Oscar Esteban

Fix images when they are badly oriented

2

https://poldracklab.github.io/mriqc/abide-reports/anatomical_sub-50277_ses-0_run-0_report.html https://poldracklab.github.io/mriqc/abide-reports/anatomical_sub-50288_ses-0_run-0_report.html https://poldracklab.github.io/mriqc/abide-reports/anatomical_sub-50328_ses-0_run-0_report.html

[ENH] Use Label encoder for site handling

3

https://github.com/poldracklab/mriqc/blob/8c2e33533637e53bbd5488b8c2ee84c7138a5fe8/mriqc/classifier/sklearn/preprocessing.py#L128 https://github.com/poldracklab/mriqc/blob/8c2e33533637e53bbd5488b8c2ee84c7138a5fe8/mriqc/classifier/nested_helper.py#L169 https://github.com/poldracklab/mriqc/blob/18c40d43d0324f4653d3ca540fb90185a1b46af9/mriqc/classifier/helper.py#L534

[BUG] ValueError: Input contains NaN, infinity or a value too large for dtype('float32').

7

Subject 50818 on ABIDE -- Original data of this subject are already **skull-stripped**. ``` Node: workflow_enumerator.anatMRIQCT1w.ComputeIQMs.datasink Working directory: /scratch/users/oesteban/abide-mriqc_0.9.6-2017-06-03-99db97c9be2e/work/workflow_enumerator/anatMRIQCT1w/ComputeIQMs/_in_file_..mnt..ABIDE..sub-50818..anat..sub-50818_T1w.nii.gz/datasink Node inputs: _outputs = {'qi_2': 0.0} acq_id = ignore_exception = False...

[ENH] Reduce HTML reports size

3

Particularly, the anatomical-individual reports can take up to 12MB and that's too much.

More appropriate parcellation+segmentation for carpetplot

1

Possibly using the [_lobes_ segmentation](https://github.com/BIC-MNI/EZminc/issues/7#event-840145988)

IQMs: adopt a "plugin" pattern

2

So measures can be easily plugged/unplugged.

Misleading WARNING of the group-level execution.

1

It is unclear whether this message ``` 2018-04-23 18:16:42,747 mriqc.cli:WARNING No IQM-JSON files were found for the T2w data type in /scratch/dataset/derivatives. The group-level report was not generated. ``` is...

Refactor nonsteady state detector and conformation of BOLD images

4

It should use nipype's [NonSteadyStateDetector](http://nipype.readthedocs.io/en/latest/interfaces/generated/nipype.algorithms.confounds.html#nonsteadystatedetector), and then workaround the EstimateReferenceImage of `fMRI_HMC_afni` to use the file generated here or pass in the number of nonsteady state frames. The n_start should...

Plot standard deviation map after motion correction

3

... in order to be able to spot other sources of variability masked by head motion.