providing it the data directory whose report must be created. It will
generate a report as a single html file which can be easily shared with
collaborators via email or dropbox. A link to the report generated from
the MNE sample dataset is provided below:
The motivation is that it can be used to generate a quick
summary/statistics of the results and do a quality check with minimal
scripting. Typically, this is something one might want to run one day
after data collection.
We also provide a python interface for more advanced users who want to
embed their own custom plots in the report.
Any feedback towards improving this is most appreciated. Thanks!
I am already really happy to see what Mainak has achieved but if you
think some relevant information is missing or not well highlighted,
it's the moment to ask so we can make it happen for the next release.
Please take a minute to give some feedback and suggest improvements.
Remove the full path from the names of each file (just put the path at
the top of the report).
No space between the colon and file type:
Events : MNE-sample-data/MEG/sample/ernoise_raw-eve.fif -> Events:
ernoise_raw-eve.fif
By default don't include events with value of 0 (maybe include it as
an option). In most cases these are not true events.
The covariance matrix scaling for gradiometers in the sample file seems flat.
At least an option to force a scaling factor for each file type
(because it can be confusing for example that the left and right
visual evoked examples have different scalings for the gradiometers).
It would be most helpful to split the Trans plot into multiple plots.
One with just the outer_skin (opaque) and array (transparent) (I would
prefer a sagittal view perhaps a view from the front is helpful too).
Another one to five plots (multiple angles) of the outer skin
(transparent) with the digitized points (scaled to indicate distance
from the surface).
Paring down the amount of information shown in the Raw examples would
be helpful.
I would also like to see a set order to the file (preferably based on
a default flow for example: raw, -ave, -cov, trans, fwd, inv).
I would love to see the -eve file information included in the
information with the raw information (in the same section). (even in
cases where there is no -eve file (as find events can get this)).
I'd also like to see some of the names changed to make a better
association with the actual filenames/filenaming conventions (this may
prove to be controversial). For example Evoked -> Average (these are
-ave.fif). Instead of MRI at the top right (maybe BEM).
But I would be happy to use this as is. It looks very good!
By default don't include events with value of 0 (maybe include it as
an option). In most cases these are not true events.
the events are actually 1 not 0. This remark suggests me that we
should not use standard y ticks/scaling for plot_events
The covariance matrix scaling for gradiometers in the sample file seems flat.
I this it's correct. It's empty room with SSP applied.
At least an option to force a scaling factor for each file type
(because it can be confusing for example that the left and right
visual evoked examples have different scalings for the gradiometers).
It would be most helpful to split the Trans plot into multiple plots.
One with just the outer_skin (opaque) and array (transparent) (I would
prefer a sagittal view perhaps a view from the front is helpful too).
Another one to five plots (multiple angles) of the outer skin
(transparent) with the digitized points (scaled to indicate distance
from the surface).
we can indeed do better. Maybe you can take a look at the plot_trans
function and improve it a bit?
Paring down the amount of information shown in the Raw examples would
be helpful.
+1
I would also like to see a set order to the file (preferably based on
a default flow for example: raw, -ave, -cov, trans, fwd, inv).
good point. +1
I would love to see the -eve file information included in the
information with the raw information (in the same section). (even in
cases where there is no -eve file (as find events can get this)).
I like this too.
I'd also like to see some of the names changed to make a better
association with the actual filenames/filenaming conventions (this may
prove to be controversial). For example Evoked -> Average (these are
-ave.fif). Instead of MRI at the top right (maybe BEM).
I suggested Evoked and MRI reason is that MRI can be displayed
without BEM and Evoked can contain other things than average (std dev
for example).
But I would be happy to use this as is. It looks very good!
I second Alex and Dan's comments? great job Mainak! I'd certainly use the current version as is. But just to make it even better, on top of the previous comments, it might be nice to see a planar plot of the evoked (averaged) response?
Thanks to everyone for pitching in with their valuable comments. I will
take them into account for the mne report. If there is anything else that
pops up later, please feel free to reply to this email thread.
reason is that MRI can be displayed