Sometime around 2002, Jonthan Taylor started writing BrainSTAT, a Python version of Keith Worsley’s FmriSTAT package.
In 2004, Jarrod Millman and Matthew Brett decided that they wanted to write a grant to build a new neuoimaging analysis package in Python. Soon afterwards, they found that Jonathan had already started, and merged efforts. At first we called this project BrainPy. Later we changed the name to NIPY.
In 2005, Jarrod, Matthew and Jonathan, along with Mark D’Esposito, Fernando Perez, John Hunter, Jean-Baptiste Poline, and Tom Nichols, submitted the first NIPY grant to the NIH. It was not successful.
In 2006, Jarrod and Mark submitted a second grant, based on the first. The NIH gave us 3 years of funding for two programmers. We hired two programmers in 2007 - Christopher Burns and Tom Waite - and began work on refactoring the code.
Meanwhile, the team at Neurospin, Paris, started to refactor their FFF code to work better with python and NIPY. This work was by Alexis Roche, Bertrand Thirion, and Benjamin Thyreau, with some help and advice from Fernando Perez.
In 2008, Fernando Perez and Matthew Brett started work full-time at the UC Berkeley Brain Imaging Center. Matthew in particular came to work on NIPY.