Clock Scan Protocol for Image Analysis: ImageJ Plugins

Summary

This paper describes two novel ImageJ plugins for 'Clock Scan' image analysis. These plugins expand the functionality of the original visual basic 6 program and, most importantly, make the program available to a large research community by bundling it with the ImageJ free image analysis software package.

Abstract

The clock scan protocol for image analysis is an efficient tool to quantify the average pixel intensity within, at the border, and outside (background) a closed or segmented convex-shaped region of interest, leading to the generation of an averaged integral radial pixel-intensity profile. This protocol was originally developed in 2006, as a visual basic 6 script, but as such, it had limited distribution. To address this problem and to join similar recent efforts by others, we converted the original clock scan protocol code into two Java-based plugins compatible with NIH-sponsored and freely available image analysis programs like ImageJ or Fiji ImageJ. Furthermore, these plugins have several new functions, further expanding the range of capabilities of the original protocol, such as analysis of multiple regions of interest and image stacks. The latter feature of the program is especially useful in applications in which it is important to determine changes related to time and location. Thus, the clock scan analysis of stacks of biological images may potentially be applied to spreading of Na⁺ or Ca⁺⁺ within a single cell, as well as to the analysis of spreading activity (e.g., Ca⁺⁺ waves) in populations of synaptically-connected or gap junction-coupled cells. Here, we describe these new clock scan plugins and show some examples of their applications in image analysis.

Introduction

The goal of this work is to present a Clock Scan protocol that is platform-free and freely available to any researcher interested in this type of image analysis. The Clock Scan protocol was originally developed in 2006¹, with the goal of improving existing methods of pixel intensity quantification within convex-shaped regions of interest (ROI), a method which has better integrative capacity and improved spatial resolution. During the acquisition, the protocol sequentially collects multiple radial pixel-intensity profiles, scanned from the ROI center to its border, or to a predetermined distance outside the ROI for the purpose of measuring the "background" pixel intensity. The protocol scales these profiles according to the cell radius, measured in the direction of the scan. Thus, the distance from the center to the ROI border of each individual radial scan is always 100% of the X scale. Finally, the program averages these individual profiles into one integral radial pixel-intensity profile. Because of scaling, the mean pixel-intensity profile, produced by the "Clock Scan" protocol, depends neither on the ROI size nor, within reasonable limits, on the ROI shape. This method allows direct comparison or, if required, averaging or subtraction of profiles of different ROIs. The protocol also allows correction of the integral pixel intensity profiles, of any object for background noise, by a simple subtraction of the average intensity of pixels located outside the object. Although it has only been tested in biological samples, our protocol provides a valuable addition to other existing image analysis tools used in studies of images of physical or chemical processes that are arranged around a point of origin (such as diffusion of substances from a point source)¹.

However, the major limitation of the original image analysis method was that the protocol was developed as a Visual Basic 6 (VB6) (code, and therefore, it was platform-dependent and difficult to distribute (requiring VB6). To address this problem and to join similar recent efforts by other investigators², we converted the VB6 Clock Scan program code into two Java-based plugins, compatible with the NIH-sponsored and freely-available open-source and platform-independent image analysis programs, ImageJ³ and Fiji ImageJ⁴. Furthermore, these plugins have now several new functions that expand the capability of the original protocol to process multiple ROIs and image stacks. Many image analysis applications are not user-friendly, with regards to performing statistical analysis of multiple objects, and thus, often only representative data are shown. With the multi Clock Scan ImageJ plugin, it is possible to facilitate the analysis of multiple objects simultaneously. A robust statistical evaluation of microscopy data, with regards to signal intensity distribution in single cells/objects, is now possible with this plugin extension. Here, we describe the Clock Scan plugins and show examples of their applications in image analysis.

Protocol

1. Software Installation

1. Install the latest versions of bundled Java and either ImageJ or Fiji ImageJ as recommended on the respective websites (see materials table for links to the corresponding websites). In the text below, both programs are referred to as "ImageJ".
2. Copy "Clock_Scan-1.0.1. jar" and "Multi_Clock_Scan-1.0.1.jar" plugin files using the link provided in materials table and paste them into the ImageJ plugin directory. Alternatively, use "Plugins | Install plugin" menu option to install these files after they have been saved on the computer hard drive.

2. Clock Scan analysis

1. Standard Clock Scan plugin (Figure 1):
   1. Use the ImageJ "File | Open" menu command to open an image of interest.
   2. Click on the 'polygon' tool, or 'segmented line selection' tool, and then draw on the image to outline the entire ROI or a segment of this region. See Figure 1A for an example of polygon selection (inner dashed outline).
      NOTE: Other selection tools, available in the software (rectangular, oval and freehand line selection), may also be used.
   3. Select "Plugins | Clock Scan" from the menu to open the standard clock scan protocol pop-up option window. Note that this command will also open the ROI Manager window with the outline automatically added to it.
   4. Use the plugin option window to do the following.
      1. Review and change the X and Y coordinates of the ROI center (automatically-calculated as coordinates of the physical mass center) by using scroll bars or changing the values in the corresponding input boxes. See Figure 1B.
      2. Depending on how much of the background region outside of the object should be covered by scanning, adjust the scan limits by using the "scan limit" scroll bar. See Figure 1A.
         NOTE: Scan limit is the fractional number representing how far the scan should proceed beyond the objects' border in any given direction; the default value is 1.20, indicating that the scan length will be 20% longer than the object radius in the direction of scan; see Figure 1A, outer dashed line).
      3. Modify the output of the plugin using "real radius", "subtract background", "polar transform" and/or the "plot with standard deviation" check-boxes.
      4. Click "OK" to run the plugin. See Figure 1C-H.
         NOTE: Examples of the output of the protocol with "plot with standard deviation" and "polar transform" or "real radius" and "polar transform" options selected are shown in Figure 1C and 1D and Figure 1E and 1F, respectively. Note that the calculated standard deviation (SD) values represent the variation between individual radial pixel intensity scans of the object. Also note the "ROI selection length" line in the plugin window, which displays the information on the ROI outline length measured in pixels.
   5. In the generated "Clock Scan Profile Plot", use the "List" command to plot values displayed in two, X and Y columns of data for grey scale images and in X and four Y columns of data for RGB images, of which Y0, Y1, Y2 and Y3 columns will be filled with integral and individual (red, green and blue) color channel pixel intensity values.
2. Multiple ROI Clock Scan plugin - working with multiple ROI (Figure 2):
   1. Open an image containing multiple ROI.
   2. Open the ROI Manager by clicking "Analyze | Tools | ROI Manager".
   3. Sequentially outline (see step 2.1.2) and add each ROI to the ROI Manager by clicking "Add" in the ROI Manager window; do this for all ROIs within the image. Use the "Analyze| Measure" command if ROI metrics are of interest.
      1. See Figure 2A for an example of multiple segmented line selections and Figure 2E for an example of multiple polygon selections.
   4. Select "Multi Clock Scan" in the "Plugins" menu to open the protocol options pop-up window.
   5. Use the protocol option window to do the following.
      1. If needed, reset the scan limit as per step 2.1.4.2; default value is 1.20.
      2. If needed, select the option to plot the mean clock scan profile with SD bars by checking the "Plot with standard deviation" box. See Figure 2C and D.
         NOTE: The calculated SD values will represent variation between integral clock scan profiles of different objects. Also, note the line in the plugin window displaying information on the "number of selected ROIs".
      3. Click "OK" to run the protocol.
   6. In the generated "Clock Scan Profile Plot", use the "List" command to plot the values displayed in the "Plot Values" window. See the "Multi Clock Scan Profile Plot" window legend for column designation by color channel.
   7. Note that the ROIs are numbered and their clock scan profiles for any given color channel are plotted in the same sequence in which the ROIs were outlined and added to the "ROI Manager".
3. Multiple ROI Clock Scan plugin - working with an image stack (Figure 3):
   1. Open an image-stack of interest.
   2. Open the ROI Manager by clicking "Analyze | Tools | ROI Manager".
   3. Outline the ROI of the images within the stack and add it to ROI manager as described in steps 2.1.2 and 2.2.3. Use the "Analyze | Measure" command if the ROI metrics are of interest.
   4. Select "Multi Clock Scan" in the "Plugins" menu to open the protocol options pop-up window.
   5. Use the protocol option window to do the following.
      1. Reset the scan limit as described in step 2.1.4.2; default value is 1.20.
      2. Select the option to plot the mean clock scan profile with SD bars by checking the 'Plot with standard deviation" box.
         NOTE: The calculated SD values will represent variation between different instances of the object selected in the image stack. Also, note the line in the plugin window displaying information on the "number of images in the stack".
      3. Click "OK" to run the protocol.
   6. In the "Clock Scan Profile Plot" window, click "List" to plot the values displayed in the "Plot Values" window, where the Y column number represents the image position within the stack - 1.

Results

The images that are used here for illustration purpose, are taken from databases created during our previous cell and tissue biological studies^5,6,7 and from the Allen Mouse Brain Atlas⁸. Both plugins were successfully tested using ImageJ 1.50i/Java 1.8.0_77, ImageJ 2.0.0-rc-44/1.50e/ Java 1.8.9_66 and Fiji ImageJ 2.0.0-rc54/1.51g/Java 1.8.0_66 program environment.

Discussion

Clock Scan Protocol: The Clock Scan protocol is a fast and simple tool of image analysis. The advantages of this protocol, compared to existing common approaches of image analysis (such as linear pixel intensity scans or calculation of mean pixel intensity of the ROI), have been described in details in previous publications^1,9. Briefly, this protocol allows the generation of integral radial pixel-intensity profiles by quantifying the intensity of...

Materials

Name	Company	Catalog Number	Comments
Computer	Any		compatible with software listed below
ImageJ or Fiji ImageJ	NIH	https://imagej.nih.gov/ij/ or https://fiji.sc/	bundled with Java 1.8 or higher
Clock-scan plugins	freeware	https://sourceforge.net/projects/clockscan/	Clock_Scan-1.0.1 jar and Multi_Clock_Scan-1.0.1/ jar
Origin 9.0	OriginLab	Northampton, MA, USA	This program was used to generate some graphs of the original Clock Scan data. Any other graphic software can be used to perform this function