Data for Visualization of a distributed synaptic memory code in the Drosophila brain
This data repository holds the raw data traces of the bouton responses presented in Bilz et al "Visualization of a distributed synaptic memory code in the Drosophila brain". We analyzed synaptic boutons of Kenyon cells of the Drosophila mushroom body γ-lobe, two to five by expressing a fluorescent Ca2+ sensor in single Kenyon cells so that axonal boutons could be assigned to distinct cells. Single files in this data set represent single cells in the described experiments. The data set consists of all cells included in the publication and is devided into a raw dataset presented here as xlsx files and an analysed data set as MatLab files. The Matlab scripts used to analyse the data can be found here: https://github.com/zerotonin/KCC_KenyonCellCorrelator . Data organization of the xlsx files =========================== Each file contains the responses of one cell. The data inside the file is split onto two sheets: Sheet 1 contains pre learning phase data, Sheet 2 contains post learning data. The data on each sheet is saved as a two mxn dimensional matrix, where m represents the number of acquired frames and n the number of identified boutons on the cell. The first row of the sheet contains stimulus inforation. All xlsx files can be found in xlsxData.zip Data organization of the MatLab files ============================== The data organization of the Matlab files is described in "Steps to reproduce" .
Steps to reproduce
Data organization of the MatLab files ============================== Matlab files already contain analyzed data. Therefore the result of the analyses is saved in multiple variables, which are explained in detail as follows: data & dataF -------------- data is a mxnx4x2 four dimensional matrix consisting of floating point numbers and is identical to the xlsxFile. m is the number of acquired frames. n is the numberof boutons. The third dimension contains the responses to the four odors in the succession: MCH, 3-Oct, 1-Oct and, MineralOil. The fourth dimension is (1)pre training or (2) post training. This is the basic organisation of many of the following variables. dataF is the filtered version of data as we did not use the filtering option this variable is idententical to the data variable. dataFM -------- This is the median values per y-lobe 2-5. The data is arranged as follows: mx4x4x2 four dimensional matrix consisting of floating point numbers and is identical to the xlsxFile. m is the number of acquired frames. Columns are the four y-lobes: y2,y3,y4 and, y5. The third dimension contains the responses to the four odors in the succession: MCH, 3-Oct, 1-Oct and, MineralOil. The fourth dimension is (1)pre training or (2) post training. ylobesIDX ----------- ylobesIDX is a column vector of integers (2 to 5), holding the information in which y-lobe the bouton was found. corrMat -------- corMat is a mxmx4x2 matrix of floating point numbers, where m is the number of boutons. The third dimension contains the responses to the four odors in the succession: MCH, 3-Oct, 1-Oct and, MineralOil. Each entry represents the cross correlation coefficient normalised to the auto-correlation of this bouton to all other boutons. It is the basis for the amplitude corrected correlation index (ACC). lobesCorr ----------- lobesCorr is the median correlation coefficient (norm. by the auto correl.) of each y-lobe to each other y-lobe. Therefore the matrix is 4x4x4x2 with the first two dimensions being y-lobes and the second to being odors and training situation. amps & ampsN ----------------- amps is a 4x4x2 matrix where the rows represent the stimuli in the described order and colums the y-lobes. Third dimension is pre and post training. Each value is the median amplitude of this ylobe. ampsN is the normalized version of amps. similarity ---------- Similarity holds the ACC index organised as in lobesCorr. simDiff -------- simDiff is the difference of the ACC-index between pre and post training situations (post values-pre values). Therefore the fourth dimension is dropped and the matrix is 4x4x4. The first two dimension coding for y-lobes, the third for odors.