Improve performance on large datasets by using bids2table

[arokem]

I learned about this today, and have not looked into it closely yet, but the benchmarks look promising: https://github.com/childmindresearch/bids2table.