Fiber Tracking

Use --action=trk to perform deterministic whole-brain or ROI-based fiber tracking in DSI Studio (Yeh et al. PLoS ONE 8(11):e80713, 2013). Additionally, you can compute tract-to-region connectomes, tract statistics, track-density images, cluster or recognize known bundles, and even perform differential analysis—all from this single command.

For automatic atlas-based bundle mapping, see --action=atk instead:
Atlas-Based Tracking Documentation

Examples

Whole-brain deterministic tracking

dsi_studio --action=trk \
           --source=subject.fib.gz \
           --output=subject_all.tt.gz

Seed-and-target ROI tracking

dsi_studio --action=trk \
           --source=subject.fib.gz \
           --seed=wholeBrain.nii.gz \
           --roi=regionA.nii.gz \
           --roi2=regionB.nii.gz \
           --output=streamlines.tt.gz

Mapping the left arcuate fasciculus and derive is tract-to-region connectome at HCP-MMP parcellation

dsi_studio --action=trk \
           --source=subject.fib.gz \
           --track_id=ArcuateFasciculusL \
           --connectivity=HCP-MMP

Core Options

Option Default Description
--action (required) Must be trk.
--source (required) Path to your .fib.gz (or .fz) file.
--output   Output tractography file (e.g. –output=tract.tt.gz) or directory (e.g. –output=/path).
--track_id (optional) Specify the name of the bundle to be mapped (e.g. –track_id=ArcuateFasciculusL ). The complete list can be found here
--tip_iteration 0 or 16 Specify pruning iterations. Default is 16 for track_id or dt_threshold_index.
--thread_count Hardware threads Number of threads to use.

Tracking Parameters

Option Default Description
--method Algorithm in file Tracking algorithm. 0=streamline, 1=RK4, etc.
--tract_count 0 Number of tracts to generate (alternative to seed count).
--seed_count Auto Number of seeds to launch. Overrides tract_count.
--track_voxel_ratio 0.5 Seeds-per-voxel ratio.
--turning_angle 0 (random) Maximum allowable turning angle (deg). Default is random between 15° and 90°.
--step_size 0 (random) Step size in mm. Default is random between 1–3 voxels.
--smoothing 0 Fraction of previous direction to mix in (0-1).
--min_length Dataset-specific Minimum fiber length (mm).
--max_length Dataset-specific Maximum fiber length (mm).
--otsu_threshold FA threshold × 0.6 Default threshold for FA-based seeding.
--threshold_index — Use a different diffusion index (e.g., QA) for termination instead of FA.
--check_ending Off for whole-brain tracking Drop tracks that do not terminate within any ROI.
--parameter_id — Load all tracking parameters from a parameter code.
--random_seed 0 Seed for random number generation in fiber tracking. Set to an integer to vary seed sequences.

ROI / Region Options

The following region types can be specified (comma-separated actions are applied in order):

--seed=<file>
--roi=<file_or_atlas:region>,<other region file>,<region actions>,... 
--roi2=…
--roa=…
--end=…
--ter=…
--nend=…
--lim=…

Region Actions

Post-Tracking Routine

Option Description
--output=<file> Save tractography result (e.g., --output=result.tt.gz).
--trk_format=<format> Set default tract format if --output is not specified (e.g., --trk_format=tt.gz).
--delete_repeat=<length> Remove duplicate streamlines within the specified voxel distance (e.g., --delete_repeat=8).
--delete_by_length=<length> Remove streamlines shorter than the specified voxel length.
--cluster=<method>,<count>,<detail>,<output> Perform clustering (e.g., --cluster=0,50,8,cluster.txt).
• method: 0 = single linkage, 1 = k-means, 2 = EM
• detail: only used in method 0
--recognize=<file> Apply atlas-based bundle recognition and save recognized tracts.
--template_track=<file> Save tracts in template voxel space.
--mni_track=<file> Save tracts in MNI space.
--end_point=<file> Save endpoint coordinates (both ends) to .txt or .mat.
--end_point1=<file> Save starting point coordinates to .txt or .mat.
--end_point2=<file> Save endpoint coordinates to .txt or .mat.
--export=<type>,<type>,... Export tract data (e.g., --export=stat,tdi,report:dti_fa:3:1)
• stat: summary statistics
• tdi: track density image
• report:<index>:<direction>:<bandwidth>: tract profile
  Example: --export=report:dti_fa:3:1 exports FA profile along fiber direction using bandwidth = 1.
• index: e.g., dti_fa, md, ad
• direction: 0 = x, 1 = y, 2 = z, 3 = along fiber, 4 = mean
• bandwidth: regression bandwidth

Connectivity Matrix (Connectome)

Option Description
--connectivity=<atlases> Comma-separated list of brain parcellation files in NIFTI format (e.g., AAL2,HCP-MMP) used to compute region-to-region (R2R) and tract-to-region (T2R) connectomes.
--connectivity_value=<metrics> Metric(s) used to compute entries of the R2R connectome matrix (e.g., count, length, qa). Multiple metrics can be specified with comma separation to generate multiple matrices in one run.
--connectivity_type=<"pass","end"> Specify whether to use streamline pass-through or endpoints to compute the R2R connectome. Multiple types can be specified with comma separation to generate multiple matrices in one run.
--connectivity_threshold=<value> Threshold for binarizing connection weights in the R2R connectome prior to graph analysis.
--connectivity_output=<types> Output type(s) for the R2R connectome. Multiple types can be specified with comma separation:
• matrix → saves *.connectivity.mat (MATLAB v4 square matrix)
• connectogram → saves *.connectogram.txt (for visualization with CIRCUS)
• measure → saves *.network_measures.txt (graph-theoretic network measures).

The connectivity command generates two types of connectomes:

Differential Tracking

Option Description
--dt_metric1 Metric for baseline tracking (e.g., qa).
--dt_metric2 Metric for comparison tracking (e.g., nqa).
--dt_threshold Percent change threshold for differential analysis.
--dt_threshold_type Type of threshold: 0=(m1−m2)/m1, 1=(m1−m2)/m2, 2=abs(m1−m2).