among neighboring voxels’ time series.

_images/reho.png
  • mask: could be whole brain mask or gray matter mask.
  • id index: identifier to find unique string for each subject
  • filetype: files in the filetype will be searched in input directories.
  • 4D nifti files: if the input data is 4D, check this item. Otherwise uncheck.
  • input dirs: directories can be input either using a *.txt file or spm select window.
  • time series: choose the time series to calculate, seperated by ‘,’.
  • nbr type: number of neighbor voxels, 6 face neighbor, 18 for edge neighbor and 26 for vertex neighbor.
  • normalize transform: in output file, a suffix of _m means the output is divided by mean intensity in the mask.
  • out dir: output directory for saving results.
  • Buttons:
    • S: Save parameters of the current panel to a *.mat file. The *.mat can be further loaded for the panel or be used in a script processing.
    • L: Load parameters from *.mat for the current panel.
    • ?: Help information.
  • Reference:
  1. Zang Y, Jiang T, Lu Y, He Y, Tian L. Regional homogeneity approach to fMRI data analysis. Neuroimage 2004; 22(1): 394-400.

FCD/FCS

FCD is short for Functional Connectivity Density and FCS is short for Functional Connectivity Strength.

A region growing algorithm was carried out to measure the local degree of each voxel under a certain threshold of Pearson’s correlation. FCD in BRANT has been implemented to calculate the local FCD (lFCD), the global FCD (gFCD) and the long-range FCD (lrFCD) at one time. The lFCD of each voxel represents the number of spatially connected voxels defined by the region growing algorithm, while the gFCD, which is also referred to as the voxel-wise degree centrality, represents the number of voxels that have higher-than-threshold correlation with the seed voxel. The lrFCD is calculated as the gFCD subtracted the lFCD.

Functional connectivity strength (FCS) measures the amount of information a node receives across whole graph or within a distance. Similar to FCD, the voxel-wise Pearson’s correlation coefficients are firstly calculated in parallel and then Fisher-z transformed to improve normality. For each voxel, the FCS is calculated as the sum of connectivity that exceeds a given threshold divided by the number of voxels.

_images/FCD_new.png
  • mask: could be whole brain mask or gray matter mask.
  • id index: identifier to find unique string for each subject
  • filetype: files in the filetype will be searched in input directories.
  • 4D nifti files: if the input data is 4D, check this item. Otherwise uncheck.
  • input dirs: directories can be input either using a *.txt file or spm select window.
  • time series: choose the time series to calculate, seperated by ‘,’.
  • compute: use OPENCL supported CPU or GPU to calculate FCD
  • r threshold: threshold of correlation (to binarize functional connectivity and sum up)
  • metrics:
    1. fcd - functional connectivity density, calculate global and region grow defined degree
    2. fcs - functional connectivity strength, calculate global-wise sum/mean of above threshold intensity
    3. fcs abs - absolute functional connectivity strength, firstly convert FC map to absolute value and calculate global-wise sum/mean of above threshold intensity
  • out dir: output directory for saving results.
  • Buttons:
    • S: Save parameters of the current panel to a *.mat file. The *.mat can be further loaded for the panel or be used in a script processing.
    • L: Load parameters from *.mat for the current panel.
    • ?: Help information.
  • Output files:
    • Raw:
      • gfcd(global fcd): count the number of voxels of voxel to whole brain correlation (rho > threshold)
      • lfcd(local fcd): count the number of voxels of voxel to neighbour voxels’ correlation (rho > threshold, with region grow method)
      • lrfcd(long-range fcd): gfcd - lfcd
      • fcs_sum: sum of above threshold voxels’ intensity
      • fcs_ave: mean of above threshold voxels’ intensity
    • Normalized:
      • gfcd: gfcd(Raw) divided by mean value of gfcd(Raw)
      • lfcd: lfcd(Raw) divided by mean value of lfcd(Raw)
      • lrfcd: lrfcd(Raw) divided by mean value of lrfcd(Raw)
      • fcs_sum_nor: fcs_sum(Raw) divided by mean value of fcs_sum(Raw)
      • fcs_ave_nor: fcs_ave(Raw) divided by mean value of fcs_ave(Raw)
  • References:
  1. Tomasi, D., & Volkow, N. D. (2010). Functional connectivity density mapping. Proceedings of the National Academy of Sciences of the United States of America, 107(21), 9885-90.
  2. Craddock R, Clark D. Optimized implementations of voxel-wise degree centrality and local functional connectivity density mapping in AFNI. GigaScience 2016; 5(suppl_1): 4-6.
  3. Qin W, Xuan Y, Liu Y, Jiang T, Yu C. Functional Connectivity Density in Congenitally and Late Blind Subjects. Cereb Cortex 2014; 25(9): 2507-16.

Utilities

We have added several frequently used functions in this module to facilitate DICOM image conversion, the process of quality control, ROI coordinates extraction and 2D/3D signal extraction.

Dicom Convert

Since in practice raw MRI data exported from an MR scanner consists of a large number of DICOM images storing slices and volumes of different sequences, by convention we convert the DICOM images to packed 3D or 4D NIFTI images before all processing steps. In BRANT, we use the dcm2nii from MRIcron/MRIcro to convert DICOM files into 4D NIfTI images by default and use wildcard characters to locate rs-fMRI image files. For the matched images, the First N timepoints removal is used to remove the first N frames that could be influenced by large motion or the instability of magnetic field.

_images/dicom_convert.png
  • operation: convert
    • parallel workers: select how many workers to start a parallel work. The default is 0.
    • convert to 4d: select to convert data to 4D format, otherwise to 3D.
    • id index: identifier to find unique string for each subject.
    • input dirs: directories can be input either using a .txt file or spm select window.
    • delete first N timepoints: delete heading fMRI volumes.
    • N: number of the heading timepoints to be deleted, the default is 10.
    • filetype: wildcard to search wanted fMRI data.
    • out dir: output directory for saving results.
  • operation: delete
    • id index: identifier to find unique string for each subject
    • filetype: files in the filetype will be searched in input directories.
    • 4D nifti files: if the input data is 4D, check this item. Otherwise uncheck.
    • input dirs: directories can be input either using a *.txt file or spm select window.
    • delete first N timepoints: delete heading fMRI volumes.
    • output fn: output filename.
    • output to another directory: output data to another directory
  • Buttons:
    • S: Save parameters of the current panel to a *.mat file. The *.mat can be further loaded for the panel or be used in a script processing.
    • L: Load parameters from *.mat for the current panel.
    • ?: Help information.
  • Reference:
  1. Rorden C, Brett M. Stereotaxic display of brain lesions. Behav Neurol 2000; 12(4): 191-200.

Head Motion Estimate

Head motion has been found having an impact on rs-fMRI signals. In preprocessing, six head motion parameters of (x-,y-,z-) translations and (pitch-,yaw-,roll-) rotations estimated during realignment are used as the inputs of the current function. By the default, the current function outputs the maximum absolute translation and rotation between frames as the exclusion criterion of large-motion subject. Additionally, the mean head displacement (the root-mean-square of translation parameters), the maximum head displacement, the number of micro displacement (>0.1mm), the mean absolute Euler angle of rotation, the framewise displacement (FD) and the number of frames with FD>0.5mm, are also exported to provide more subject exclusion criteria.

_images/head_motion_est.png
  • id index: identifier to find unique string for each subject
  • filetype: files in the filetype will be searched in input directories.
  • input dirs: directories can be input either using a .txt file or spm select window.
  • out dir: output directory for saving results.
  • Meaning of results:
    • max-abstranslation(mm): maximum translation. Estimated as the max absolute value of the first 3 columns from rp*.txt
    • max-absrotation(deg): maximum rotation. Estimated as the max absolute value of the last 3 columns from rp*.txt multiple by 180/pi.
    • From van Dijk et al., Neuroimage 2012
      • max-motion-Dijk(mm): maximum root-mean-square of translation.
      • mean-motion-Dijk(mm): mean root-mean-square of translation.
      • num-movements-Dijk(>0.1mm): number of micro-movement. The number of root-mean-square of translation that is greater than 0.1mm
      • mean-rotation-Dijk(deg): mean absolute Euler angle
    • From Power et al., Neuroimage 2012
      • mean-FD(mm): frame-wise displacement. Estimated using translation and rotation.
      • num-FD>0.5: number of frame-wise displacement that is greater than 0.5mm.
_images/head_motion_est_result.png
  • Buttons:
    • S: Save parameters of the current panel to a *.mat file. The *.mat can be further loaded for the panel or be used in a script processing.
    • L: Load parameters from *.mat for the current panel.
    • ?: Help information.
  • References:
  1. Van Dijk KR, Sabuncu MR, Buckner RL. The influence of head motion on intrinsic functional connectivity MRI. Neuroimage 2012; 59(1): 431-8.
  2. Power JD, Barnes KA, Snyder AZ, Schlaggar BL, Petersen SE. Spurious but systematic correlations in functional connectivity MRI networks arise from subject motion. Neuroimage 2012; 59(3): 2142-54.

Visual Check

The current function provides batch operations to visually inspect artifacts and normalization quality, by calling Display from SPM. We’ve added keyboard operations to the Display figure that users can press up/down to switch fMRI volumes of one subject and press left/right to switch subjects. Before running the frame-by-frame inspection, the current function exports screenshots of selected slices overlaid by a semi-transparent brain mask for a glimpse of the overall image quality.

_images/visual_check.png

The function calls SPM Check Reg to visualize input volume, a keyboard callback is added to enable left/right and up/down to switch timepoint and subjects.

  • mask: could be whole brain mask or gray matter mask.
  • id index: identifier to find unique string for each subject
  • filetype: files in the filetype will be searched in input directories.
  • 4D nifti files: if the input data is 4D, check this item. Otherwise uncheck.
  • input dirs: directories can be input either using a *.txt file or spm select window.
  • display orthogonal view: check to display orthogonal view of selected slices and save screenshots. uncheck will only save screenshots to out dir.
  • mask color: color of the transparent overlaid mask.
  • slices: modified image n-th slice of x,y,z to save.
  • out dir: output directory for saving results.
  • Keyboard Operation:
    • Up/Down: last/next timepoint of the same subject.
    • Left/Right: same timepoint of last/next subject.
  • Buttons:
    • S: Save parameters of the current panel to a *.mat file. The *.mat can be further loaded for the panel or be used in a script processing.
    • L: Load parameters from *.mat for the current panel.
    • ?: Help information.

TSNR (Temporal Signal to Noise Ratio)

Influenced by the magnetic field inhomogeneity at air-tissue interfaces, rs-fMRI signals at orbitofrontal and temporal medial and polar areas suffer from a certain degree of distortions and signal loss. To exclude spurious voxels, we use the thresholded voxel-wise TSNR, which is calculated as the average intensity of time series divided ty the standard deviation, to generate subject-level or group-leval whole-brain mask.

_images/tsnr_new.png
  • mask: could be whole brain mask or gray matter mask.
  • id index: identifier to find unique string for each subject
  • filetype: files in the filetype will be searched in input directories.
  • 4D nifti files: if the input data is 4D, check this item. Otherwise uncheck.
  • input dirs: directories can be input either using a *.txt file or spm select window.
  • threshold: intensity threshold for mean TSNR (to generate a binary mask).
  • out dir: output directory for saving results.
  • Buttons:
    • S: Save parameters of the current panel to a *.mat file. The *.mat can be further loaded for the panel or be used in a script processing.
    • L: Load parameters from *.mat for the current panel.
    • ?: Help information.
  • References:
  1. Tomasi D, Volkow ND. Functional connectivity density mapping. Proceedings of the National Academy of Sciences of the United States of America 2010; 107(21): 9885-90.
  2. Yeo BT, Krienen FM, Sepulcre J, Sabuncu MR, Lashkari D, Hollinshead M, et al. The organization of the human cerebral cortex estimated by intrinsic functional connectivity. J Neurophysiol 2011; 106(3): 1125-65.
  3. Welvaert M, Rosseel Y. On the definition of signal-to-noise ratio and contrast-to-noise ratio for FMRI data. PLoS One 2013; 8(11): e77089.

ROI coordinates

To visualize the topological strcture of network connections, ROI coordinates are expected as the centers of spheres. In the current function, coordinate of each number-tagged ROI is calculated as the center of mass with weights and then exported to a *.csv table.

_images/roi_coordinates.png

The current function extracts coordinates for each cluster and output to a table.

  • input type: seperated binary clusters or labeled clusters
  • mask*: optional. mask to do AND operation with.
  • cluster size: threshold of cluster size.
  • roi file: input ROI file
  • out dir: output directory for saving results.
  • Buttons:
    • S: Save parameters of the current panel to a *.mat file. The *.mat can be further loaded for the panel or be used in a script processing.
    • L: Load parameters from *.mat for the current panel.
    • ?: Help information.

Extract Value

_images/extract_value.png
  • data: matrix
    • symmetric matrix: check to extract the upper right matrix’s elements, uncheck to extract all elements
    • corr mask: e.g. a matrix mask used to find significant links instead of all links
    • roi info*: a .csv with at least one column start with “label”
    • id index: identifier to find unique string for each subject
    • filetype: filetype
    • data dir: Input directories of matrices.
  • data: volume
    • roi file: ROI file used for extracting mean intensity in each roi tagged by number.
    • roi info: labels of tagged ROIs. (optional)
    • mask: could be whole brain mask or gray matter mask.
    • id index: identifier to find unique string for each subject
    • filetype: files in the filetype will be searched in input directories.
    • 4D nifti files: if the input data is 4D, check this item. Otherwise uncheck.
    • input dirs: directories can be input either using a .txt file or spm select window.

  • string removal*: optional. remove partial string from string parsed by id index.

  • output as: (only works for data type ‘matrix’) choose to output data matrix as ASCII(can be edited with text reader) or binary(more hard drive friendly with large matrix). if binary is chosen, you need to handle the matrix by:

    fid = fopen('brant_extract_links.txt', 'rt');
outmat = fread(fid, sizeofmat, 'single');
fclose(fid);

  • out dir: output directory for saving results.

  • Buttons:
    • S: Save parameters of the current panel to a *.mat file. The *.mat can be further loaded for the panel or be used in a script processing.
    • L: Load parameters from *.mat for the current panel.
    • ?: Help information.

Reslice

_images/reslice.png

The current function reslice MRI images to a reference image (internally calls SPM-Coregister-Reslice).

  • reference: reference image for header information.
  • id index: identifier to find unique string for each subject
  • filetype: files in the filetype will be searched in input directories.
  • 4D nifti files: if the input data is 4D, check this item. Otherwise uncheck.
  • input dirs: directories can be input either using a *.txt file or spm select window.
  • output prefix: prefix of output file, by default the program will output to the same directory as input file.
  • output to another directory: check to output file to another directory.
  • out dir: output directory for saving results.
  • Buttons:
    • S: Save parameters of the current panel to a *.mat file. The *.mat can be further loaded for the panel or be used in a script processing.
    • L: Load parameters from *.mat for the current panel.
    • ?: Help information.

NET

Network metrics depict the properties of information flow among predefined nodes. Regarding brain networks, ROIs are defined as nodes and the connections between pairs of ROIs are defined as edges. In the current module, connectivity matrices are firstly thresholded by intensity or sparsity to weighted or binary networks.For group comparisons under a vector of thresholds, Student’s t-tests are provided.

Threshold Estimation

Find out for each subject, the sparsity under each threshold of correlation

_images/threshold_estimation.png
  • filetype: files in the filetype will be searched in input directories.
  • data dir: directory where all *.txt correlation matrix results are stored.
  • use absolute value of input matrics: as it means
  • intensity threshold: vector of thresholds for matrix intensity. e.g. correlation coefficient
  • out dir: output directory for saving results.
  • Buttons:
    • S: Save parameters of the current panel to a *.mat file. The *.mat can be further loaded for the panel or be used in a script processing.
    • L: Load parameters from *.mat for the current panel.
    • ?: Help information.

Network Calculation

_images/network_calculation.png
  • parallel workers: number of workers used for parallel computing.
  • filetype: files in the filetype will be searched in input directories.
  • data dir: directory where all *.txt correlation matrix results are stored.
  • use absolute value of input matrics:
    • raw value: use raw value to construct binary matrix.
    • absolute value: use absolute value to construct binary matrix.
  • intensity threshold: vector of thresholds for matrix intensity. e.g. correlation coefficient
  • sparsity threshold: a vector of sparsity threshold, for each element, threshold the input matrix using the fraction of the matrix’s largest number of connection n*(n+1)/2;
  • minimum spanning tree: a process to avoid unconnected network. To label the backbone of the network’s nodes.
  • matrix type: binarized and weighted network. The binarized networks comes from thresholded input matrics, while the weighted network comes from a dot product operation of binarized network and the original network.
  • Network properies: a panel to select network properties. In the option panel, (*) means the calculation of the property is slow.
  • out dir: output directory for saving results.
  • zero value in clustering coefficient: if the network is connected in a way (e.g. Hamilton path) that neighbour nodes are not connected; or the current node has only one neighbour node.
  • Inf in small worldness: The small-worldness is calculated as real-network / random-network, and in real and random case, small-worldness is calculated as clustering coefficient/shortest-path length. If the mean clustering coefficient of the random-network is zero, then the small-worldness of random-network is zero, and it will cause the real-network / random-network to be Inf(any non-zero divided by 0). The solution is set smaller thresholds that avaliable for all subjects.
  • Buttons:
    • S: Save parameters of the current panel to a *.mat file. The *.mat can be further loaded for the panel or be used in a script processing.
    • L: Load parameters from *.mat for the current panel.
    • ?: Help information.
  • Reference:
  1. Rubinov M, Sporns O. Complex network measures of brain connectivity: uses and interpretations. Neuroimage 2010; 52(3): 1059-69.

Network Statistics

_images/network_statistics.png

  • filetype: files in the filetype will be searched in input directories.

  • data dir: directory where *.mat result of NETWORK CALCULATION is stored.

  • stat type: one sample t-test, two sample t-test and paired t-test.

  • groups:
    • for two sample t-test, e.g. SZ,NC -> will do two-sample t-test for SZ and NC
    • for paired t-test, e.g. SZ,NC -> will do paired t-test for SZ and NC
    • for one sample t-test, e.g. SZ;NC -> will do one-sample t-test for both SZ and NC group
  • grouping info:

    • table: A comma-seperated values (csv) table, which is used for parsing subject names and covariates. The parsed names/ids will be matched to search results conducted with datadir and filetype. Before matching, BRANT will remove specified strings.

    • For one and two sample t-tests:
      name group filter age
      subj1 SZ center1 28
      subj2 SZ center1 27
      subj3 NC center1 30
      subj4 NC center2 25
    • For paired t-test, another column of paired_t_idx is required to specify paired subjects in each group:
      name group filter age paired_t_idx
      subj1 stage1 center1 28 1
      subj2 stage1 center1 27 2
      subj3 stage2 center1 30 1
      subj4 stage2 center2 25 2

  • string removal*: optional. remove partial string from string parsed by id index.

  • regressors*: optional. title of regressors which will be regressed out before statistical analysis. e.g. age

  • filter*: optional. use the control for subject in different state or center, fill in center1 here and subject4 won’t be included in the analysis.

  • discard subjects without info: when checked, if subjects’ information are not found in the table, a warning message will be shown; when unchecked, an error message will be shown.

  • out dir: output directory for saving results.

  • If the output figure is empty, check whether there is NaN or Inf in the output *.csv files, where the raw global network properties are extracted.

  • Buttons:
    • S: Save parameters of the current panel to a *.mat file. The *.mat can be further loaded for the panel or be used in a script processing.
    • L: Load parameters from *.mat for the current panel.
    • ?: Help information.

STAT

The current module provides Student’s t-tests for sample mean comparisons and several methods for image-based meta-analysis (IBMA). For multi-comparison correction, we use the Benjamini-Hochberg and the Benjamini-Yekutieli procedures to control the false discovery rate (FDR) of dependent and independent cases, and the Bonferroni procedure to control the familywise error rate (FWER).

T-tests

_images/t_tests.png
  • data: matrix
    • symmetric matrix: check to extract the upper right matrix’s elements, uncheck to extract all elements
  • data: volume
    • mask: could be whole brain mask or gray matter mask.

  • filetype: filetype

  • data dir: Input directories of matrices.

  • grouping info:

    • stat type: one sample t-test, two sample t-test and paired t-test.

    • groups:
      • for two sample t-test, e.g. SZ,NC -> will do two-sample t-test for SZ and NC
      • for paired t-test, e.g. SZ,NC -> will do paired t-test for SZ and NC
      • for one sample t-test, e.g. SZ;NC -> will do one-sample t-test for both SZ and NC group

    • table: A comma-seperated values (csv) table, which is used for parsing subject names and covariates. The parsed names/ids will be matched to search results conducted with datadir and filetype. Before matching, BRANT will remove specified strings.

    • For one and two sample t-tests:

      name group filter age
      subj1 SZ center1 28
      subj2 SZ center1 27
      subj3 NC center1 30
      subj4 NC center2 25

    • For paired t-test, another column of paired_t_idx is required to specify paired subjects in each group:

      name group filter age paired_t_idx
      subj1 stage1 center1 28 1
      subj2 stage1 center1 27 2
      subj3 stage2 center1 30 1
      subj4 stage2 center2 25 2

    • string removal*: optional. remove strings from search results parsed by id index.

    • regressors*: optional. title of regressors which will be regressed out before statistical analysis. e.g. age

    • filter*: optional. use the control for subject in different state or center, fill in center1 here and subject4 won’t be included in the analysis.

    • discard subjects without info: when checked, if subjects’ information are not found in the table, a warning message will be shown; when unchecked, an error message will be shown.

  • Multiple comparison correction methods (voxel-wise)
    • threshold: the level of MULCC
    • fdrID: false discovery rate (independent input)
    • fdrN: false discovery rate (inputs not independent)
    • bonf: Bonferroni correction for family wise error rate

  • out dir: output directory for saving results.

  • Buttons:
    • S: Save parameters of the current panel to a *.mat file. The *.mat can be further loaded for the panel or be used in a script processing.
    • L: Load parameters from *.mat for the current panel.
    • ?: Help information.

IBMA (Image-based meta-analysis)

With statistical maps of different datasets tested using same analysis pipeline, and the demography of each sample, users can perform meta-analysis to merge the multisite statistics using image-based or matrix-based meta-analysis. We have implemented Stouffer’s z-score method, Fisher’s method, fixed/mixed effects model, Worsley and Friston’s method and Nichols’ method.

_images/ibma.png
  • data: matrix
    • filetype: files in the filetype will be searched in input directories.
    • data dir: directory where all ttest2*.mat results are stored.
  • data: volume
    • center info: number of subjects for different centers. A csv format table is required. N1 and N2 is the number of subjects in each group.
      center N1 N2
      ttest2_center1_a_vs_b 40 39
      ttest2_center2_a_vs_b 38 37

    • mask: could be whole brain mask or gray matter mask.

    • id index: identifier to find unique string for each subject

    • filetype: files in the filetype will be searched in input directories.

    • data dir: directories can be input either using a *.txt file or spm select window.

  • Multiple comparison correction methods (voxel-wise)
    • threshold: the level of MULCC
    • fdrID: false discovery rate (independent input)
    • fdrN: false discovery rate (inputs not independent)
    • bonf: Bonferroni correction for family wise error rate
  • IBMA Methods:
  • out dir: output directory for saving results.
  • Buttons:
    • S: Save parameters of the current panel to a *.mat file. The *.mat can be further loaded for the panel or be used in a script processing.
    • L: Load parameters from *.mat for the current panel.
    • ?: Help information.
  • References:
  1. Stouffer’s z-score

    Stouffer, S.A., Suchman, E.A., DeVinney, L.C., Star, S.A. and Williams Jr, R.M., 1949. The American soldier: Adjustment during army life.(Studies in social psychology in World War II), Vol. 1. Princeton University Press, Princeton,.

  2. Fisher

    Fisher, R.A. (1925). Statistical Methods for Research Workers. Oliver and Boyd (Edinburgh). ISBN 0-05-002170-2.

  3. Fixed/mixed Effects Model

    Hedges, L.V. (1992). Meta-Analysis. Journal of Educational and Behavioral Statistics. 17(4), 279-296. doi: 10.3102/10769986017004279.

    Konstantopoulos, S. (2006). Fixed and mixed effects models in meta-analysis. Iza Discussion Papers.

  4. Worsley and Friston’s method

    Worsley, K.J., and Friston, K.J. (2000). A test for a conjunction. Statistics & Probability Letters. 47(2), 135-140. doi: 10.1016/S0167-7152(99)00149-2.

  5. Nichols’s method

    Nichols, T., Brett, M., Andersson, J., Wager, T., and Poline, J.B. (2005). Valid conjunction inference with the minimum statistic. Neuroimage. 25(3), 653-660. doi: 10.1016/j.neuroimage.2004.12.005.

  6. Salimi-Khorshidi G, Smith SM, Keltner JR, Wager TD, Nichols TE. Meta-analysis of neuroimaging data: a comparison of image-based and coordinate-based pooling of studies. Neuroimage 2009; 45(3): 810-23.

  7. Benjamini Y, Hochberg Y. Controlling the False Discovery Rate - a Practical and Powerful Approach to Multiple Testing. J Roy Stat Soc B Met 1995; 57(1): 289-300.

  8. Benjamini Y, Yekutieli D. The control of the false discovery rate in multiple testing under dependency. Ann Stat 2001; 29(4): 1165-88.

  9. Lazar NA, Luna B, Sweeney JA, Eddy WF. Combining brains: a survey of methods for statistical pooling of information. Neuroimage 2002; 16(2): 538-50.

Third Party

DiffusionKit

A batch processing GUI-interface for DiffusionKit

_images/diffusionkit.png
  • D.K. dir: path of installed DiffusionKit, e.g. C:/Program Files (x86)/DiffusionKit
  • id index: identifier to find unique string for each subject
  • filetype: files in the filetype will be searched in input directories.
  • 4D nifti files: if the input data is 4D, check this item. Otherwise uncheck.
  • input dirs: directories can be input either using a *.txt file or spm select window.
  • Choose below options to perform in each directory.
    • eddy correction
    • reconstruction–DTI
    • tracking–DTI
    • reconstruction–HARDI(ODF)
    • reconstruction–HARDI(FOD)

Results will be saved in each input directory.

  • Buttons:
    • S: Save parameters of the current panel to a *.mat file. The *.mat can be further loaded for the panel or be used in a script processing.
    • L: Load parameters from *.mat for the current panel.
    • ?: Help information.

For more information, please visit http://diffusion.brainnetome.org/

  • Reference:
  1. Sangma Xie, Liangfu Chen, Nianming Zuo and Tianzi Jiang. DiffusionKit: A Light One-Stop Solution for Diffusion MRI Data Analysis. Journal of Neuroscience Methods. vol. 273, pp. 107-119, 2016.

Circos

_images/circos.png
  • circos dir: Path of unzipped circos folder. /circos_path/bin
  • conf dir: Path of circos configure file.
  • roi info: CSV table which defines sub-areas and lobes, there is an example in brant-master/circos/brant_circos_3mm_273.csv
  • There should be at least four columns in the table:
    • label: label of each sub-areas.
    • module: to which lobe/module does the sub-area belong.
    • index_module: order of the arranged module.
    • index_node: order of sub-areas within one module.

Before using the function, please download and install circos 0.69 or higher.

  • Buttons:
    • S: Save parameters of the current panel to a *.mat file. The *.mat can be further loaded for the panel or be used in a script processing.
    • L: Load parameters from *.mat for the current panel.
    • ?: Help information.
  • References:
  1. Krzywinski M, Schein J, Birol I, Connors J, Gascoyne R, Horsman D, et al. Circos: an information aesthetic for comparative genomics. Genome Res 2009; 19(9): 1639-45.
  2. Irimia A, Chambers MC, Torgerson CM, Van Horn JD. Circular representation of human cortical networks for subject and population-level connectomic visualization. Neuroimage 2012; 60(2): 1340-51.
  3. Fan L, Li H, Zhuo J, Zhang Y, Wang J, Chen L, et al. The Human Brainnetome Atlas: A New Brain Atlas Based on Connectional Architecture. Cereb Cortex 2016; 26(8): 3508-26.

VIEW

To visualize voxel intensities, we have implemented the ROI Mapping to extract and render the surface of 3D clusters, the Surface Mapping to project voxel intensity to vertices on a surface. To visualize ROI-ROI connectivity, Network Visualization is implemented to draw spheres and rods within a rendered brain surface, to present nodes and edges of the input network.

Surface Mapping

Besides shading each ROI/cluster, we can also project the voxel intensities to the surface. By the default, we use a rendered human brain surface constructed from vertices and triangular faces loaded from a pregenerated file. To draw another surface, users can input a binarized 3D mask, with which BRANT can extract the generate vertices and faces and render a new surface. When projecting a 3D volume to surface, the vertices on the surface are shaded as the intensity of the nearest voxel, while the material of the surface, the color maps of positive and negative intensities, the lighting and shading algorithm can be adjusted.

_images/surface_mapping.png
  • show colorbar: display colorbar
  • discrete value: the intensity of the input volume has float or integer datatype
  • alpha: degree of opaque
  • max val radius(mm): radius for maximum neighbour interpolation. if the radius is greater than the size of a voxel, the program will search for maximum value within a sphere for each vertex, otherwise (leave empty or smaller than the size of a voxel) use the default 1-voxel interpolation.
  • display: mode of display
  • material: sets the lighting characteristics of surface and patch objects
    • shiny: sets the reflectance properties so that the object has a high specular reflectance relative to the diffuse and ambient light, and the color of the specular light depends only on the color of the light source
    • dull: sets the reflectance properties so that the object reflects more diffuse light and has no specular highlights, but the color of the reflected light depends only on the light source
    • metal: sets the reflectance properties so that the object has a very high specular reflectance, very low ambient and diffuse reflectance, and the color of the reflected light depends on both the color of the light source and the color of the object
  • lighting: selects the algorithm used to calculate the effects of light objects on all surface and patch objects in the current axes
    • flat: produces uniform lighting across each of the faces of the object. Select this method to view faceted objects
    • gouraud: calculates the vertex normals and interpolates linearly across the faces. Select this method to view curved surfaces
    • phong: sets the lighting to phong
    • none: turns off lighting
  • shading: controls the color shading of surface and patch graphics objects
    • flat: each mesh line segment and face has a constant color determined by the color value at the endpoint of the segment or the corner of the face that has the smallest index or indices
    • faceted: flat shading with superimposed black mesh lines
    • interp: varies the color in each line segment and face by interpolating the colormap index or true color value across the line or face
  • colormap: controls the colors used in displaying the surface
  • surface: surface file
  • brain vol: volume to map to the surface
  • thr vol: set the range to generate a mask for input volume, seperated by ‘,’.

e.g:

thr vol: 1, 5, 9, 15
  • Buttons:
    • S: Save parameters of the current panel to a *.mat file. The *.mat can be further loaded for the panel or be used in a script processing.
    • L: Load parameters from *.mat for the current panel.
    • ?: Help information.
_images/surface_mapping_result.png

Slice Mapping

This function is used to visualize 3D volume by cut it into no more than 19 slices. All slices will be displayed in a single window, overlapped with the background volume. We can choose the range of displaying if necessary.

_images/slice_mapping.png
  • view angle: direction of slice viewing
  • slice order: choose which slice(s) to display
  • bg: directory of backgroud volume file
  • brain vol: directory of displayed volume file
  • white background: the background of the display window will be white if checked, otherwise it will be black.
  • only positive: only voxels with positive value will be displayed if checked
  • only negative: only voxels with negetive value will be displayed if checked
  • expand display range: if checked, the voxels with value out of range will be displayed as the threshold value
  • colormap: choose the color map of slice viewing
  • thr vol: set the range to generate a mask for input volume, seperated by ‘,’.
  • col title: title of the colorbar
  • Buttons:
    • S: Save parameters of the current panel to a *.mat file. The *.mat can be further loaded for the panel or be used in a script processing.
    • L: Load parameters from *.mat for the current panel.
    • ?: Help information.

ROI Mapping

When visualizing ROIs from an atlas or clusters from a user-defined 3D volume (e.g., clusters with significant difference between sample means), we can use the current function to extract and shade the surface of each number-tagged ROI/cluster in random or user defined colors. The ROIs/clusters of the input 3D image should be tagged with positive-integers. With an additional input of a reference *.csv table containing number-label pairs (as described in Utilities -> DICOM Convert), we can further parse the labels of each shaded ROI/cluster and present them in a legend.

_images/roi_mapping.png

  • alpha: degree of transparency.

  • display: mode of display.

  • display surface: show surface.

  • surface: surface file.

  • display legend: display legend.

  • roi file: extract mean intensity in the roi tagged by numbers.

  • roi info*: optional. two columns of information for each labeled cluster in a *.csv file. For example:

    1,SFG
2,MFG
3,IFG

  • roi vals: select which roi to display.

  • color: optional. use random color or input color file.

  • color file: the input color could be (ROI tag, R, G, B):

    1,255,155,100
2,1,1,1

  • output color: output color of current image.

  • out dir: output directory for saving results.

  • Buttons:
    • S: Save parameters of the current panel to a *.mat file. The *.mat can be further loaded for the panel or be used in a script processing.
    • L: Load parameters from *.mat for the current panel.
    • ?: Help information.
_images/roi_mapping_result.png

Network Visualization

Using a *.txt file storing symmetric connectivity matrix and a *.csv table with nodal information (such as coordinate, label, module and color) as input, we can draw spheres and rods to visualize nodes and edges.

_images/network_visualization.png
  • surface: surface file
  • alpha: degree of opaque
  • display: mode of display
  • node: node file defined as csv table. All columns are optional except for ‘x’, ‘y’ and ‘z’.

For example:

 x     y    z  size  module   r   g   b   label
 -1 , 20 , 20 , 4 , module1 , 5 , 5 , 5 , node1
-10 , 22 , 20 , 4 , module1 , 5 , 5 , 5 , node2
 12 , 20 , 20 , 4 , module2 , 5 , 5 , 5 , node3
  • show node labels: check to show labels defined in input node file.
  • same size: use same size for all node, uncheck to use user defined size in input node file.
  • user defined node color: use color defined in input node file.
  • same node color: use same color for all node.
  • module color: use different color for each module. Modules are defined in input node file.
  • edge: edge matrix for input file, the number of rows and columns should be the same as input file.
  • display edges: display or not edges.
  • hide node without edge: select not to show nodes without edge
  • thickness: relative thickness for all edges
  • adjust edge color: use different color for positive and negative edge.
  • threshold: an expression that compatible with matlab syntax to filter out unwanted edges in edge matrix.
  • use summed weight as node radius: sum up node’s degree and define node size.
  • threshold: nodes with degree smaller than the threshold will not be shown.
  • Buttons:
    • S: Save parameters of the current panel to a *.mat file. The *.mat can be further loaded for the panel or be used in a script processing.
    • L: Load parameters from *.mat for the current panel.
    • ?: Help information.

Examples

STEP 1: DICOM Convert

  • Open Directories Selection GUI:
    Click the Utilities, DICOM Convert and then select the input file by clicking the … icon.
  • Input Directories of DICOM Files:
    Directories can be navigated by operating folders in the left side panel or directories on the top. Folders matching the filter (Filt) are shown in the panel on the right and can be selected by clicking or dragging. Use the right mouse button if you would like to select all files. The panel at the bottom shows files that are already selected. Clicking a selected file will un-select it.
  • Usage of ID Index:
    Using id index properly can help BRANT to find unique string for each subject. When id index is 1,it means the data folder contains subject string itself. 2 means subject information can be found in data folder’s parent upper one. e.g. if your data are stored in G:/TestData/1_NC001, set id index to 1, if your data are stored in G:/TestData/1_NC001/fMRI, set id index to 2. Both output files will be put in the folder out dir/1_NC001.
_images/step1_dicom_convert.png

fig.1 Utilities => DICOM Convert

_images/step1_dicom_convert_input.png

fig.2 Selection GUI

STEP 2: Preprocessing

  • Input Directories for Preprocessing:
    Click the Preprocessing button. Select the folders as data dirs where STEP 1 outputs. You can check the from text file and select a brant_preprocess_paths.txt file which has already automatically created.
  • Preprocessing Settings:
    Check the checkboxes you need, remember to modify parameters. You donnot need to check the Coregister checkbox without structural images. Remember to change the source parameter in Normalise to co*.nii if you’re preprocessing images with structural images.

Note

If an error occours during preprocessing, the R (refresh) button can recover the run button.

  • Further information about preprocessing can be found in:

Slice Timing

  • slice order: the order of scans in one volume, seperated by comma or space.
  • TR(s): repetition time, cannot be 0.
  • reference slice: normally be the number of scan in the middle of the order.(when dealing with task-fMRI, note that selecting the middle timepoint as reference will change the timing of task TR)
  • prefix: output prefix.
_images/slicetiming.png

e.g:

slice order: 1:2:33,2:2:32
TR: 2
reference slice: 33

Realign

_images/realign.png
  • Estimate: Estimate parameters for realignment.
    • quality: Highest quality (1) gives most precise results, whereas lower qualities gives faster realignment.
    • sep: The separation (in mm) between the points sampled in the reference image. Smaller sampling distances gives more accurate results, but will be slower.
    • fwhm: Full width at half maximum of the Gauss smoothing kernel (mm).
    • rtm: 1 indicates images are registered to the mean image, while 0 indicates images are registered to the first image in each subject’s folder.
    • wrap: 3 dimensions of wrapping, e.g. [1 1 1] for wrapping in X, Y and Z direction, [0 0 0] for no wrapping.
    • interp: Indexes of interpolation methods. (1 for Trilinear; 2-7 for 2nd-7th Degree B-Spline).
  • Write:
    • which: The first parameter allows 0, 1 and 2 as input (0: create only a mean resliced image; 1: don’t reslice the first image. 2: reslice all the image). The second parameter indicates whether to output a resliced mean image (0 for false and nonzero for true).
    • interp: Interpolation methods for write option. (0 for Nearest Neighbor; 1 for Trilinear; 2-7 for 2nd-7th Degree B-Spline; Inf for Fourier Interpolation).
    • mask: Mask output images (true/false). If any image has a zero value at a voxel, then all images will have zero (or NaNs if allowed) values at that voxel.
    • wrap: 3 dimensions of wrapping, e.g. [1 1 1] for wrapping in X, Y and Z direction, [0 0 0] for no wrapping.
    • prefix: Output images will have a prefix of r by default.
  • Reference: spm manual.

Coregister

  • Subject info:

    • reference: Filetype of reference image stored in each subject’s folder to register.

    • source: Filetype of image to match the reference image stored in each subject’s folder.

    • seg & bet: Segment and skull stripe structural T1/T2 image. Using skull striped T1/T2 image for coregistration is recommended.

      • options:

        1: segment using new segment and bet based on tissue probability maps;

        2: bet only (there should be segmented c1-c3*.nii files in the directory);

        other number: do not segment nor bet; we recommend using co*.nii instead of bet*.nii to normalise.

_images/coregister.png
  • Estimate:
    • object fun: Methods to maximise or minimise objective function.
    • sep: The average distance between sampled points (in mm).
    • tol: The accuacy for each paramters.
    • fwhm: Kernel of gaussian smooth to apply to the 256*256 joint histogram.
  • Reference: spm manual.

Normalise SPM12

  • Subject info:
    • source: Filetype of images for normalization are stored. Default is the mean*.nii generated from realign. Users can also change it to T1/T2 structural image of each subject stored in each subject’s folder. If Coregister is checked, remember to add prefix of Coregister to the source filetype and change to template to the same modality of source.
_images/normalise_spm12.png
  • Estimate:
    • biasreg: bias regularisation.
    • biasfwhm: FWHM of Gaussian smoothness of bias.
    • tpm: Tissue probability map which the source image will be registered to.
    • affreg: affine regularisation.
    • reg: the amount of regularization for the nonlinear part of the spatial normalization.
    • fwhm: option for smoothness.
    • samp: sampling distance.
  • Write:
    • bb: Bounding box of the volume.
    • vox: The voxel sizes of the normalized images.
    • interp: Interpolation methods for write option. (0 for Nearest Neighbor; 1 for Trilinear; 2-7 for 2nd-7th Degree
    • prefix: Output images will have a prefix of w by default.
  • Reference: spm manual.

Normalize

  • Subject info:
    • source: Filetype of images for normalization are stored. Default is the mean*.nii generated from realign. Users can also change it to T1/T2 structural image of each subject stored in each subject’s folder. If Coregister is checked, remember to add prefix of Coregister to the source filetype and change to template to the same modality of source.
    • weight: weighting image of the template.
_images/normalise.png
  • Estimate:
    • template: A standard template image which the source image will be registered to.
    • smosrc: Smoothing to be applied to the copy of the source image. (Source image and the template should have the same smoothness)
    • smoref: Smoothing to be applied to the copy of the source image. (The default templates of spm already have been smoothed by 8mm)
    • regtype: mni (affine registration into MNI space), subj (Registering to an image that has an almost same size of the source image.) and none (No registration)
    • cutoff: Cutoff of DCT bases.
    • nits: Number of nonlinear wrapping iterations.
    • reg: The amount of regularization for the nonlinear part of the spatial normalization.
  • Write:
    • preserve: 0 (The warped images preserve the intensities of the original images) and 1 (Spatially normalised images are “modulated” in order to preserve the total amount of signal in the images.)
    • bb: Bounding box of the volume.
    • vox: The voxel sizes of the normalized images.
    • interp: Interpolation methods for write option. (0 for Nearest Neighbor; 1 for Trilinear; 2-7 for 2nd-7th Degree B-Spline; Inf for Fourier Interpolation).
    • wrap: 3 dimensions of wrapping, e.g. [1 1 1] for wrapping in X, Y and Z direction, [0 0 0] for no wrapping.
    • prefix: Output images will have a prefix of w by default.
  • Reference: spm manual.

Denoise

_images/denoise_1.png
  • Masks & Motion

    Input mask files for whole brain, global mean signal, white matter signal and CSF signal;

    If common space is selected, one mask file for each mask type should be input.

    If indivisual space is selected, wildcard for individual mask should be input, BRANT will search the mask within each subject’s directory.

    The threshold for common space mask is 0.5 by default, while for individual space the threshold can be altered.

    (common space masks normally are binarized, however in individual space are probability)

    • reslice masks with: if the header information (size, FOV, originator, orientation and etc.) is different between data and mask, BRANT will: in common space reslice masks to the first input data

      in individual space reslice masks to the each subject’s input data accordingly

      If the masks are stored as binarized value, the suggested method for reslice is nearest neighbour, otherwise 4th degree B-spline.

    • motion filetype: BRANT will search estimated headmotion file in each subject’s folder, normally by spm the file is rp*.txt

_images/denoise_2.png
  • Regerssion Model
    • linear trend: regressor of 1:T
    • quadratic trend: regressor of [1,22,32:T2]
    • T: a gross regressors for selected global signal, white matter signal and CSF signal
    • T2: element-wise square of T
    • T’: temporal derivatives of T, zero padded
    • T’2: element-wise square of T’
    • Tt-1: 1-frame lagged T, zero padded
    • Tt-1 2: element-wise square of Tt-1
    • R: a gross matrix for motion, should be 6 columns, loaded from rp*.txt
    • R2: element-wise square of R
    • R’: temporal derivatives of R, zero padded
    • R’2: element-wise square of R’
    • Rt-1: 1-frame lagged R, zero padded
    • Rt-1 2: element-wise square of Rt-1
    • Spike Handling:

    As suggested in 1. , spikes are censored with a threshold of FD.

    FD is calculated as:

    motion_diff = diff(R);
FD = [0; sum([abs(motion_diff(:, 1:3)), 50 * abs(motion_diff(:, 4:6))], 2)];

    The scrubbing will take place before the regression model. For example, if order of denoise is selected as Filter first and Regression, the scrubbing will not affect Filter.

_images/denoise_3.png
  • Filter:
    • tr: repetition time.
    • lower cutoff (Hz): lower cutoff for band pass filter.
    • upper cutoff (Hz): upper cutoff for band pass filter.
  • Output Options:
    • Regression + Filter: do regression first and then filter
    • Filter + Regression: do filter first and then regression
    • Regression only: do only regression
    • Filter only: do only filter
    • Run with and without GSR: if checked and mask of global signal is specified, BRANT will run selected process twice with and without global signal as regressor in T, and output to different file in the current and following processes.
    • Save only last results: if checked, no middle files will be saved.
    • Output to *.gz format: check to output to *.gz files (note that the following smooth process from SPM would require uncompressed files, if it’s checked, an error will occur in smooth); choolse if smooth is done beforehand or not required.
    • multi-regression prefix: prefix for output of GLM
    • filter prefix: prefix for output of filter
  • References:
  1. Power, J. D., Mitra, A., Laumann, T. O., Snyder, A. Z., Schlaggar, B. L., & Petersen, S. E. (2014). Methods to detect, characterize, and remove motion artifact in resting state fMRI. Neuroimage, 84(1), 320-341.

Smooth

_images/smooth.png
  • fwhm: full width half maximum for gauss kernel.
  • prefix: output prefix.
  • Further information about output files can be found in the Filename part.
_images/step2_preprocessing.png

fig.1 Preprocessing

STEP 3: Visual Check and Head Motion Estimation

  • Input Directories for Visual Check:
    Click the Utilities button, then Visual Check. Select the folders as input directories where STEP 2 outputs.
  • Visual Check:
    Check the output. Arrow keys can switch the images among different timepoints or subjects during Visual Check. The images will be saved automatically.
  • Head Motion Estimation:
    Click the Utilities button, then Head Motion Est. Select the folders as input directories where STEP 2 outputs. The results will be saved as brant_headmotion_result.csv and brant_headmotion_exclusions.txt automatically.
_images/step3_check.png

fig.1 Utilities => Visual Check

_images/step3_check_output1_2.png

fig.2 Visual Check Output

_images/step3_check_2.png

fig.3 Utilities => Head Motion Estimation

_images/step3_check_output2.PNG

fig.4 brant_headmotion_result.csv

_images/step3_check_output3.PNG

fig.5 brant_headmotion_exclusions.txt

STEP 4: Functional Connectivity

  • Input Coordinates for ROI Drawing:
    Click FC, Draw ROI. Both making a *.csv with coordinates (input type: file) or input coordinates divided with “;”(input type: manual) are admitted in BRANT. Fig.2 shows the format of the csv file.
  • Choose Ref&mask:
    Choose a *.nii file as an example. The output files will follow its extracting information of origin, voxel size, bounding box, etc..
  • Extract ROI:
    Click Merge/extract ROIs in FC GUI, switch the operation to extract. Input roi index to choose which part you need in the roi file.
  • Input Directories for ROI Calculation:
    Click ROI Calculation in FC GUI. Select the folders as input directories where STEP 2 outputs.
  • ROI Calculation:
    Change filetype to fdnoGSR*.nii, input roi file and roi index or just use the default settings. Select output directories.
_images/step4_roi_drawing.png

fig.1 FC => Draw ROI

_images/step4_roi_drawing_input.png

fig.2 example for csv format input

_images/step4_roi_extract.png

fig.3 FC => Merge/extract ROIs

_images/step4_roi_calculation.png

fig.4 FC => ROI Calculation

STEP 5: SPON

  • AM:
    Click SPON button, then AM. Select the folders as input directories where STEP 2 outputs. Change filetype to dnoGSR*.nii. Select output directories.
  • ALFF/fALFF:
    Click ALFF/fALFF button in SPON GUI. Select the folders as input directories where STEP 2 outputs. Change filetype to dnoGSR*.nii. Remember that TR cannot be 0. Select output directories.
  • ReHo:
    Click ReHo button in SPON GUI. Select the folders as input directories where STEP 2 outputs. Change filetype to dnoGSR*.nii. Change nbr type to 6 (face neighbor), 18 (edge neighbor) or 26 (vertex neighbor) if necessary. Select output directories.
  • FCD/FCS:
    Click FCD/FCS button in SPON GUI. Select the folders as input directories where STEP 2 outputs. Change filetype to dnoGSR*.nii. The computation can use either OpenCL supported CPU or GPU. Select output directories. You can also calculate FCS with absolute value if necessary by selecting the abs fcs in metrics.

Important

Make sure your graphics drivers are up-to-date before using FCD/FCS function.

_images/step5_AM.png

fig.1 AM

_images/step5_ALFF.png

fig.2 ALFF/fALFF

_images/step5_ReHo.png

fig.3 ReHo

_images/step5_FCD_new.png

fig.4 FCD/FCS

STEP 6: Statistics for Differences among Groups

  • Input Directories of T-Tests:
    Click the STAT button, then T-Tests. Open the directories selection GUI and find the folder where ROI Calculation in STEP 4 outputs, then select the roi2roi_z_pearson_correlation folder as input directories. Remove strings from search results parsed by id index by typing the strings into the string removal.
  • Group Table:
    Create a *.csv file, input filenames without extensions, group strings, covariates as fig 2. Select this file for table. If your table file contains other information such as age or sex, you can input those titles in the regressors.
_images/step6_T_Tests.png

fig.1 STAT => T-Tests

_images/step6_T_Tests_Table.png

fig.2 Tabel input

STEP 7: Network Properties

  • Input Directories of Network Calculation:
    Click the NET button, then Network Calculation. Open the data directories selection GUI and find the folder where ROI Calculation in STEP 4 outputs.
  • Network Calculation:
    Click the … button of Network Properties, select those you need in the Brant Net Measure Options GUI. Those options with (*) will slow down the speed of calculation.
  • Input Directories of Network Statistics:
    Click Network Statistics button in NET GUI. Open the directories selection GUI and find the folder where Network Calculation above outputs.
  • Network Statistics
    Remove strings from search results parsed by id index such as _corr_z_network by typing the strings into the string removal. Select the *.csv file created in STEP 6 as input of table. If your table file contains other information such as age or sex, you can input those titles in the regressors to ignore them.
_images/step7_Network_Calculation.png

fig.1 NET => Network Calculation

_images/step7_Network_Statistics.png

fig.2 STAT => Network Statistics

_images/clustering_coefficient_corr_ttest2_grp1_vs_grp2.png

fig.3 Network Statistics Result 1

_images/clustering_coefficient_spar_ttest2_grp1_vs_grp2.png

fig.4 Network Statistics Result 2

STEP 8: Visualization

  • Surface Mapping
    Click the VIEW button, then Surface Mapping. Select material, lighting and shading if necessary.
  • ROI Mapping
    Click ROI Mapping button in VIEW GUI. Select material, lighting and shading if necessary.
  • Network Visualization
    Click Network Visualization button in VIEW GUI. Open the node directories selection GUI and find the brant_roi_info.csv file where ROI Calculation in STEP 4 outputs. Open the edge directories selection GUI and find the ttest2_grp1_vs_grp2_h_unc.txt file where T-Tests in STEP 6 outputs.
  • Circos
    Click the Embedded button, then Circos. Select the circos’ directories as input or circos dir, e.g. D:/circos-0.69-5/bin. ROI info can use the example brant_circos_3mm_273.csv file in */Matlab/toolbox/brant-master/circos. Open the edge directories selection GUI and find the ttest2_grp1_vs_grp2_h_unc.txt file where T-Tests in STEP 6 outputs.
_images/step8_Surface_Mapping.png

fig.1 VIEW => Surface Mapping

_images/step8_ROI_Mapping.png

fig.2 VIEW => ROI Mapping

_images/surface_mapping_result.png

fig.3 Result of Surface Mapping

_images/step8_ROI_Mapping_Result.png

fig.4 Result of ROI Mapping

_images/step8_Network_Visualization.png

fig.5 VIEW => Network Visualization

_images/step8_Network_Visualization_Result.png

fig.6 Result of Network Visualization

_images/step8_circos.png

fig.7 Embedded => Circos

_images/circos_result.png

fig.8 Result of Circos

File Name

Operation Name File/Folder Name
DICOM Convert brant_4D.nii
slice timing abrant_4D.nii
abrant_4D.mat
realign rabrant_4D.nii
meanabrant_4D.nii
rp_abrant_4D.txt
coregister wholebrain.nii
betStruchImg.nii
normalise wrabrant_4D.nii
denoise dGSRwrabrant_4D.nii
dnoGSRwrabrant_4D.nii
fdGSRwrabrant_4D.nii
fdnoGSRwrabrant_4D.nii
smooth s6fdGSRwrabrant_4D.nii
s6fdnoGSRwrabrant_4D.nii
Visual Check axial (folder)
coronal (folder)
ortho (folder)
sagital (folder)
Head Motion Est brant_headmotion_exclusions.txt
brant_headmotion_result.csv
Draw ROI brant_2_sphere_rois.nii/brant_2_cube_rois.nii
roi_info_sphere_2_rois.csv/roi_info_cube_2_rois.csv
ROI Calculation mean_ts(folder)
roi2roi_r_pearson_correlation (folder)
roi2roi_z_pearson_correlation (folder)
roi2wb_r_pearson_correlation (folder)
roi2wb_z_pearson_correlation (folder)
brant_roi_info.csv
roi_history.txt
T-Tests grp1_grp2_group_info.csv
group_info.mat
ttest2_diary.txt
ttest2_grp1_vs_grp2.mat
ttest2_grp1_vs_grp2_h_unc.txt
ttest2_grp1_vs_grp2_pval_left_unc.txt
ttest2_grp1_vs_grp2_pval_right_unc.txt
ttest2_grp1_vs_grp2_tval.txt
Network Calculation *_corr_z_network.mat
Network Statistics clustering_coeffecient_corr_ttest2_grp1_vs_grp2.png
clustering_coeffecient_spar_ttest2_grp1_vs_grp2.png
network_clustering_coeffecient.csv
network_clustering_coeffecient_ttest2_grp1_vs_grp2_stat.csv
readme.txt

Result of rs-fMRI data analysis

To validate the efficacy of our toolkit, we used the same preprocessing pipelines provided by BRANT and DPABI v2.3, to compare ReHo, fALFF and FCs using a rs-fMRI dataset consists of 18 patients with mild cognitive impairment (MCI), 17 patients with mild Alzheimer’s disease (mAD), 18 patients with severe Alzheimer’s disease (sAD) and 21 normal controls (NC). This dataset is available online.

AD subjects were diagnosed using standard operationalized criteria (Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV); American Psychiatric Association 1994 and National Institute of Neurological and Communicative Disorders and Stroke - Alzheimer’s Disease and Related Disorders Association (NINCDS-ADRDA)). The severity of dementia was assessed using the Clinical Dementia Rating (CDR) scale. Patients with a diagnosis of AD and CDR score of 1 were classified as mild AD and those with a CDR score of 2 or 3 were diagnosed as severe AD. MCI was diagnosed according to standard criteria, which included subjective memory loss with objective evidence of memory impairment in the context of normal or near-normal performance on other domains of cognitive functioning; minimal impairment of activities of daily living with a CDR score of 0.5. Normal volunteers have a CDR score of 0.

The MR images were acquired on a 3.0-T MR scanner (Magnetom Trio, Siemens, Germany). Functional MRI data were acquired using an echo planar imaging (EPI) sequence sensitive to BOLD contrast: Repetition time (TR) = 2000 ms, echo time (TE) = 30 ms, flip angle (FA) = 90°, matrix = 64 × 64, field of view (FOV) = 220 mm × 220 mm, slice thickness = 3 mm with inter-slice gap = 1 mm. Each brain volume comprised 32 axial slices, and each scanning session lasted for 360 s. Sagittal T1-weighted MR images were acquired by a magnetization-prepared rapid gradient-echo sequence (TR/TE = 2000/2.6 ms, FA = 9°, matrix = 256 × 224, FOV = 256 mm × 224 mm, 176 continuous sagittal slices with 1 mm thickness).

No significant differences (P>0.05) of age (two-tailed two sample t-test), gender (chi-squared test) and education (two-tailed two sample t-test) were found between each patient group and NC group. T-statistic maps of fALFF, ReHo and results of FCs (P<0.001, uncorrected) based on different toolkits have quite similar patterns, which suggest BRANT is another optimal toolkit for rs-fMRI research community. In the results, the minor differences can be induced by 489 different implementations of trends removal, covariance regression and band-pass filter. For the results, we didn’t draw a strong conclusion, since the interpretation requires multiple comparison corrections and is not the main point of the current study.

We have listed the differences among BRANT and other Matlab-based toolboxes in FAQs. We can also find that t-statistic maps of fALFF, ReHo and results of FCs (P<0.001, uncorrected) based on different toolkits have quite similar patterns, which suggest BRANT is another optimal toolkit for rs-fMRI research community.

_images/result2.png

fig.1 T-statistic maps of fALFF and ReHo based on different toolkits

_images/result3.png

fig.2 results of FCs based on different toolkits

  • References:
  1. Yan CG, Wang XD, Zuo XN, Zang YF. DPABI: Data Processing & Analysis for (Resting-State) Brain Imaging. Neuroinformatics 2016; 14(3): 339-51.
  2. Wang J, Wang X, Xia M, Liao X, Evans A, He Y. GRETNA: a graph theoretical network analysis toolbox for imaging connectomics. Front Hum Neurosci 2015; 9: 386.
  3. Whitfield-Gabrieli S, Nieto-Castanon A. Conn: a functional connectivity toolbox for correlated and anticorrelated brain networks. Brain Connect 2012; 2(3): 125-41.
  4. Zuo XN, Kelly C, Di Martino A, Mennes M, Margulies DS, Bangaru S, et al. Growing together and growing apart: regional and sex differences in the lifespan developmental trajectories of functional homotopy. J Neurosci 2010; 30(45): 15034-43.
  5. Liu Y, Yu C, Zhang X, Liu J, Duan Y, Alexander-Bloch AF, et al. Impaired long distance functional connectivity and weighted network architecture in Alzheimer’s disease. Cereb Cortex 2014; 24(6): 1422-35.
  6. He X, Qin W, Liu Y, Zhang X, Duan Y, Song J, et al. Abnormal salience network in normal aging and in amnestic mild cognitive impairment and Alzheimer’s disease. Human brain mapping 2014; 35(7): 3446-64.
  7. Liu J, Zhang X, Yu C, Duan Y, Zhuo J, Cui Y, et al. Impaired Parahippocampus Connectivity in Mild Cognitive Impairment and Alzheimer’s Disease. J Alzheimers Dis 2016; 49(4): 1051-64.
  8. McKhann G, Drachman D, Folstein M, Katzman R, Price D, Stadlan EM. Clinical diagnosis of Alzheimer’s disease: report of the NINCDS-ADRDA Work Group under the auspices of Department of Health and Human Services Task Force on Alzheimer’s Disease. Neurology 1984; 34(7): 939-44.
  9. Morris JC. The Clinical Dementia Rating (CDR): current version and scoring rules. Neurology 1993; 43(11): 2412-4.
  10. Petersen RC, Smith GE, Waring SC, Ivnik RJ, Tangalos EG, Kokmen E. Mild cognitive impairment: clinical characterization and outcome. Arch Neurol 1999; 56(3): 303-8.
  11. Petersen RC, Doody R, Kurz A, Mohs RC, Morris JC, Rabins PV, et al. Current concepts in mild cognitive impairment. Arch Neurol 2001; 58(12): 1985-92.
  12. Choo IH, Lee DY, Youn JC, Jhoo JH, Kim KW, Lee DS, et al. Topographic patterns of brain functional impairment progression according to clinical severity staging in 116 Alzheimer disease patients: FDG-PET study. Alzheimer Dis Assoc Disord 2007; 21(2): 77-84.

Download

  • You can get the latest version of BRANT from
    https://github.com/YongLiuLab/brant-stable/archive/master.zip
  • Some functions are based on SPM, you can get it from
    http://www.fil.ion.ucl.ac.uk/spm/software/spm12
  • Circos in STEP 8 can be downloaded from
    http://circos.ca/distribution/circos-0.69-5.tgz
  • Unzip the brant-master.zip file and move both brant and spm12 folders to /path/of/toolbox. (The path can be anywhere in your computer as long as it’s in English)
  • Configure SPM paths:
    1. Click Set Path in MATLAB
    2. Click add folder
    3. Select SPM’s root folder
    4. Run spm fmri in MATLAB’s Command Window to let spm add its subfolders. (Note that some SPM’s subfolders are called internally in SPM and should not be added in MATLAB’s search path. Scripts in those folders are conflicting with MATLAB functions, and could cause untraceable errors)
  • Configure BRANT paths:
    1. Click Set Path in MATLAB
    2. Click Add with subfolders
    3. Select the unzipped BRANT folder
    4. Click Save.

Tip

  • An alternative way to configure both SPM and BRANT Paths:
    • Run in MATLAB’s Command Window:
      • cd('/path/of/unzipped/brant/'); % to set current working directory to the unzipped brant
      • brant_configure_paths; % to add BRANT paths (same as add with subfolders)
      • brant_configure_paths('/path/of/spm12/'); % to add SPM paths

FAQs

How can I Download and Install BRANT

  • You can get the latest version of BRANT from
    https://github.com/YongLiuLab/brant-stable/archive/master.zip
  • Some functions are based on SPM, you can get it from
    http://www.fil.ion.ucl.ac.uk/spm/software/spm12
  • Circos in STEP 8 can be downloaded from
    http://circos.ca/distribution/circos-0.69-5.tgz
  • Unzip the brant-master.zip file and move both brant and spm12 folders to /path/of/toolbox. (The path can be anywhere in your computer as long as it’s in English)
  • Configure SPM paths:
    1. Click Set Path in MATLAB
    2. Click add folder
    3. Select SPM’s root folder
    4. Run spm fmri in MATLAB’s Command Window to let spm add its subfolders. (Note that some SPM’s subfolders are called internally in SPM and should not be added in MATLAB’s search path. Scripts in those folders are conflicting with MATLAB functions, and could cause untraceable errors)
  • Configure BRANT paths:
    1. Click Set Path in MATLAB
    2. Click Add with subfolders
    3. Select the unzipped BRANT folder
    4. Click Save.

Tip

  • An alternative way to configure both SPM and BRANT Paths:
    • Run in MATLAB’s Command Window:
      • cd('/path/of/unzipped/brant/'); % to set current working directory to the unzipped brant
      • brant_configure_paths; % to add BRANT paths (same as add with subfolders)
      • brant_configure_paths('/path/of/spm12/'); % to add SPM paths

What is BRANT’s Advantage Compared with Other Matlab-based Toolboxes

Compared with DPABI v2.3, GRETNA v2.0.0 and CONN v17.f, the differences are listed in the table below.

In conclusion, BRANT can directly support *.gz for most postprocessing functions, use OPENCL-based parallel computing for time consuming FCD/FCS calculation to save time. Other than focusing on several specific types of rs-fMRI data processing, functions of BRANT cover a wide range of rs-fMRI data processing methods. Also, GUIs are created automatically with a few lines of MATLAB code instead of drawn manually.

_images/result1.png

fig.1 differences among four Matlab-based toolboxes

How to Format Tables Correctly

  • In some fuctions, you may have to use some *.csv files to input data. If you cannot find any proper existed *.csv files, you can create a *.csv file by Excel or Notepad. e.g.

    • STEP 4 => FC => Draw ROI => Coordinates in txt or csv file (input type: file)
      x y z label
      40 -16 50 ROI1
      -40 -16 50 ROI2
    • STEP 6 => STAT => T-Tests => table
      name group age sex
      NC_02_0001_fMRI grp1 31 0
      NC_02_0002_fMRI grp2 27 0
      NC_02_0003_fMRI grp1 36 0
      NC_02_0004_fMRI grp2 37 1

      If you select paired t-test, the table should contain an additional column of paired_t_idx

      name group filter age paired_t_idx
      subj1 stage1 center1 28 1
      subj2 stage1 center1 27 2
      subj3 stage2 center1 30 1
      subj4 stage2 center2 25 2
    • STEP 8 => VIEW => Network Visualization => node
      x y z label vox_num index
      -4.6213 15.6 53.2468 A8m_l 235 1
      … … … … … …

      This brant_roi_info.csv file can be found in the folder as the out dir of ROI Calculation in STEP 4.

    • STEP 8 => Embedded => Circos => roi info
      x y z label_old label vox_num index module index_module index_node
      -4.6213 15.6 53.2468 SFG_L_7_1 A8m_l 235 1 SFG 1 1
      … … … … … … … … … …

      This brant_circos_3mm_273.csv file can be found in the */Matlab/toolbox/brant-master/circos.

What can I Do When an Error Occurs

  • Here are some common errors with their reasons and solutions below. You can also modify parameters with the help of examples and rerun former steps. If these all cannot work, please contact us.
_images/faqs_error_new_2.png

fig.2 error during DICOM Convert

  • Reason: Misuse id index.
  • Solution: Delete files created before the error occurs. Read example to use id index properly.
_images/faqs_error_new_1.png

fig.3 error after preprocessing

  • Reason: The preprocessing may be interrupted and create wrong *.nii files.
  • Solution: Delete files created during preprocessing, examine your preprocessing settings and run preprocessing again.
_images/faqs_error_new_5.png

fig.4 error during preprocessing

  • Reason: TR(s) in Preprocessing is zero.
  • Solution: Set tr(s) in Silce Timing to a legal value.
_images/faqs_error_new_3.png

fig.5 error during STAT

  • Reason: The table file is incorrect.
  • Solution: Confirm that the table file contains a row with group information and contains only one row. Open it with notepad to check if the data in this file is divided by comma.
_images/faqs_error_new_4.png

fig.6 error after DICOM Convert

  • Reason: More than one *.nii files were found in above directories.
  • Solution: Uncheck the delete first N timepoints or convert just one set of data per time.

Publications using Brant

[17] P. Wang et al., “Aberrant Hippocampal Functional Connectivity Is Associated with Fornix White Matter Integrity in Alzheimer’s Disease and Mild Cognitive Impairment,” J Alzheimers Dis., 2020; 75:1153-1168.

[16] D. Jin et al., “Grab-AD: Generalizability and reproducibility of altered brain activity and diagnostic classification in Alzheimer’s Disease,” Human Brain Mapping, 2020; 41:3379-3391.

[15] Z. Zhou et al., “A toolbox for brain network construction and classification (BrainNetClass),” Human Brain Mapping, hbm.24979, Mar. 2020

[14] T. Liebe, J. Kaufmann, M. Li, M. Skalej, G. Wagner, and M. Walter, “In vivo anatomical mapping of human locus coeruleus functional connectivity at 3 T MRI,” Human Brain Mapping, p. hbm.24935, Jan. 2020.

[13] A. Li et al., “A neuroimaging biomarker for striatal dysfunction in schizophrenia,” Nature Medicine, Mar. 2020.

[12] Quan M, Zhao T, Tang Y, Luo P, Wang W, Qin Q, Li T, Wang Q, Fang J, Jia J. “Effects of gene mutation and disease progression on representative neural circuits in familial Alzheimer’s disease,” Alzheimers Res Ther. 2020;12(1):14.

[11] Zhu W, Huang H, Yang S, Luo X, Zhu W, Xu S, Meng Q, Zuo C, Zhao K, Liu H, Liu Y, Wang W. “Dysfunctional Architecture Underlines White Matter Hyperintensities with and without Cognitive Impairment,” J Alzheimers Dis, 2019;71(2):461-76.

[10] C. Vries, R. T. Staff, G. D. Waiter, M. O. Sokunbi, A. L. Sandu, and A. D. Murray, “Motion during Acquisition is Associated with fMRI Brain Entropy,” IEEE Journal of Biomedical and Health Informatics, pp. 1–1, 2019.

[9] J. Li et al., “ASAF: altered spontaneous activity fingerprinting in Alzheimer’s disease based on multisite fMRI,” Science Bulletin, Apr. 2019.

[8] N. Luo et al., “Brain function, structure and genomic data are linked but show different sensitivity to duration of illness and disease stage in schizophrenia.,” Neuroimage Clin, vol. 23, pp. 101887–101887, 2019.

[7] R. Pang et al., “Altered Regional Homogeneity in Chronic Insomnia Disorder with or without Cognitive Impairment,” AJNR Am J Neuroradiol, vol. 39, no. 4, pp. 742–747, Apr. 2018.

[6] H. Sun et al., “Regional homogeneity and functional connectivity patterns in major depressive disorder, cognitive vulnerability to depression and healthy subjects,” Journal of Affective Disorders, vol. 235, pp. 229–235, Aug. 2018.

[5] Y. Zhang, X. Liu, K. Zhao, L. Li, and Y. Ding, “Study of altered functional connectivity in individuals at risk for Alzheimer’s Disease.,” Technol Health Care, vol. 26, no. S1, pp. 103–111, 2018.

[4] S. Peeters et al., “Reduced specialized processing in psychotic disorder: a graph theoretical analysis of cerebral functional connectivity,” Brain Behav, vol. 6, no. 9, p. e00508, 2016.

[3] M. Xiao et al., “Attention Performance Measured by Attention Network Test Is Correlated with Global and Regional Efficiency of Structural Brain Networks,” Front. Behav. Neurosci., vol. 10, 2016.

[2] S. Yang et al., “Altered Intranetwork and Internetwork Functional Connectivity in Type 2 Diabetes Mellitus With and Without Cognitive Impairment,” Sci Rep, vol. 6, no. 1, p. 32980, Dec. 2016.

[1] Y. Wang et al., “Using Regional Homogeneity to Reveal Altered Spontaneous Activity in Patients with Mild Cognitive Impairment,” BioMed Research International, vol. 2015, pp. 1–8, 2015.