Automated Sholl Analysis of Digitized Neuronal Morphology at Multiple Scales

This video demonstrates a procedure to analyze the morphology of dendrites and axons, collectively known as neurites, using a semi-automated program in order to observe local changes. First neuron J software is used to trace eight bit TIFF images of neurons to identify the position of neurite segments. Neuron studio software is then used to define structural information about the connectivity between neurites.

Finally, the bonfire MATLAB program is run to extract morphological data from the cells, including shoal analysis, as well as the number of neurites branch points and neurite tips. Results show local changes in neurite morphology based on the performance of shoal analysis on sub regions of the neurotic arbor. The main advantage of this technique over existing methods, such as shoal analysis by hand, is that the program is semi-automated, which greatly increases the efficiency of analysis and allows the data to be audited so that the accuracy of the analysis can be verified.

Demonstrating the procedure will be Hirsch AWA undergraduate student working in the Firestein lab for the summer. In order to use the bonfire program to analyze the morphological characteristics of the Neurites eight bit tiff, images of individual neurons must be obtained after downloading the bonfire program. The program settings must be adjusted based on the image resolution of the images that you wish to analyze in the bonfire parameters portion of the bonfire program replace the current value for the variable picks conversion with the value of the image resolution of your images.

In order for bonfire to analyze the data, the files must be organized in a specific structure, including a master folder, a bonfire folder containing the bonfire MATLAB M files, sub folders containing each of the different conditions and the cell image files inside the different condition folders begin by preparing the image for tracing. For this experiment, only the dendrites will be traced in order to analyze changes in dendritic morphology. Begin by preparing the image for tracing.

Open the image by selecting the open button on the neuron J toolbar and select the image you would like to trace. Now, resize the image by selecting the maximize button to adjust the brightness and contrast of the image so that you can visualize all of the Neurites select image on the neuron J toolbar, and then select adjust brightness contrast. To begin tracing, select the add traces button on the neuron J toolbar Trace around the perimeter of the cell body.

Then select the label tracings button on the neuron J toolbar and select N one from the tracing ID dropdown menu, select type zero six in the neuron J attributes window and select Okay. To begin tracing the neurites, select the add traces button on the neuron J toolbar and add a trace along each neurite branch of interest. It is best if the segments you draw stop at each branch point with each daughter branch point, starting at that point as a new trace.

Select the save tracings button on the neuron J toolbar and save the tracing you have just created in the same folder as the original image file. Next, export the trace files and the trace identifier files from neuron J.Select the export trace button on the neuron J toolbar. Now select the tab delimited text files separate file for each tracings option on the neuron J export dialogue box and select okay allow neuron J to choose the names of the files and the save location.

Select the measure tracings button on the neuron J toolbar, and then select the display tracings measurements option on the neuron J measurements window and select run Select file in the neuron J tracings window, and then select save as and save the file. The file name must exactly match the name of the original image file, followed by underscore info and should not contain a three letter file extension. Check that your computer does not automatically add a XLS file extension.

If it does, the file extension must be manually deleted. First, reorganize the folders using bonfire load by opening matlab, and clicking on the triple dot button in the upper right of the command window. Select the bonfire folder in your master folder in the browse for folder window type bonfire load into the command window and press enter.

Select the condition folder that you wish to analyze in the browse for folder, window, and select.Okay. This will reorganize the folder structure by creating cell sub folders containing all of the data for each individual cell. Now to create file name underscore premium swc files type bonfire underscore NDF two s WC into the command window and press enter.

Select the same condition folder that was just reorganized with bonfire load in the browse for folder window, and select.Okay. This will create a WC file in each cell folder for the chosen condition. Each cell folder will contain five files.

The original TIF image, the NDF file, the underscore info, trace identifier file, a TXT file, and a swc file. Before using Neuron Studio, we recommend that you become acquainted with the program's features and shortcuts. They have an excellent online user manual, and knowing the keyboard shortcuts will save you a great deal of time.

To begin open the neuron studio program and select file open on the neuron studio toolbar, find the TIF image of the neuron. You want to edit and open it. Select run settings and enter one into each of the three boxes.

In the voxel size window, select file import s wc. Select the appropriate s WC file that matches the file you want to work with. The image file will now be overlaid with the trace image.

The cell soma should be overlaid with a red circle to link neurites. Use the neurite tool to join nodes so that each branch point represented by yellow nodes can only create two branches. All traces must be continuous with the soma.

Now to export data from Neuron Studio, select file, save neurites, and save as the default name to check for errors in the swc files type bonfire trace. Check into the command window and press enter. Select the condition folder containing the data that was just processed in the browse for folder window and select okay if there are any errors in any of the images in the folder.

The program will output images displaying where the error is located to extract morphological data from the dot S WC files. Type bonfire into the command window and press enter. Select the condition folder that you want to analyze in the browse for folder window, and select okay for each of the neurons of interest.

The bonfire analysis will generate a graph of neuronal morphology along with the shoulder rings used in the analysis. Additionally, the bonfire command will generate the MAT file that contains all of the morphological information that has been taken from the analysis. To view the preliminary graphs of the data type bonfire results into the MATLAB command window, select the condition folder for the condition to view in the browse for folder window and select.Okay.

The browse for folder window will close temporarily and reopen allowing additional conditions to be selected. Once the conditions have been selected, select cancel to exit selection process. Bonfire result will return summary charts including the data from the condition folders you selected to export the morphological data into Excel type bonfire Export into the MATLAB command window.

Select the condition folder containing the data you would like to export in the browse for folder window and select. Okay, bonfire export will create Excel files of the morphological data and will place them in the condition folder that was selected. Now we'll show you an example of the data generated by the bonfire program on a data set containing two conditions.

These are example inverted images of both conditions. It can be seen from these images that condition one contains more dendrites than condition two. This phenomenon can also be observed in a variety of graphs that are generated by the bonfire program.

Because the bonfire program performs its analysis on sub regions of the dendrites, the increased dendrites in condition one can be readily identified, plotting the shoal curve of the total dendritic arbor shows that there are more dendrites distal to the cell body for condition one neurons plotting the average number of branch points and terminal points per cell shows that condition one contains both more branch points and more terminal points than condition two. The average number of processes per cell is shown for primary, secondary, tertiary, or higher order dendrites. There are more tertiary and higher order dendrites for condition one plotting the average number of processes per cell for root, intermediate, and terminal dendrites shows that condition one has more intermediate or terminal processes than condition two.

Shown here are the segment identity specific shoal analysis curves in which the segments are grouped as primary, secondary, or tertiary or greater. The increase in tertiary dendrites occurs distal to the cell body for condition one neurons. The segment identity specific shoal analysis curves are shown in which the segments are grouped as root segments, intermediate segments, or terminal segments.

The increase in both intermediate and terminal dendrites in condition one occurs distal to the cell body. After watching this video, you should have a good understanding of how to perform shoal analysis using the bonfire program in order to observe local changes in neuro morphology.