[FIX] Change to Linear for upsampling (#820)

* Change to Linear for upsampling

* Update docs/quickstart.rst

Co-authored-by: Taylor Salo <[email protected]>

---------

Co-authored-by: Taylor Salo <[email protected]>

docs/quickstart.rst

@@ -17,9 +17,14 @@ One way to process these data would be to call ``qsiprep`` like this::
 
   qsiprep \
     /path/to/inputs /path/to/outputs participant \
-    --output-resolution 1.2 \
+    --output-resolution X \
     --fs-license-file /path/to/license.txt
 
+.. warning::
+   The above example sets the ``--output-resolution`` to ``X``, where in
+   real usage this number should be set to a value in millimeters.
+   See :ref:`output_resolution` for specifics.
+
 
 Grouping scans
 ==============
@@ -53,15 +58,14 @@ Specifying outputs
 ==================
 
 .. note::
-   This section covers ``--output-resolution 1.2``, and
+   This section covers ``--output-resolution X``, and
    ``--skip-t1-based-spatial-normalization``.
 
 Unlike with fMRI, which can be coregistered to a T1w image and warped to a
 template using the T1w image's spatial normalization, the T1w images do not
 contain enough contrast to accurately align white matter structures to a
 template. For this reason, spatial normalization is typically done *after*
-models are fit. Therefore we omit the ``--output-spaces`` argument from
-preprocessing: i.e. **no template warping takes place by default**.
+models are fit.
 
 All outputs will be registered to the T1w image (or the
 AC-PC aligned b=0 template if ``--dwi-only`` was specified) but will have
@@ -79,15 +83,19 @@ brain graphs. If you want to save ~20 minutes of computation time, this
 normalization can be disabled with the
 ``--skip-t1-based-spatial-normalization`` option.
 
+.. _output_resolution:
+Output Resolution and Resampling
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
 The ``--output-resolution`` argument determines the spatial resolution of the
 preprocessed dwi series. You can specify the resolution of the original data
 or choose to upsample the dwi to a higher spatial resolution. Some
 post-processing pipelines such as fixel-based analysis recommend resampling
 your output to at least 1.3mm resolution. By choosing this resolution here,
 it means your data will only be interpolated once: head motion correction,
 susceptibility distortion correction, coregistration and upsampling will be
-done in a single step. If your are upsampling your data by more than 10%,
-QSIPrep will use BSpline interpolation instead of Lanczos windowed sinc
+done in a single step. If you are upsampling your data by more than 10%,
+QSIPrep will use Linear interpolation instead of Lanczos windowed Sinc
 interpolation.
 
 

qsiprep/interfaces/images.py

@@ -501,8 +501,8 @@ def _run_interface(self, runtime):
             resolution_cutoff = 0.9 * np.array(nb.load(input_file).header.get_zooms()[:3])
             print(output_resolution, resolution_cutoff)
             if np.any(output_resolution < resolution_cutoff):
-                interpolator = "BSpline"
-                LOGGER.warning("Using BSpline interpolation for upsampling")
+                interpolator = "Linear"
+                LOGGER.warning("Using Linear interpolation for upsampling")
                 break
         self._results["interpolation_method"] = interpolator
         return runtime