High-quality neuroscience requires clear hypotheses, preregistered methods, open data, and reproducible analysis. Exploratory data mining is insufficient for theory development.
Recommended practices:
1) Poldrack et al. (2017) – Scanning the horizon: Towards transparent and reproducible neuroimaging research
https://www.nature.com/articles/s41562-016-0021
2) The Brain Imaging Data Structure (BIDS) for organizing neuroimaging data:
https://bids.neuroimaging.io
3) NeuroKit2, for reproducible biosignal analysis in Python:
https://neurokit2.readthedocs.io
