buzsakilab / buzcode Goto Github PK

@dlevenstein , any ideas?

The example structs in buzcode have problems, for example, the struct inside the mat files has a name that doesn't match the mat file. Also, there is no example of any actual data, e.g. a spike train. Dan L. asked me to open this issue, but it will be helpful for me also because I want to polish the buzcode<->TStoolbox conversion scripts.

MakeClassicFet.m

SleepScoreMaster output figures need to be updated.
-Yuta had bug where figure marked states did not match the actual output states
-Clustering figure is not created
-Channel selection figure could be a single figure

-Bring the documentation up to date while you're at it.....

different batches of neuronexus probes (of the same design) have different channel maps. This needs to be accounted for in the file naming (See NRX_Buzsaki64_8X8_H64LP.xlsx as a working example)

If you could include in the FindRipples to create a rip.evt file that can be loaded into Neuroscope to manually check the detection, it would make analysis it a lot smoother for me.

assumes int16 for all files

There are a suite of interval utilities, i.e. InIntervals, FindInIntervals, RestrictInts, which have somewhat overlapping functionality.

1. A census/guide should be created of these functions and any that can be merged (maintaining all functionality) should be merged

2. Calls to these functions within buzcode will need to be updated to reflect any mergers.

This way we could remove 'noisy' segments of recording that often mess up the scoring results...

To do this, we'll probably need to add functionality to LoadBinary.m as well, to load interval sets

LoadParameters has moved far enough away from the original FMA function that is should be renamed bz_LoadParameters and all calls in buzcode should be updated.

bz_getSessionInfo was accidentally deleted, so calls to bz_getSessionInfo were replaced with LoadParameters.

related to issue #86

This is really more of a place to start discussion than code production. Currently, we have two places where metadata is stored at the individual session level (animal metadata is duplicated elsewhere) 1) the .xml and 2) the *.sessionInfo.mat file. Going forward we should establish a single internal lab format that is easy for newcomers to adopt and extensible to new data types (imaging, opto, patch, etc).

Some things we should figure out:

• Should entry be text or GUI based?
• Do we store the data in a xml or .mat file?
• Should we have different priorities for different parts of the metadata (i.e. required, suggested, optional)?
• others?

check that all functions reflect the new behavior format

pos2behav.m
bz_plottrials.m

create a new function to convert original behavior to behavior 2.0 format.

It may be better to have naming convention be

baseName.events.eventsName.mat
baseName.cellinfo.cellinfoName.mat

etc etc....

that way when you browse basePath, all your events/cellinfo/etc files are next to each other.

Could be implemented with a single conversion script or via the I/O functions.

Thoughts? too late? worth it?

While some data types have conversion functions (i.e. bz_GetSpikes), others still need to be created made and conventions need to be decided upon.

Script should check if .spk files exist, if so then there is no need to read spikes from .dat file to resave them

I'd like to propose some folder restructuring.

1. that we remove the statistics folder (currently empty) and statistical tests can be put in the utilities folder or the relevant analysis subfolder.
2. we rename the plotting folder to visualization, and it can subsume the eventexplorer tool @dlevenstein has been working on.
3. buzcode/preprocessing/fileConversions/ moves to buzcode/utilities/fileConversions/

Thoughts?

I am trying to test PhaseModulation, using

PhaseModulation([1,2,3,4,5],linspace(0,6,5*1250),[5 10])

in octave and I am coming across the following error:

error: element number 2 undefined in return list
error: called from
    Filter at line 123 column 10
    FilterLFP at line 87 column 10
    PhaseModulation at line 85 column 13

The implicated line is [b a] = fir1(order,passband/nyquist); in
externalPackages/FMAToolbox/General/Filter.m where the code expects 2 outputs from fir1. That function only gives one output argument in Octave's signal toolbox and in MATLAB's most recent release. Is this code expecting an earlier version of this function? In the current release, this can be fixed by setting a = 1

	case 'fir1',
		a = 1;
		if ~isempty(passband),
			if passband(1) == 0,
				b = fir1(order,passband(2)/nyquist,'low');
			else
				b = fir1(order,passband/nyquist);
			end
		else
			b = fir1(order,stopband/nyquist,'stop');
		end

This error can also be avoided by taking out the optional argument specifying fir1 in PhaseModulation line 85:

fil = FilterLFP(lfp,'passband',passband);

SetCurrentSession is the heart of FMA and the current buzcode version is named the same but breaks backwards compatibility. I recommend renaming the buzcode version or fixing the backwards compatibility.

Let's get preliminary matlab functions uploaded that can read from the NYU share and update the database as new recordings are added, as well as query the DB for subsets of data.

relevant tools can be put here:
https://github.com/buzsakilab/buzcode/tree/master/database

seems to be for any input value for the 'smooth' parameter

-isCellinfo
-isBehavior
-isState
-isPopInfo
-isLFP
-isEvent
-etc, etc

One feature that I rely heavily on in the code I'm using is restricting the ripple detections to NREM sleep from the states.mat file. And as a side note, I don't use the output from the automated sleep scoring (epileptic lfp doesn't do so well with the automated detections) so I'm still using the states.mat file from TheStateEditor.

a reminder for me, or anyone interested, to rework this function into buzcode format at some point

A prompt is opened for .regions and other fields even when the 'noprompts' field is set to true.

-Update all the code to load the LFP with bz_GetLFP, and to use the buzcode-style lfp structure throughout. This has the benefit that the sampling frequency is associated with the lfp structure (in lfp.samplingRate), and so nothing has to be hard coded but can use whatever sf the lfp is.
-Update bz_GetLFP to have an optional input ‘targetSamplingRate’.

input arguments
-default is all neurons from recording
-by UID
-by shank/channel group
-by region (metadata file required?)

additional optional inputs

• mean waveforms
• max amplitude channel #
• ACG's

Would make sense to start with GetSpikeTimes.m for this

reminder to add warning and deprecate MTCoherence.m

@dlevenstein, could we move wavespec.m into buzcode? there are a few functions that call it

move all functions to timestamps naming conventions

if anyone wants to take a look and fix this, feel free :) I don't have time at the moment

functions to check: (delete once a pair has been checked/merged, close issue when all are crossed off)

./externalPackages/fieldtrip-20150209/external/eeglab/sobi.m
./externalPackages/eeglab14_0_0b/functions/sigprocfunc/sobi.m

./externalPackages/fieldtrip-20150209/external/eeglab/jader.m
./externalPackages/eeglab14_0_0b/functions/sigprocfunc/jader.m

./externalPackages/fieldtrip-20150209/external/eeglab/binica.m
./externalPackages/eeglab14_0_0b/functions/sigprocfunc/binica.m

./externalPackages/fieldtrip-20150209/external/eeglab/COPYING
./externalPackages/eeglab14_0_0b/eeglablicense.txt

./externalPackages/fieldtrip-20150209/external/eeglab/floatwrite.m
./externalPackages/eeglab14_0_0b/functions/sigprocfunc/floatwrite.m

./externalPackages/fieldtrip-20150209/external/eeglab/varimax.m
./externalPackages/eeglab14_0_0b/functions/miscfunc/varimax.m

./externalPackages/fieldtrip-20150209/external/eeglab/binica.sc
./externalPackages/eeglab14_0_0b/functions/sigprocfunc/binica.sc

./externalPackages/fieldtrip-20150209/external/eeglab/ica_linux
./externalPackages/eeglab14_0_0b/functions/resources/ica_linux

./externalPackages/fieldtrip-20150209/external/eeglab/floatread.m
./externalPackages/eeglab14_0_0b/functions/sigprocfunc/floatread.m

./externalPackages/fieldtrip-20150209/external/eeglab/pcsquash.m
./externalPackages/eeglab14_0_0b/functions/miscfunc/pcsquash.m

/externalPackages/fieldtrip-20150209/external/neuroscope/private/isradians.m
./externalPackages/FMAToolbox/Helpers/isradians.m

./externalPackages/fieldtrip-20150209/external/neuroscope/private/wrap.m
./externalPackages/FMAToolbox/Helpers/wrap.m

./externalPackages/fieldtrip-20150209/external/neuroscope/LoadSpikeAmplitudes.m
./externalPackages/FMAToolbox/IO/LoadSpikeAmplitudes.m

./externalPackages/eeglab14_0_0b/sample_locs/GSN-HydroCel-32.sfp
./externalPackages/fieldtrip-20150209/template/electrode/GSN-HydroCel-32.sfp

/externalPackages/eeglab14_0_0b/plugins/dipfit2.3/standard_BEM/skin/standard_skin_1222.vol
./externalPackages/fieldtrip-20150209/template/headmodel/skin/standard_skin_1222.vol

./externalPackages/eeglab14_0_0b/plugins/dipfit2.3/standard_BEM/skin/standard_skin_14038-1.bps
./externalPackages/fieldtrip-20150209/template/headmodel/skin/standard_skin_14038-1.bps

./externalPackages/eeglab14_0_0b/plugins/dipfit2.3/standard_BEM/skin/standard_skin_5054-1.bps
./externalPackages/fieldtrip-20150209/template/headmodel/skin/standard_skin_5054-1.bps

./externalPackages/eeglab14_0_0b/plugins/dipfit2.3/standard_BEM/skin/standard_skin_1222-1.bnd
./externalPackages/fieldtrip-20150209/template/headmodel/skin/standard_skin_1222-1.bnd

./externalPackages/eeglab14_0_0b/plugins/dipfit2.3/standard_BEM/skin/standard_skin_5054-1.bpl
./externalPackages/fieldtrip-20150209/template/headmodel/skin/standard_skin_5054-1.bpl

./externalPackages/eeglab14_0_0b/plugins/dipfit2.3/standard_BEM/skin/standard_skin_1222-1.bps
./externalPackages/fieldtrip-20150209/template/headmodel/skin/standard_skin_1222-1.bps

./externalPackages/eeglab14_0_0b/plugins/dipfit2.3/standard_BEM/skin/standard_skin_14038.vol
./externalPackages/fieldtrip-20150209/template/headmodel/skin/standard_skin_14038.vol

./externalPackages/eeglab14_0_0b/plugins/dipfit2.3/standard_BEM/skin/standard_skin_14038-1.bnd
./externalPackages/fieldtrip-20150209/template/headmodel/skin/standard_skin_14038-1.bnd

./externalPackages/eeglab14_0_0b/plugins/dipfit2.3/standard_BEM/skin/standard_skin_5054-1.bnd
./externalPackages/fieldtrip-20150209/template/headmodel/skin/standard_skin_5054-1.bnd

./externalPackages/eeglab14_0_0b/plugins/dipfit2.3/standard_BEM/skin/standard_skin_14038-1.bpl
./externalPackages/fieldtrip-20150209/template/headmodel/skin/standard_skin_14038-1.bpl

./externalPackages/eeglab14_0_0b/plugins/dipfit2.3/standard_BEM/skin/standard_skin_1222-1.bpl
./externalPackages/fieldtrip-20150209/template/headmodel/skin/standard_skin_1222-1.bpl

./externalPackages/eeglab14_0_0b/plugins/dipfit2.3/standard_BEM/skin/standard_skin_5054.vol
./externalPackages/fieldtrip-20150209/template/headmodel/skin/standard_skin_5054.vol

./externalPackages/eeglab14_0_0b/plugins/dipfit2.3/standard_BEM/elec/standard_alphabetic.elc
./externalPackages/fieldtrip-20150209/template/electrode/standard_alphabetic.elc

./externalPackages/eeglab14_0_0b/plugins/dipfit2.3/standard_BEM/elec/standard_prefixed.elc
./externalPackages/fieldtrip-20150209/template/electrode/standard_prefixed.elc

./externalPackages/eeglab14_0_0b/plugins/dipfit2.3/standard_BEM/elec/standard_primed.elc
./externalPackages/fieldtrip-20150209/template/electrode/standard_primed.elc

./externalPackages/eeglab14_0_0b/plugins/dipfit2.3/standard_BEM/elec/standard_1020.elc
./externalPackages/fieldtrip-20150209/template/electrode/standard_1020.elc

./externalPackages/eeglab14_0_0b/plugins/dipfit2.3/standard_BEM/elec/standard_postfixed.elc
./externalPackages/fieldtrip-20150209/template/electrode/standard_postfixed.elc

./externalPackages/eeglab14_0_0b/plugins/dipfit2.3/standard_BEM/elec/standard_1005.elc
./externalPackages/fieldtrip-20150209/template/electrode/standard_1005.elc

just found a potential issue, when given channel #'s that aren't actually in the recording for 'SWChannels' or 'ThetaChannels', sleepscoremaster doesn't throw an immediate error but fails later on when picking channels.

Should we add an input checking at the beginning to verify the channel inputs actually exist?

all the struct type checking should also verify if a file has been manually examined by a human

currently bz_LFPfromDat just downsamples. Should also low pass filter like NDManager...

instead of writing to nohup.out logfile, this should be put in the klg file...

a single channel is selected for each shank/spike group (0-indexed neuroscope format) and sent directly to Loadbinary.m which takes 1-indexed inputs. added 1 to channel list and pushed.

(this is also a test of the slack/code-dev integration)

we'll need to fix the code so that it can parse multiple XML files and select the correct one. Is there a more general function we could write the parses XML files (bz_getSessionXml) or something?

A warning that this function isn't working, will look/fix post SFN