Skip to content

Latest commit

 

History

History
129 lines (79 loc) · 4.25 KB

README.md

File metadata and controls

129 lines (79 loc) · 4.25 KB

imaging_decisionMaking_exc_inh

The scripts correspond to my postdoc project "Excitatory and Inhibitory Subnetworks Are Equally Selective during Decision-Making and Emerge Simultaneously during Learning", published in Neuron, 2019: https://www.ncbi.nlm.nih.gov/pubmed/31753580" (PDF file of the paper is available here: https://www.researchgate.net/profile/Farzaneh_Najafi4/research).

The data are available at CSHL repository: http://repository.cshl.edu/36980/

Codes to generate the Figures in the paper

Below, you can find a description of the codes to generate the Figures in the paper.
(Note: figure numbers may not match those in the final published paper; however, a description for each figure is provided below to help identify the figures.)

Figure 1

PMF:

farznaj/imaging_decisionMaking_exc_inh/behavior/PMF_allmice.m

FOV images:

fni18, 151217 (scale: 50 um: (50*512)/580])

PSTHs:

farznaj/imaging_decisionMaking_exc_inh/imaging/avetrialAlign_plotAve_trGroup.m

Inferred spikes:

farznaj/imaging_decisionMaking_exc_inh/utils/lassoClassifier/excInh_Frs.py

Heatmap of sorted neural activity:

fni17, 151015 avetrialAlign_plotAve_trGroup (in imaging_postproc.m, set mouse name, and run the imaging_prep_analysis section, stopping at avetrialAlign_plotAve_trGroup, line 1056)

FRs:

farznaj/imaging_decisionMaking_exc_inh/utils/laassoClassifier/excInh_FRs.py

Figure 2

ROC:

Set vars in: farznaj/imaging_decisionMaking_exc_inh/imaging/choicePref_ROC_exc_inh_plots_setVars.m

Example session AUC histogram: fni16, 151029 code gets called in line 340 of: choicePref_ROC_exc_inh_plotsEachMouse.m outcome2ana = 'corr'; %''; % 'corr'; 'incorr'; ''; doChoicePref = 0; %2;

Fraction choice tuned:

choicePref_ROC_exc_inh_plotsAllMice.m section: Fractions of significantly choice-tuned neurons ~line 340.

Example mouse absDevAUC time course: fni16 (doChoicePref = 2;

Example day AUC of corr, incorr: fni16; '151029_1-2' (last day, day 45) (doChoicePref = 0;) run choicePref_ROC_exc_inh_plots_setVars.m once with outcome2ana = corr, another time with incorr. Then run the first section of choicePref_ROC_exc_inh_plotsEachMouse (which calls choicePref_ROC_exc_inh_plotsEachMouse_corrIncorr)

ROC controlling for FR values of exc and inh:

run script: choicePref_ROC_exc_inh_plotsAllMice_sameFR

Fraction choice selective neurons; time course;

fni16 the following section Plot fraction choice-selective neurons averaged across days in code choicePref_ROC_exc_inh_plotsEachMouse

Figure 3

svm_excInh_trainDecoder_eachFrame_plots.py

panel B (example class accuracies): Figure: curr_chAl_day151015_exShfl3_171010-112112_sup.pdf fni17, 1 session (151015); average and st error across cross validation samples; for exc, only one example excitatory sample is used (exShfl3) so the error bar matches that of inh and allN.

Event time distributions: eventTimesDist.py

Weights: svm_excInh_trainDecoder_eachFrame_plotWeights.py

Supp Figure:

Corr, incorr (SVM trained on correct trials, tested on incorrect trials):

svm_excInh_trainDecoder_eachFrame_testIncorrTrs_plots.py Stimulus category decode: svm_excInh_trainDecoder_eachFrame_plots.py

Figure 4 (stability):

svm_excInh_trainDecoder_eachFrame_stabTestTimes_plots.py

Figure 5

Panel A: example session: fni16, 151029 PW correlations: corr_excInh_plots.m ; read comments on the top of the script for the scripts you need to run beforehand.

svm_excInh_trainDecoder_eachFrame_plots addNs_ROC = 1 set shflTrsEachNeuron first to 1, and run the code; then to 0, and run the code.

Use the following script for the plots of all mice (change in CA after breaking noise correlations): svm_excInh_trainDecoder_eachFrame_addNs1by1ROC_sumAllMice_plots

Figure 6

svm_excInh_trainDecoder_eachFrame_plots

Supp Figs:

Running and licking:

tracesAlign_wheelRev_lick_classAccur_plots.m Fraction choice selective for early and late days: choicePref_ROC_exc_inh_plotsAllMice.m (at the end of the script)

Example sessions for classification accuracy (time course): svm_excInh_trainDecoder_eachFrame_plots.py fni17, sessions: 151014 – 151029 – 151022 – 151008 – 151020 - 150903 (not used, but pretty good: 151026 – 151021 – 151013 – 150918)

Temporal epoch tuning:

temporalEpochTuning_allSess.m