Progress and Paris

C-Day plus 13

·        I’ve spent the past couple of weeks have getting my hands dirty with the underlying physics equations behind solving for forward solutions in source imaging (mostly with Mosher et al., 1999 and Hämäläinen &Sarvas, 1989). The forward solution is a matrix that relates each point on the cortex to the MEEG sensors. Once you have the forward solution, you can use some pretty fancy mathematics (including Tikinov regularization) to find a pseudoinverse for this matrix – called the inverse solution. Then you’re able to relate the sensor measurements with estimates of what areas of the brain are active, which is the fundamental motivation for source imaging.

·        One component of my project is focused on modifying this forward solution, so learning something about the way it was originally formulated currently has been a useful endeavor. I also found out from the algorithm’s creator that no material exists to help bridge the gap between the idealized and published equations and the optimized and cryptic code. Therefore, I’ve added quite a few comments and docstring improvements to make this easier for the next programmer wrestling with these equations. Later in the summer, I’m hoping to use this knowledge to find the relationship between the cortical surface and the multipolar moment space I started describing in my last post (see the discussion on SSS and spherical harmonics). This should provide a number of benefits that I’ll discuss when I pick this portion of the work back up.

·        For now, I’m going to try to make some headway on the first aim of my project: Maxwell filtering. Again, the Maxwell filter implemented in SSS is just an elegant way to exclude noise from MEEG recordings using physics (see my earlier post about steam bowls and floating frogs for more description or Taulu 2005 for one of the SSS papers). 

·        Last thing: the heads of the MNE-Python project have generously offered to fly me to Paris for a weeklong coding sprint in July! I’m pretty excited to finally meet all of the MNE-Python crew and learn more about how the Europeans view science and research.