Support single-PLD Q2TIPS PASL and multi-PLD PCASL

[tsalo]

Closes #250 and closes #251.

To do:

• Decide between Juttukonda 2021 unweighted average and Dai 2012 weighted average for multi-PLD PCASL CBF calculation.
  • I don't know how to do the weighted average yet.
• Define T1tissue based on magnet strength.
  • I'm not aware of any papers with formulae, but Manuel Taso pointed me toward Wright et al. (2008). This unfortunately limits us to 1.5, 3, and 7T data.
• Double-check the code and formulae.

Changes proposed in this pull request

• Estimate ATT for multi-PLD PCASL data using the Dai 2012 method. Use this to calculate CBF for those data.
• Use formula from Noguchi 2015 to calculate CBF for single-PLD Q2TIPS data.
• Add license info for all copied, modified, translated, or otherwise external code.
• Update test datasets by removing hidden files.
• Also, fix test_002 dataset:
  • Drop second deltam volume. It's just a duplicate of the first, and was a conversion bug.
  • Add necessary metadata fields, including RepetitionTimePreparation for the M0 scan and LabelingEfficiency, BackgroundSuppression, and BackgroundSuppressionNumberPulses for the asl scan.
• Add modified version of FSL Split interface that enforces a list output.
• Modify _select_first_in_list to only return the first item if a list is provided.
• Rename aslprep.interfaces.cbf_computation to aslprep.interfaces.cbf.