Name: quantifying intermembrane distances with serial image dilations
Uploaded: 2018-09-28T13:30:00
Duration: 7 min 45 s
Description: Watch this Scientific Journal Video about Quantifying Intermembrane Distances with Serial Image Dilations at JoVE.com

The authors would like to thank Kathy Lowe at the Virginia-Maryland College of Veterinary Medicine for processing and staining TEM samples. 
Funding: 
National Institutes of Health R01-HL102298 
National Institutes of Health F31-HL140873-01

NOTE: Software required are ImageJ (or similar image-modifying software) and Matlab R2015. The user may encounter compatibility issues with other versions of Matlab.
1. Pre-Processing Images
<ol>
	<li>For any grayscale image, ensure the maximum intensity value of any given pixel is &#60;255. This is typically done by subtracting a value of 1 from the image in the custom Matlab program &#34;ImageSub.m&#34;, included in supplemental file S1.</li>
</ol>
2. Outlining the Perinexus
<ol>
	<li>Outline the perinexus in ImageJ or other image-processing software.</li>
	<li>Ensure that the outline is one pixel thick and is set to the highest intensity value in an image (255 in a grayscale image from 0 to 255).</li>
	<li>Identify the GJ by its pentalaminar structure11,12, and define the beginning of the perinexus as the point at which the two opposing cell membrane bilayers diverge, as shown in Figure 1A. Begin ~200 nm from the edge of the GJ, tracing along the inner membrane of the first cell and back along the inner membrane of the second cell. In ImageJ, release the pen to automatically close the outline. This artificial closing will be cropped out later. 
	NOTE: It is crucial to outline the perinexus with great care at as high a magnification as possible, as even small mis-applications of the outline can result in several nanometers of error in the final measurement.</li>
</ol>
3. Setting up Algorithm and Selecting Perinexus of Interest
NOTE: The pathfinding algorithm requires the AutoGraph, Edge, Graph, Node , and Pathfinding functions9 to be in the same directory as the MembraneSepDist m-file. All files can be found in supplemental file S1.
<ol>
	<li>Select save locations for data and figures. These are currently hard-coded into the m-file. 
	NOTE: The first line of the program is a function to clear all variables, close all windows and clear the command window. Save any desired variables or figures before running the m-file. 
	NOTE: Software screenshots are included in supplemental file S2 for all hard-coded values.</li>
	<li>Run the program &#34;MembraneSepDist.m&#34;.</li>
	<li>Set Parameters. 
	NOTE: A GUI will pop up with default parameters for the gradient threshold, scale, region of interest, and manual start. The default values can be changed in the m-file, or they can be changed for each individual image.
	<ol>
		<li>Set spatial derivative gradient threshold. 
		NOTE: Higher values result in more points selected in centerline isolation. Values that are too high or too low (outside of a range of approximately 3.0-7.9) may result in computational inefficiency or an imprecise selection of centerline points yielding imprecise centerline isolation (See Figure 2A-C).</li>
		<li>Set scale in pixels/scale units.</li>
		<li>Set the spatial lower and upper limits for the region of interest. 
		NOTE: By our laboratory&#39;s convention, the defined region of interest is between 30 and 105 nm from the edge of the GJ2,3,10.</li>
		<li>Set automatic/manual start. In the majority of cases, the algorithm accurately detects the starting point where the gap junction ends and perinexus begins. However, in some cases of irregularly-shaped perinexi, the user must identify the start point manually. Set this value to 0 for automatic, 1 for manual.</li>
	</ol>
	</li>
	<li>Select desired image. 
	NOTE: The file select folder can be changed in the m-file.</li>
	<li>Crop the image to select the perinexus of interest.
	<ol>
		<li>When the image comes up, the cursor will automatically change to a crosshair. Crop the image by dragging a box around the perinexus of interest (See Figure 3). The crop box can be adjusted by using the squares on the corners and sides to make it larger or smaller.</li>
		<li>When cropping, ensure the &#34;open&#34; end of the perinexus (furthest from the GJ, see Figure 3) is cropped so that the two membrane outlines reach the edge of the cropped image. 
		NOTE: It is recommended to make the image full-screen to more easily see the perinexus of interest and crop appropriately.</li>
	</ol>
	</li>
	<li>Select the final crop by double-clicking with the curser in between the opposing edges to be measured. 
	NOTE: It is critical that the double-click be performed inside the perinexus. If the program fails to identify a centerline, restart the program and make sure the click occurs within the perinexus.</li>
	<li>Observe the final centerline after all dilations and erosions pop up for a final user evaluation of the program&#39;s efficacy. 
	NOTE: A dialog box will appear on the screen while the program is running to inform the user that Matlab will be unable to process any additional commands until the program has finished. How long this process takes depends on array (image) size and computer processing power.</li>
	<li>If manual start point is enabled, observe the image of the centerline pop up over the original anatomical image, along with a cross-hair cursor (see Figure 2E). Select a point outside of the perinexus near the desired start point. 
	NOTE: The program will find the centerline point closest to the selected pixel and use that as the start point.</li>
	<li>Record data. 
	NOTE: Once the program has finished, the program will return a mapped centerline, plot of perinexal width as a function of distance from the edge of the GJ. Additionally, the program will return the average perinexal width up to 150 nm from the edge of the GJ as well as the average from within the defined region of interest at the Matlab command line. Wp values and distances from the GJ are stored in the variable &#34;WpList&#34; or the user can manually record them separately.</li>
</ol>
4. Algorithm Troubleshooting
<ol>
	<li>If the centerline is not properly identified (Figure 2A), open the figure &#34;Gmag&#34; and use the Index to identify an appropriate gradient threshold (Figure 2C).</li>
	<li>If the start point is not properly identified, set the start point manually (See Protocol 3.8).</li>
</ol>

<ol>
	<li>Rhett, J. M., Gourdie, R. G. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=The+perinexus:+a+new+feature+of+Cx43+gap+junction+organization.">The perinexus: a new feature of Cx43 gap junction organization.</a> Heart Rhythm. 9 (4), 619-623 (2012).</li><li>Veeraraghavan, R., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Sodium+channels+in+the+Cx43+gap+junction+perinexus+may+constitute+a+cardiac+ephapse:+an+experimental+and+modeling+study.">Sodium channels in the Cx43 gap junction perinexus may constitute a cardiac ephapse: an experimental and modeling study.</a> Pflugers Archiv: European Journal of Physiology. , (2015).</li><li>George, S. A., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Extracellular+sodium+dependence+of+the+conduction+velocity-calcium+relationship:+evidence+of+ephaptic+self-attenuation.">Extracellular sodium dependence of the conduction velocity-calcium relationship: evidence of ephaptic self-attenuation.</a> American Journal of Physiology - Heart and Circulatory Physiology. 310 (9), 1129-1139 (2016).</li><li>Veeraraghavan, R., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Potassium+channels+in+the+Cx43+gap+junction+perinexus+modulate+ephaptic+coupling:+an+experimental+and+modeling+study.">Potassium channels in the Cx43 gap junction perinexus modulate ephaptic coupling: an experimental and modeling study.</a> Pflugers Archiv: European Journal of Physiology. , (2016).</li><li>Rhett, J. M., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Cx43+associates+with+Na(v)1.5+in+the+cardiomyocyte+perinexus.">Cx43 associates with Na(v)1.5 in the cardiomyocyte perinexus.</a> Journal of Membrane Biology. 245 (7), 411-422 (2012).</li><li>Raisch, T. B., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Intercalated+Disc+Extracellular+Nanodomain+Expansion+in+Patients+with+Atrial+Fibrillation.">Intercalated Disc Extracellular Nanodomain Expansion in Patients with Atrial Fibrillation.</a> Frontiers in Physiology. , (2018).</li><li>Yan, J., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Novel+methods+of+automated+quantification+of+gap+junction+distribution+and+interstitial+collagen+quantity+from+animal+and+human+atrial+tissue+sections.">Novel methods of automated quantification of gap junction distribution and interstitial collagen quantity from animal and human atrial tissue sections.</a> PLoS One. 9 (8), 104357 (2014).</li><li>Papari, G., Petkov, N. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Adaptive+pseudo+dilation+for+gestalt+edge+grouping+and+contour+detection.">Adaptive pseudo dilation for gestalt edge grouping and contour detection.</a> IEEE Transactions on Image Processing. 17 (10), 1950-1962 (2008).</li><li>George, S. A., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Extracellular+sodium+and+potassium+levels+modulate+cardiac+conduction+in+mice+heterozygous+null+for+the+Connexin43+gene.">Extracellular sodium and potassium levels modulate cardiac conduction in mice heterozygous null for the Connexin43 gene.</a> Pflugers Archiv: European Journal of Physiology. , (2015).</li><li>Revel, J. P., Karnovsky, M. J. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Hexagonal+array+of+subunits+in+intercellular+junctions+of+the+mouse+heart+and+liver.">Hexagonal array of subunits in intercellular junctions of the mouse heart and liver.</a> Journal of Cell Biology. 33 (3), 7-12 (1967).</li><li>Huttner, I., Boutet, M., More, R. H. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Gap+junctions+in+arterial+endothelium.">Gap junctions in arterial endothelium.</a> Journal of Cell Biology. 57 (1), 247-252 (1973).</li><li>Makowski, L., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Gap+junction+structures.+II.+Analysis+of+the+x-ray+diffraction+data.">Gap junction structures. II. Analysis of the x-ray diffraction data.</a> Journal of Cell Biology. 74 (2), 629-645 (1977).</li><li>Entz, M., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Heart+Rate+and+Extracellular+Sodium+and+Potassium+Modulation+of+Gap+Junction+Mediated+Conduction+in+Guinea+Pigs.">Heart Rate and Extracellular Sodium and Potassium Modulation of Gap Junction Mediated Conduction in Guinea Pigs.</a> Frontiers in Physiology. 7, 16 (2016).</li><li>Sild, M., Chatelain, R. P., Ruthazer, E. S. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Improved+method+for+the+quantification+of+motility+in+glia+and+other+morphologically+complex+cells.">Improved method for the quantification of motility in glia and other morphologically complex cells.</a> Neural Plasticity. 2013, 853727 (2013).</li><li>Rhett, J. M., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=The+perinexus:+Sign-post+on+the+path+to+a+new+model+of+cardiac+conduction.">The perinexus: Sign-post on the path to a new model of cardiac conduction.</a> Trends in Cardiovascular Medicine. , (2013).</li><li>Lang, R. M., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Recommendations+for+cardiac+chamber+quantification+by+echocardiography+in+adults:+an+update+from+the+American+Society+of+Echocardiography+and+the+European+Association+of+Cardiovascular+Imaging.">Recommendations for cardiac chamber quantification by echocardiography in adults: an update from the American Society of Echocardiography and the European Association of Cardiovascular Imaging.</a> Journal of the American Society of Echocardiography. 28 (1), 1-39 (2015).</li><li>Kawai, J., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Left+ventricular+volume+and+ejection+fraction+by+the+axius+auto+ejection+fraction+method:+comparison+with+manual+trace+method+and+visual+assessment+of+ejection+fraction.">Left ventricular volume and ejection fraction by the axius auto ejection fraction method: comparison with manual trace method and visual assessment of ejection fraction.</a> Journal of Cardiology. 49 (3), 125-134 (2007).</li><li>Frederiksen, C. A., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Clinical+utility+of+semi-automated+estimation+of+ejection+fraction+at+the+point-of-care.">Clinical utility of semi-automated estimation of ejection fraction at the point-of-care.</a> Heart, Lung and Vessels. 7 (3), 208-216 (2015).</li><li>Foster, F. S., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=A+new+ultrasound+instrument+for+in+vivo+microimaging+of+mice.">A new ultrasound instrument for in vivo microimaging of mice.</a> Ultrasound in Medicine and Biology. 28 (9), 1165-1172 (2002).</li><li>Moran, C. M., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=A+comparison+of+the+imaging+performance+of+high+resolution+ultrasound+scanners+for+preclinical+imaging.">A comparison of the imaging performance of high resolution ultrasound scanners for preclinical imaging.</a> Ultrasound in Medicine and Biology. 37 (3), 493-501 (2011).</li><li>Papademetris, X., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Estimation+of+3D+left+ventricular+deformation+from+echocardiography.">Estimation of 3D left ventricular deformation from echocardiography.</a> Medical Image Analysis. 5 (1), 17-28 (2001).</li><li>Hosny, A., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Unlocking+vendor-specific+tags:+Three-dimensional+printing+of+echocardiographic+data+sets.">Unlocking vendor-specific tags: Three-dimensional printing of echocardiographic data sets.</a> Journal of Thoracic Cardiovascular Surgery. 155 (1), 143-145 (2018).</li><li>Cretoiu, D., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Human+cardiac+telocytes:+3D+imaging+by+FIB-SEM+tomography.">Human cardiac telocytes: 3D imaging by FIB-SEM tomography.</a> Journal of Cellular and Molecular Medicine. 18 (11), 2157-2164 (2014).</li><li>Risi, C., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Ca(2+)-induced+movement+of+tropomyosin+on+native+cardiac+thin+filaments+revealed+by+cryoelectron+microscopy.">Ca(2+)-induced movement of tropomyosin on native cardiac thin filaments revealed by cryoelectron microscopy.</a> Proceedings of the National Academy of Sciences of the United States of America. 114 (26), 6782-6787 (2017).</li><li>Dhindwal, S., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=A+cryo-EM-based+model+of+phosphorylation-+and+FKBP12.6-mediated+allosterism+of+the+cardiac+ryanodine+receptor.">A cryo-EM-based model of phosphorylation- and FKBP12.6-mediated allosterism of the cardiac ryanodine receptor.</a> Science Signaling. 10 (480), (2017).</li><li>Reddy, V. Y., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Integration+of+cardiac+magnetic+resonance+imaging+with+three-dimensional+electroanatomic+mapping+to+guide+left+ventricular+catheter+manipulation:+feasibility+in+a+porcine+model+of+healed+myocardial+infarction.">Integration of cardiac magnetic resonance imaging with three-dimensional electroanatomic mapping to guide left ventricular catheter manipulation: feasibility in a porcine model of healed myocardial infarction.</a> Journal of the American College of Cardiology. 44 (11), 2202-2213 (2004).</li><li>van Heeswijk, R. B., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Three-Dimensional+Self-Navigated+T2+Mapping+for+the+Detection+of+Acute+Cellular+Rejection+After+Orthotopic+Heart+Transplantation.">Three-Dimensional Self-Navigated T2 Mapping for the Detection of Acute Cellular Rejection After Orthotopic Heart Transplantation.</a> Transplant Direct. 3 (4), 149 (2017).</li><li>Valinoti, M., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=3D+patient-specific+models+for+left+atrium+characterization+to+support+ablation+in+atrial+fibrillation+patients.">3D patient-specific models for left atrium characterization to support ablation in atrial fibrillation patients.</a> Magnetic Resonance Imaging. 45, 51-57 (2018).</li></ol>

The authors have nothing to disclose.

The algorithm uses serial image dilations to count the number of pixels between two opposing 2D edges in a binary image, which in this case is the inter-membrane separation of the perinexus2,3,14. A spatial derivative and a pathfinding algorithm are then used to isolate the centerline, followed by a secondary dilation and erosion sequence to fill gaps in the centerline, similar to what has been done before15. The centerline is then combined with the final dilation-count image to represent perinexal width as a function of distance from the beginning of edge separation, in this case the end of the GJ and the beginning of the perinexus16.
Four primary parameters are user-defined in a GUI at the start of the program:
<ol>
	<li>Gradient threshold</li>
	<li>Scale</li>
	<li>Region of interest range</li>
	<li>Start point selection method (automatic or manual)</li>
</ol>
The most common mechanism of failure for the algorithm is the failure of the centerline to reach the edge of the image, which is how the endpoint is determined for the pathfinding algorithm. In order to fix such an issue, the user can increase the gradient threshold described in step 3.3.1, which will cause the program to select more points out of the spatial derivative image, which will increase the computation time required by the pathfinding algorithm. Therefore, this algorithm requires a compromise between computation speed and centerline integrity. It is important to note that so long as all the points of the centerline are identified from the spatial derivative, along with an appropriate start point, the spatial derivative threshold will have no effect on the edge separation measurement.
Image orientation appears to affect dilation values, because the kernel dilates in 90 degree steps, which can introduce an error if the majority of the region of interest is at an angle 45&#176; to the axes of dilation matrices. Therefore, the dilation count may not always be an accurate representation of the space between the edges. This limitation has been addressed by a trigonometric correction factor, but could potentially be ignored if all images in a dataset are aligned at the same orientation. Furthermore, caution should be used in interpreting results, as it is possible that section planes are not perfectly perpendicular to the two membranes. In Figure 9B, we use GJW to suggest that our perinexus images were in-plane. Still, it is imperative that sample sizes be sufficiently large to account for any sectioning variations between images. Additionally, our perinexal width measurements should not be interpreted to reflect in vivo spaces, but this approach is used to measure mean differences in perinexal width relative to some intervention or disease state.
The current algorithm also requires a manually-traced outline of the edges as an input. It is important to note that so long as the scale is set correctly, spatial resolution has no effect on the algorithm&#39;s measurements, as demonstrated by the varying resolutions of images in Figure 6 and an additional low-resolution image in supplemental file S6. The next step in improving the algorithm is removing human intervention from outline generation along with a tool that can select the area of interest. These features would likely enhance the precision of the measurement and reduce user bias.
This computationally efficient algorithm provides a faster method, requiring approximately one fifth the man-hours, of quantifying the perinexus with no detectable penalty to reproducibility when compared to the manual segmentation process. Additionally, the manual segmentation process utilizes one measurement every 15 nanometers to quantify perinexal width, which can lead to under sampling as the membrane separation of the perinexus can change substantially within that 15 nm range. In contrast, the automated program has a spatial resolution equal to that of the imaging modality, in this case 2.9 pixels per nanometer along the length of the perinexus, therefore delivering a more finely resolved average of perinexal width.
While the applications in the field of cardiac structural biology are promising and exciting, this algorithm&#39;s uses are not confined to TEM images. Any field requiring a precise, high-resolution measurement of two quasi-parallel 2D edges can make use of this algorithm. The algorithm could be used to track anything from riverbank erosion and flood patterns from satellite images to vascular development with brightfield or fluorescent microscopy. One of the most promising potential applications is in the field of cardiology and measuring ventricular ejection fraction (EF) with point-of-care cardiac echocardiography. Currently, the standard technique is the biplane method of disks17, though a newer algorithm, AutoEF, is currently the cutting edge EF-quantifying method18,19. For biplane method of disks, the chamber in question is manually traced and quantified using a modified Simpson&#39;s method, whereby overall volume is automatically calculated by the summation of stacked elliptical disks. The main limitation with this method is that it can only return the total cross-sectional area of the desired chamber, with no resolution to identify specific regions of interest, and also requires substantial human input and expertise. The newer method, AutoEF, identifies and outlines the edge of the ventricle using a 2D speckling algorithm and then computes the ventricular cross-sectional area. This process, while precise and efficient for measuring gross ventricular area, also has a similar inherent limitation of only measuring total cross-sectional area. This primary drawback limits clinicians&#39; diagnostic and treatment abilities. In contrast, the algorithm presented in this manuscript can identify a midline and has a resolution equal to the resolution of the imaging modality to pinpoint specific regions of interest. This is important because ultrasound scanners with micrometer spatial resolution are commercially available20,21, implying that this algorithm could detect localized wall motion abnormalities at the resolution of micrometers instead of centimeters. While this application needs to be experimentally validated, it is one of the most immediately promising applications of this algorithm. In fact, it could easily be combined with the speckle tracking capabilities of AutoEF or the manual traces utilized in manual planimetry to provide higher-resolution information in parallel with conventional EF data.
As versatile and applicable as the current algorithm is, it was developed for 2D images. However, as imaging technologies continue to improve, there is an increasing demand for 3 and 4D quantification technologies. Therefore, the next iteration of the algorithm is to adapt the same approach, serially dilating a binary image, to a 3-dimensional object, where automatically defining a centerline is currently beyond the capabilities of current imaging programs. Such an algorithm would have wide applications both clinically and experimentally in the cardiac field alone, including 3D cardiac echocardiograms22,23, 3D electron microscopy24,25,26, and 3D magnetic resonance imaging27,28,29.

The following algorithm was developed to measure the intermembrane distance between two structurally coupled cardiomyocytes at the point they separate from each other at the edge of a gap junction plaque in a nanodomain called the perinexus1, which has been implicated in ephaptic coupling2,3,4,5. In the process of analyzing hundreds of transmission electron microscopy (TEM) images of perinexi using a manual segmentation method in a previous study6, the need was identified for a higher-throughput method that sampled perinexal width in higher spatial resolution while preserving the accuracy of the previous manual segmentation process During manual segmentation, lines are drawn at 15 nm intervals, approximately orthogonal to the centerline, in order to measure perinexal width.The new algorithm takes a one pixel thick binary outline of two parallel lines and uses serial image dilations to count the number of pixels between the two membranes.While image dilations have commonly been used in a myriad of image processing applications, including contour or edge detection7,8, this algorithm uses dilations as a counting mechanism. The centerline is then isolated using a pathfinding algorithm9 and perinexal width is then measured at a resolution along the length of perinexus equal to the resolution of the image. The difference in resolution in this case is 1 measurement per 15 nm for manual segmentation and 1 measurement per 0.34 nm with the new algorithm, a 44-fold increase in spatial sampling frequency.Furthermore, this increased sampling frequency is accomplished in approximately 1/5th the time needed for manual segmentation.
This algorithm will be used in its current form to measure perinexal width at the conventional 0-150 nm from the edge of a gap junction plaque5 (GJ) as well as within a specified region of interest, where the perinexus plateaus between 30 and 105 nm2,3,10. The increased sampling frequency reduces the variability in individual perinexus measurements compared to manual segmentation and significantly shortens analysis time, allowing for efficient processing of large data sets.However, this program is not limited to nanoscale TEM images of cardiac intercalated disks.The same approach could be used to quantify vascular diameter, ventricular ejection fraction, or even non-biological phenomena such as river erosion or flooding. This algorithm is appropriate for quantifying the distance between any two quasi-parallel edges.

<table border="0" cellpadding="0" cellspacing="0" style="width:867px;" width="866">
	<colgroup>
		<col />
		<col />
		<col />
		<col />
	</colgroup>
	<tbody>
		<tr height="21">
			<td height="21" style="height:21px;width:216px;">Touchscreen Monitor</td>
			<td style="width:79px;">Dell</td>
			<td style="width:119px;">S2240T</td>
			<td style="width:453px;">Needs soft-tipped stylus</td>
		</tr>
		<tr height="21">
			<td height="21" style="height:21px;width:216px;">Desktop</td>
			<td style="width:79px;">Dell</td>
			<td style="width:119px;">Precision T1650</td>
			<td>8GB RAM</td>
		</tr>
		<tr height="21">
			<td height="21" style="height:21px;width:216px;">Operating System</td>
			<td style="width:79px;">Microsoft</td>
			<td style="width:119px;">Windows 7 Enterprise</td>
			<td>64-bit OS</td>
		</tr>
		<tr height="21">
			<td height="21" style="height:21px;width:216px;">Program platform</td>
			<td style="width:79px;">Mathworks</td>
			<td style="width:119px;">Matlab R2015b</td>
			<td>Program may be incompatible with newer/older versions of Matlab</td>
		</tr>
	</tbody>
</table>

quantifying intermembrane distances with serial image dilations

A recently-described extracellular nanodomain, termed the perinexus, has been implicated in ephaptic coupling, which is an alternative mechanism for electrical conduction between cardiomyocytes. The current method for quantifying this space by manual segmentation is slow and has low spatial resolution.We developed an algorithm that uses serial image dilations of a binary outline to count the number of pixels between two opposing 2 dimensional edges.This algorithm requires fewer man hours and has a higher spatial resolution than the manual method while preserving the reproducibility of the manual process.In fact, experienced and novice investigators were able to recapitulate the results of a previous study with this new algorithm.The algorithm is limited by the human input needed to manually outline the perinexus and computational power mainly encumbered by a pre-existing pathfinding algorithm.However, the algorithm&#39;s high-throughput capabilities, high spatial resolution and reproducibility make it a versatile and robust measurement tool for use across a variety of applications requiring the measurement of the distance between any 2-dimensional (2D) edges.

Watch this Scientific Journal Video about Quantifying Intermembrane Distances with Serial Image Dilations at JoVE.com

Quantifying Intermembrane Distances with Serial Image Dilations

The purpose of this algorithm is to continuously measure the distance between two 2-dimensional edges using serial image dilations and pathfinding. This algorithm can be applied to a variety of fields such as cardiac structural biology, vascular biology, and civil engineering.

A recently-described extracellular nanodomain, termed the perinexus, has been implicated in ephaptic coupling, which is an alternative mechanism for ...

Research

JoVE Journal

Bioengineering

Preparation of Complaint Matrices for Quantifying Cellular Contraction

The regulation of cellular adhesion to the extracellular matrix (ECM) is essential for cell migration and ECM remodeling.  Focal adhesions are ...

Quantifying the Mechanical Properties of the Endothelial Glycocalyx with Atomic Force Microscopy

Our  understanding of the interaction of leukocytes and the vessel wall during  leukocyte capture is limited by an incomplete understanding of the ...

Observing and Quantifying Fibroblast-mediated Fibrin Gel Compaction

Cells embedded in collagen and fibrin gels attach and exert traction forces on the fibers of the gel. These forces can lead to local and global ...

Manufacture of Concentrated, Lipid-based Oxygen Microbubble Emulsions by High Shear Homogenization and Serial Concentration

Gas-filled microbubbles have been developed as ultrasound contrast and drug delivery agents. Microbubbles can be produced by processing surfactants using ...

Hybrid &#181;CT-FMT imaging and image analysis

Hybrid Ãƒâ€šÃ‚ÂµCT-FMT imaging and image analysis

Fluorescence-mediated tomography (FMT) enables longitudinal and quantitative determination of the fluorescence distribution in vivo and can be used to ...

Custom-designed Laser-based Heating Apparatus for Triggered Release of Cisplatin from Thermosensitive Liposomes with Magnetic Resonance Image Guidance

Liposomes have been employed as drug delivery systems to target solid tumors through exploitation of the enhanced permeability and retention (EPR) effect ...

Image-guided, Laser-based Fabrication of Vascular-derived Microfluidic Networks

This detailed protocol outlines the implementation of image-guided, laser-based hydrogel degradation for the fabrication of vascular-derived microfluidic ...

Quantifying Microorganisms at Low Concentrations Using Digital Holographic Microscopy (DHM)

Quantifying Microorganisms at Low Concentrations Using Digital Holographic Microscopy DHM

Accurately detecting and counting sparse bacterial samples has many applications in the food, beverage, and pharmaceutical processing industries, in ...

A Rapid Image-based Bacterial Virulence Assay Using Amoeba

Traditional bacterial virulence assays involve prolonged exposure of bacteria over the course of several hours to host cells. During this time, bacteria ...

Meso-Scale Particle Image Velocimetry Studies of Neurovascular Flows In Vitro

Meso-Scale Particle Image Velocimetry Studies of Neurovascular Flows In Vitro

Particle image velocimetry (PIV) is used in a wide variety of fields, due to the opportunity it provides for precisely visualizing and quantifying flows ...