search

Neuroinflab / kCSD-python

Kernel Current Source Density
https://kcsd-python.readthedocs.io/en/latest/
Other
15 stars 23 forks source link

Reliability and error maps. #75

Closed ccluri closed 5 years ago

ccluri commented 5 years ago

All through use 300 dpi resolution

6x6_broken_electrodes_subplotwith_electrodes

1) Please upload the scripts to generate these figures 2) Use gridspec and move the cbar to occupy the whole two rows. 3) Use figure_properties and for using consistent font size and font styles. 4) The 'x labels' 'depth' can be omitted for the first row 5) The subplot labels 'A', 'B' etc must be figure labels and not inside the title itself.

Side question - are these same as the error calculation plots in the tutorial? If so we should use consistent cmap.

reliability_map_random_symm 1 )The numbers here on the contours are not visible at all - either remove them or use a different color - like blue. 2) The cbar needs units 3) Can you try transparency instead of overlay of two different contours?

kcsd_with_error_mask 1) Use figure properties for the font size and font consistency 2) D you can reduce the cbar range so that the darks are more visible. You can use colorbar's extend parameter (https://matplotlib.org/gallery/images_contours_and_fields/image_masked.html#sphx-glr-gallery-images-contours-and-fields-image-masked-py)

m-kowalska commented 5 years ago

First picture is going to be removed. This is not the standard reliability map but the difference between reliability map for the complete setup and the setup with missing contacts.

@ccluri What do you think about adding the third row to the error calculation plots in the tutorial to show the reliability map for random sources (both small and large) and three plots for calculated differences (5, 10 and 20 missing contacts) in the grey scale?

m-kowalska commented 5 years ago

reliability_map_random_symm

@ccluri Does it look better?

ccluri commented 5 years ago

@m-kowalska Blue outlines look better. Can you point me to the code of this - I'd like to play around with it a bit.

And regarding 'What do you think about adding the third row to the error calculation plots in the tutorial to show the reliability map for random sources (both small and large) and three plots for calculated differences (5, 10 and 20 missing contacts) in the grey scale?' Sure go ahead, lets see how that would look! :) - Code for this figure is here - https://github.com/Neuroinflab/kCSD-python/blob/master/figures/kCSD_properties/tutorial_broken_electrodes.py

ccluri commented 5 years ago

@m-kowalska https://www.dropbox.com/sh/6kykj4d3dx3fp5s/AAACtN49VCbAHA9otOfNXbnOa?dl=0 Has the files pre-computed for your convenience and future reference. Download the folders in this into kcsd-python/figures/kCSD_properties folder, and the script from above should work.

(Wait 1hr after this post for it to finish uploading all the files)

m-kowalska commented 5 years ago

@ccluri Thank you for providing me with your data :)

In my fork (https://github.com/m-kowalska/kCSD-python/tree/master/figures/kCSD_properties) you can find scripts for figure 6, 7, and 9 with different error calculation (consistent with the formulation from the paper).