.. _new:

What is new
===========

.. admonition:: 0.1.1: News on the interface front.

    **Design matrices:**

    * Whereas previously you had to provide `fmrifit` with a formal
      description of the experimental design, you may now provide
      `fmrifit` with a custom design matrix. Have a look at the new
      command line function (CLF) `csv2design`.

    * You would like to use FSL together with fmristats? You can now
      provide `fmrifit` with FSL's own design matrix. Have a look at the
      new CLF `fsl2design`.

    * You would like to use SPM together with fmristats? You can now
      provide `fmrifit` with SPM's own design matrix. Have a look at the
      new CLF `spm2design`.

    **BOLD contrasts:**

    * You can now calculate any BOLD contrast post-hoc after fitting
      your model to the FMRI data. This means you don't need to provide
      `fmrifit` with a fixed, a-priory defined list of BOLD contrasts
      any more.

    * You can now also specify BOLD contrasts in the form of a *named
      dictionary*. This named dictionary can be re-use when analysing
      multible subjects. Have a look at the files:

        - If you are using FSL: :download:`fsl-parameter-dict.txt`.
        - If you are using SPM: :download:`spm-parameter-dict.txt`.

    **Head movements and reference spaces:**

    * The interface has grown and now gives you even more options. Have
      a look at `fmririgids`.

    * You would like to use the head movement estimates as they were
      estimated by FSL? Have a look at the new `fsl2rigids`.

    * You would like to use the head movement estimates as they were
      estimated by SPM? Have a look at the new `spm2rigids`.

    Together, your analysis may now look as follows:

    .. code:: shell

        # Use the head movements as estimated by SPM
        spm2rigids -v \
            --study-name abc --cohort run0 --id 0 --datetime 2019-01-22-1300 \
            --paradigm ckq
            --par rp_run0_5.txt \
            --par-path raw3/fmristats_01/run0/realign \

        # Analyse your data in native space (the space isometric to the subject)
        fmripop -v \
            --study-name abc --cohort run0 --id 0 --datetime 2019-01-22-1300 \
            --paradigm ckq
            --vb-name isospm \
            --rigids spm --diffeomorphism-nb identity --resolution 0. \

        # Fit the data
        fmrifit -v \
            --study-name abc --cohort run0 --id 0 --datetime 2019-01-22-1300 \
            --paradigm ckq \
            --vb-name isospm --diffeomorphism identity --rigids spm \
            --stimulus-block left --control-block right --fwhm 5 \
            --use-custom-design --design spm --parameter-dict spm-parameter-dict.txt

    Last but not least, the in-line documentation has further been
    polished.

.. admonition:: Everything new in 0.1.0: Well, not everything but well
    enough to justify a version increase.

    * Most importantly, fmristats is a lot faster than before. You may
      now expect the calculations to be finished in around 4h.

    * The API now consistently builds around the *study instance* that
      has been introduced in 0.0.4. This makes the whole user interface
      much more streamlined and, well, consistent. To get an idea of the
      changes, please have a look at the :ref:`getting-started` and
      :ref:`study-interface` chapters.

    * Last but not least, the documentation has been polished a lot.

    Unfortunately, version 0.1.0 is *not* backwards compatible with
    versions <0.1.0. Please upgrade to the most recent version.