To begin, initiate MATLAB and execute the EZcalcium toolbox to access the initial GUI. Within the initial GUI, select Motion Correction to open the motion correction GUI. Use the Add Files option to upload a TIF file containing the imaging data.
Next, set the non-rigid motion correction to blank, upsampling factor to 50, maximum shift to 15, initial batch size and bin width to 200. Click on Run Motion Correction to initiate the correction. Within the initial GUI, activate automated ROI detection to access the ROI detection GUI.
Use the Add Files feature to import the motion-corrected imaging data. Set initialization to greedy, search method to ellipse, deconvolution to constrained FOOPSI-SPGL1, and autoregression to decay. Then set the estimated number of ROIs to 60.
Assign the estimated ROI width to 17, merge threshold to 0.95, fudge factor to 0.95, spatial and temporal downsampling to one, and temporal iterations to five. Then click Run ROI Detection to initiate the detection process. In the initial GUI, select ROI Refinement to launch the ROI refinement GUI.
Use the Low Data button to import ROI data. Select the ROIs with low activity frequency situated beneath the skull or those overlapping with other neurons/neurites. Click Exclude ROI to exclude those ROIs from subsequent analysis.
Calculate the delta F by F values using this equation. Choose XLSX as the data export format and execute export data to generate an Excel file populated with the raw delta F by F values. Compute the Pearson's correlation coefficient for the delta F by F values among the ROIs and construct a matrix of the correlation coefficients.
Use Fiji software to delineate the barrel boundaries from the TCA-RFP image. Then assign the ROIs to their corresponding barrels or septa. Compare the pairwise correlation coefficients within identical barrels and distinct barrels.
Generate 1, 000 to 10, 000 surrogate datasets by randomly permuting the association between ROI positions and calcium ion traces. In each surrogate dataset, compute the mean correlation coefficient within barrels individually and determine the statistical significance of the correlation. Higher pairwise correlation coefficients were observed within the same sensory processing units than across different units.
Activities demonstrated stronger synchrony within the same units despite longer distances, surpassing the correlation with physically closer neurons from different units. The average of correlation coefficients within the same barrels was significantly higher than that calculated from 10, 000 sets of surrogate data. The correlation within the same barrels was significantly strong among three different time windows.
