Behavioral and 2P Calcium Imaging Data for "A neural circuit architecture for rapid learning in goal-directed navigation"

Behavioral and 2P calcium imaging data for the following paper:Dan C, Hulse BK, Kappagantula R, +Jayaraman V & +Hermundstad AM (2024), A neural circuit architecture for rapid learning in goal-directed navigation.Data files for different groups of flies:Preproc_*_FlightBehaviorData.mat: preprocessed behavioral dataPreproc_*_CAImaging_FlightBehaviorData.mat: preprocessed behavioral data and raw calcium imaging dataSegmented_*_FlightBehaviorData.mat: processed behavioral data segmented into fixations and saccadesProcJumpData_*.mat: processed Calcium imaging data to track jumps in the compass heading over timeFile descriptors:Punishment ('Laser'; 'NoLaser'): indicates whether or not flies underwent laser trainingGenotype ('WT', 'SS00090 Parent/Kir', 'SS00096 Parent/Kir', '60D05')File organization:Each file contains MATLAB data structures indexed by fly i and trial jPreProcData*.fly(i).trial(j) contains the following fields:x: frame position of visual scene in arena (can take values from 0 to 95, which span 360deg)wba: difference in wing-beat amplitude between left and right wingswbf: wingbeat frequencytime: time in secondslaser: binary number indicating whether laser punishment was deliveredpuff: binary number indicating whether an airpuff was deliveredtrialType: 'naive', 'train', or 'probe'bgsubF: background-subtracted fluorescence (only for 60D05 flies)SegData*.fly(i).trial(j) contains the following fields:saccades: structure containing properties of saccadesfixations: structure containing properties of fixationsdrift: structure containing properties of drift periods that were not classifies as either fixations or saccadesPI: performance score for trialPIinds: indices of timepoints used to compute PI scores and other behavioral measuresnlaser: number of timepoints at which laser punishment was delivered.residency: number of timepoints spent at each of 48 frame positions (i.e., collapsed across the two-fold symmetry of the visual scene)shifted: data structure containing data aligned to the preferred heading of each fly, defined over the 48 frame positions in the collapsed span of the arenaresidencyFull: number of timepoints spent at each of 96 frame positionsshiftedFull: data structure containing data aligned to the preferred heading of each fly, defined over all 96 frame positionstSeriesProcAll(i).jumpData(j) contains the following fields:dffMat: fluorescence for 32 wedges in the ellipsoid body, tracked across timearena: orientation of visual scene in arena, defined from -pi to pi radianscorrectedPVA_EB: original population vector average (PVA) of activity in the EB after correcting for multiple offsetsorigPVA_EB: original population vector average (PVA) of activity in the EB, without correcting for multiple offsetsPVA_EB_strength: strength of corrected PVAFRmaps.FRMapHeading: average fluorescence in each of 32 wedges of the EB for each of 48 orientations of the visual scene (i.e., collapsed across the two-fold symmetry of the visual scene)bumpjumps: structure containing properties of offset between the compass bump and the visual scenethe following fields are copied over from the data structures above:saccades: structure containing properties of saccades (copied from SegData*.fly(i).trial(j))fixations: structure containing properties of fixations (copied from SegData*.fly(i).trial(j))setpoint: preferred arena heading, defined from 0 to 95 pixelsPIinds: indices of timepoints used to compute PI scores and other behavioral measuresx: frame position of visual scene in arena (can take values from 0 to 95, which span 360deg)wba: difference in wing-beat amplitude between left and right wingswbf: wingbeat frequencytime: time in secondstrialType: 'naive', 'train', or 'probe'Associated code:Code for FPGA camera-based wingbeat analyzer is available at: https://github.com/ChuntaoDan/FPGA_WingbeatAnalyzer (DOI: 10.5281/zenodo.11060747)Code for all behavioral analyses, calcium imaging analyses, modeling, and plotting is available at https://github.com/HermundstadLab/flyVisualLearning (DOI: 10.5281/zenodo.11062515)