These are chat archives for thunder-project/thunder

Feb 2015
Jeremy Freeman
Feb 18 2015 00:27
Yeah if keys are constant, sorting will be constant
Just put up a PR that expands and refactors support for sorting to handle some of these concerns, great if you could take a look thunder-project/thunder#112
Jason Wittenbach
Feb 18 2015 21:20
Checking it out now!
Feb 18 2015 23:01

Hey, we're doing ECoG neurolinguistics and I just switched from matlab to python. We're continually scaling up our data collection and mining and I wanted to see if Thunder is the right tool for us. Some general things we do now or would like to do:

  • compute the Hilbert transform of band-passed ECoG signal
  • other types of temporal ECoG frequency decomposition: STFT and wavelet decomposition
  • computing spectro-temporal receptive fields (from potentially days worth of neural and auditory data- requires spectro-temporal decomposition of both)
  • event-related responses (I see event series are on the to-do list)
  • categorical and continuous decoding

How easy would it be to use Thunder to do these things? Are there relevant examples for these types of things? I know it's a different direction from a lot of the examples, but the math overlaps quite a bit