Index ← Previous Topic Next Topic →

Settings

After loading a dataset you should set the program parameters by selecting Run >> Settings... from the main menu. The Settings window contains a number of tabbed windows used to enter parameters used in the analysis of the data.

Voxel Size

The Voxel Size window (shown below) allows you to enter the dimensions of each voxel in microns (μm). The X and Y dimensions relate to the lateral resolution of your scanned data. The Z dimension relates to the axial resolution or step size used in the image acquisition.

Voxel Size Settings

Neurites

The Neurites window (shown below) allows you to enter parameters used to control the pruning of false-positive branches in a model. The Attach Ratio and Min Length are the two settings used in conjunction to accomplish this. A branch in the model is kept if it meets the following simple formula.

length of the branch >= Min Length + radius of the attachment point × Attach Ratio

Neurites Settings

IMPORTANT: It is assumed that the correct voxel dimensions have been entered in the Voxel Size window.

TIP: In most cases it is only necessary to adjust the Min Length setting. You can keep increasing this value and then selecting the Apply button to iteratively refine the model. See example below.

Example: (voxel size 0.098μm x 0.098μm x 0.081μm)

1) Set Attach Ratio = 1.5 and Min Length = 0.0μm. The Apply button is pressed. Result shown below.

Settings Example 1 Image 1

2) Set Attach Ratio = 1.5 and Min Length = 0.75μm. The Apply button is pressed. Result shown below.

Settings Example 1 Image 2

3) Set Attach Ratio = 1.5 and Min Length = 1.5μm. The Apply button is pressed. Result shown below.

Settings Example 1 Image 3

Spines

The Spines window (shown below) controls the spine detection component of the program. Here the minimum and maximum expected values for width, height, and volume of spines must be provided. You should exercise care while changing these values since they not only affect the result but can also have a dramatic effect on performance. Values are expressed in microns (μm) for width and height, or # of voxels for stubby and non-stubby size. It is important to know that the spine detection will only work in sections of the dataset where a correct model is present. This model can have been created either by tracing methods or imported as a file. Therefore you should work on getting the best possible model first before running the spines. IMPORTANT: These settings depened on the correct voxel dimensions being set.

Spines Settings

TIP: You can keep adjusting the spine parameters and then select Run >> Build Spines to iteratively refine the results. See examples below.

Example: (voxel size 0.05μm x 0.05μm x 0.1μm)

1) Increase the Max Height setting (from 1.0μm to 3.0μm) to capture more of the taller spines. Run >> Build Spines selected. Result shown below.

Settings Example 2 Image 1 Settings Example 2 Image 2

2) Increase the Min Height setting (from 0.0μm to 0.2μm) to remove unwanted short spines, which may be considered small bumps on the dendrite. Run >> Build Spines selected. Result shown below.

Settings Example 2 Image 3 Settings Example 2 Image 4

3) Increase the Max Width setting (from 1.0μm to 3.0μm) to capture fatter spines. Run >> Build Spines selected. Result shown below.

Settings Example 2 Image 5 Settings Example 2 Image 6

4) Increase the Min Non-Stubby Size setting (from 5 to 10) to remove unwanted tiny spines, which may just be noise. Run >> Build Spines selected. Result shown below.

Settings Example 2 Image 7 Settings Example 2 Image 8

Threshold

The Threshold window (shown below) is used to specify if dynamic or fixed threshold is used for defining object and background voxels. When fixed threshold is used, voxels with intensity at or above the specified value are considered part of the object while those below are considered part of the background. If Use Dynamic is checked the program will adjust the threshold used according to the local intensity distribution of the data.

Threshold Settings

TIP: Unless you have a valid reason, it is recommended to use dynamic thresholding since it can compensate for non-uniform staining of the cell.

In addition, the Spine Correction setting allows the threshold (dynamic or fixed) to be adjusted downwards by a specified percent value in order to capture possibly darker spines. If the percentage entered is negative, the value is adjusted upwards. The default is zero percent adjustment.

TIP: Similar to the neurites and other spine parameters, you can keep adjusting the Spine Correction and then select Run >> Build Spines to iteratively refine the results. See example below.

Example: (voxel size 0.05μm x 0.05μm x 0.1μm)

1) Increase the Spine Correction setting (from 0.0% to 15.0%) to capture more of the darker spines. Run >> Build Spines selected. Result shown below.

Settings Example 3 Image 1 Settings Example 3 Image 2

TIP: Setting the Spine Correction setting too high is not recommended in most cases as this could cause unrealistic clumping of spines.

Defaults

The Defaults window (shown below) allows you to save the current settings as the program defaults. This is useful for cases were the imaging parameters remain constant. The Save Current button stores the current settings as the defaults for subsequent runs of the program. The Load Initial button resets all the settings to initial program defaults. Please note that the program stores all the settings for all previously run stacks in case the same stack is reloaded so the default settings only apply to stacks that have not been opened before.

Defaults Settings

Index ← Previous Topic Next Topic →
Send questions or comments to: alfredo.rodriguez@mssm.edu