FilamentTracer is the most advanced software product for the automatic detection of neurons (dendritic trees, axons and spines), microtubules, and other filament-like structures in 2D, 3D and 4D. When combined with ImarisTrack, detection of temporal changes in length and volume of developing spines and dendrites will help researchers understand alterations caused by developmental and environmental changes. FilamentTracer utilizes multiple automatic, semi-automatic, and manual segmentation methods that can be used in any combination, to successfully segment, then visualize and quantify the detected structures.
Accurate morphological description of filamentous structures requires the ultimate resolution available from light microscopy today with structures often of the size that is near or at the theoretical limit of resolution of the best objective lens. The best possible imaging is needed to capture the finest of structures of interest. These are fundamental reasons why the analysis of neuron and filament morphology is so challenging. Purely manual or purely automated strategies are therefore likely to fail. Imaris FilamentTracer uses strategies that involve an optimal combination of automated analysis and operator decision. The choice of the best method is an essential element in optimizing efficiency. FilamentTracer allows the choice between four complementary tracing strategies ranging from manual to automatic. FilamentTracer utilizes a creation wizard to easily guide the user through the creations steps.
Equally essential to optimizing efficiency is the ability to overlay the segmented structure (tracing) and the 3D original data in real time for all time points of 3D time lapse. FilamentTracer is fully integrated with Imaris, enabling users to leverage its impressive range of 3D and 4D visualization capabilities as well as its image editing and processing functions. With FilamentTracer there is no need to switch between programs, windows, or views, costing the user valuable time. Instead FilamentTracer allows the creation, visualization, editing, and analysis all in one window. By having an interactive overlay of the original image and the resulting tracing the user can easily assess the validity of the tracing and make any necessary changes in 3D space.
FilamentTracer has been specifically designed to deal with the complex problem of spine analysis over time. This is a challenging process that to date has been a time-consuming task that often did not produce a reliable 4D analysis of spines’ fate. There is no need to manually analyze spines anymore: With just a few clicks of a button, FilamentTracer allows 4D tracking and classification of detected structures such as spines and dendrites. Researchers can examine statistics that change over time for both dendrites and spines including information on counts, branching, size, shape, and intensity. Researchers can being to explore spine evolution, e.g. during brain development, memory formation, stroke and other vital life processes.
Together with ImarisXT, analysis can be taken a step further and spines can be classified by diameter, size, shape, or any other parameter that is calculated by FilamentTracer. The ImarisXT module allows custom configuration of any user defined classification rules. The ideal combination for FilamentTracer is Imaris, Imaris MeasurementPro (measurement of volume and intensity changes over time and manual object segmentation), ImarisTrack (for tracing the temporal changes in spines and dendrites) and ImarisXT (automated spine classification).
The FilamentTracer functionality requires Imaris.
FilamentTracer offers the largest choice of tracing methods on the market to ensure the fast and accurate tracing of filament like structures in 2D and 3D microscopy images. The creation wizard helps the user choose the optimal parameters for any detection method. Detection and segmentation methods can be freely combined on any image.
Fully Automatic – FilamentTracer offers the choice between two fully automatic detection methods; “AutoPath” (creates trees without loops) or automatic detection based on intensity thresholds (may contain loops). Just put simple parameters into the creation wizard and FilamentTracer completes the tracing.
Semi Automatic – FilamentTracer allows semi automatic tracing (using “AutoPath”) by manually selecting starting and end points. Click on a starting location in 3D space or choose a segment or group of segments as the starting point. FilamentTracer performs a calculation. Then you can simply click on end points and FilamentTracer instantly provides a preview tracing along the best path back to the starting point as you move the mouse in the image. Clicking the end point sets the tracing.
Advanced Manual – FilamentTracer’s advanced manual tracing mode allows you to directly trace onto a 3D volume image with automatic z depth placement (“AutoDepth”) and achieve a fast, accurate and reproducible manual tracing result. This enables you to quickly trace in 3D space. For more flexible and real-time filament analysis use the AutoNetwork option that maps out the entire network in one step and cuts down on time otherwise needed to recalculate an alternative filament trace. By using Waypoints during the tracing process, you can define the key locations for the trace to follow en route from the start to end point.
Manual – FilamentTracer provides a tool to fully trace structures manually when focusing on specific details in complex structures. Tracing can be completed on individual planes.
Region of interest processing (ROI) – FilamentTracer allows you to process single or multiple regions of interest to target specific areas of the image or to test tracing parameters which can then be used automatically on the entire image.
Iterative Processing – FilamentTracer allows the user to trace a small section of the dataset automatically or manually and then the user can “build” upon the existing data in an iterative fashion using the automatic methods. This provides excellent functionality for data sets that are extremely large or where there is no clear starting point.
Rebuild – Choose between keeping and deleting existing data when re-entering the automatic creation process of the FilamentTracer Creation Wizard. This allows you to keep the structures you have traced manually while adjusting the automatic parameters that were originally set.
With its novel Spine and Dendrite tracking in 3D over time, FilamentTracer offers you a powerful tool with which you can analyze images of neurons undergoing continual changes. Utilizing FilamentTracer's intuitive wizard-driven detection algorithms, the user can easily define which spines and dendrites are going to be segmented, then which terminal, branch or attachment points will be tracked and finally visualized in 4D.
Automatic Tracking - After segmentation with one of the automatic methods described above, choose to track dendrites and / or spines within the creation wizard.
- Dendrites may be tracked based on the position of dendrite branch points or dendrite terminal points.
- Spines may be tracked based on the position of spine attachment points, terminal points, or branch points.
Autoregressive Motion Model - For the most efficient tracking of directional objects like Spines and Dendrites the autoregressive motion model is automatically used. The user simply has to input the maximum distance they believe a spine and / or dendrite attachment point will move between two adjacent time points. If due to imaging problems a spine or dendrite attachment point is not detected at a particular time point, the user has the option to allow a specified number of time frame to pass where the attachment point is missing and still complete the tracking properly. In this mode, FilamentTracer can automatically deal with changes in acceleration / deceleration of growth or shrinkage of spines and dendrites and improves tracking accuracy for those objects.
Manual Track Editing – Once segmentation is complete and the image is automatically tracked, the user can use the track editor to manually correct any errors that may have occurred during automatic tracking. The track editor for filaments performs the same way that is does for spot(s), surface(s), and cell(s) objects making learning the software easier.
A rich selection of interactive 3D and 4D display modes is important for the creation, editing, and presentation of tracings. This allows multiple viewing combinations in which the tracing can be displayed individually as an overlay on the original data or together with other cellular components that may have been imaged in other channels
Display – FilamentTracer allows the display of 4D traced segments and preview for all time points as lines, constant-width cylinders or real-diameter cylinders (showing the actual diameter measured from the structure). Each part of a filament tracing can be displayed with any color or transparency.
Surpass View - Combine Filament tracings with any Imaris “Surpass Object” to show surrounding cellular or tissue components and superimpose filaments over the original volume rendering to verify your segmentation.
FilamentTracer has a versatile range of methods to allow you to select parts of the tracing for editing and classification following the creation process.
Selection – FilamentTracer provides for selection of individual segments, entire branches, or single points as well as for the selection of loops or segments based on absolute or relative size. Multiple selections, inverse selections, and clearing selections can also be completed. FilamentTracer has an option to automatically extend your dendrite Selection to highlight attached Spines associated with that dendrite.
Selection Processing – Once selected, traced components can be deleted or joined. More importantly any segment (such as manually or semi-manually drawn segments) can be properly centered to the original data and can also be “grown” so the tracing has the accurate diameter and shape when compared to the original data.
Classifications – Any selected segment can be classified as a dendrite or spine. This is important for proper analysis of neuroscience images. When combined with Bitplane’s ImarisXT spine classifier, spines can be classified into different groups. Both spines and dendrites can also be classified into beginning points, branches, branching points and terminal points.
Filament Processing – Choose between manual and automatic merging of Filaments, split filaments into unconnected sub-filaments and manually connect broken or unconnected segments. Created filament segments can now be adjusted in post-process with an option that allows to select an input channel for filament diameter's re-calculation.
FilamentTracer provides interactive measurements and data export. This feature requires Imaris MeasurementPro.
Statistical Values – FilamentTracer provides a huge selection of Neuroscience specific statistical values: Dendrite Volume, Dendrite Length, Number of Dendrite Segments, Number of Dendrite Branches, Dendrite Area, Spine Volume, Spine Length, Number of Spines, Spine Density, Points, Filaments, Branch Level, Scholl Analysis, and many more. Now in addition to these parameters the users can select specific values related to the 3D tracking over time of Dendrite and Spine's edges, branch and attachment points This tracking data allows researchers to examine changes over time in any statistical parameter calculated by FilamentTracer. These statistics can also be used on non-neuroscience related filament like structures.
Data Export – Easy export of data to selected modeling programs (Neuron) and to spreadsheet programs (Excel, CSV format).
Filtering – Utilize the identical filtering methods with FilamentTracer as described in Imaris MeasurementPro.
Segment, edit, display, and measure filament like structures
Works on 3D and 4D structures
Segmented (traced) and original data displayed in the same window
Preview and display 4D-traced segments and for all time points during filament creation process
For neuronal images, segments may be classified as spines or dendrites
Any combination of automatic, semi-automatic, and manual methods may be used together
Fully automatic detection
- Wizard based interface guides automatic creation
- AutoPath algorithm
- Produces a tree-like filament with no loops
- Based on local intensity contrast
- Input the largest and smallest diameter segments
- Interactively include or exclude start and end points
- Start and end points may be manually added or deleted
- Automatically connects large start- and small end-points to form tree
- Dynamic control over the maximum gap allowed between disconnected filament segments. This tool (Remove Disconnected Components) is particularly useful when segmenting images with sub-optimal quality (e.g. overall low signal to noise ratio and/or in the presence of high intensity (non-filament) points in the background).
- Can be optimized for spine detection
- Segments may be classified as spines as part of the creation wizard
- Threshold based algorithm
- Produces a filament that can have loops
- Utilize an edge preserving filter to maintain filamentous structures
- Enable feature preprocessing to remove noise that would cause erroneous tracings but maintain structure above the diameter entered
- Based on an connective intensity threshold
- Interactively in real time select the double threshold
- Fill cavities in the tracing to create fully enclosed volumes as necessary
- Eliminate small segments if desired as part of the creation
- Original data may be kept as part of the rebuilding process to keep manually or semi manually added features.
- Region of Interest (ROI) processing – Automatically trace on one or more ROI’s.
- Iterative creation of tracings - Automatically or manually trace a small region then use this traced area as a starting point to continue automatic, semi-automatic or manual tracing.
-Interactive volume / size determination - Use the preview (similar to surfaces) overlaid on the image in 3D to interactively dial in the correct threshold used to calculate the size characteristics of both the dendrites and spines.
-Independent dendrite and spine segmentation- Utilize different segmentation parameters for dendrites first and then optionally spines second.
- Manually select a start point
- Point to end locations and a tracing along the path of highest intensity is calculated as a real time interactive preview
- Click on end points to complete the tracing
- Cursor is automatically positioned in 3D to the most intense point for selections
- Existing segments can be used as the starting point (i.e. all new segments must connect to the existing segment)
- Manually draw onto the 3D image
- Automatic placement (the location of the highest data intensity) in the depth of the screen is performed to accurately assist with the tracing in 3D
- map out the entire network in one step
- cut down on the time needed to recalculate an alternative filament trace
- use Waypoints during the tracing process - define the key locations for the trace to follow en route from the start to end point
- Segments may be drawn with an initial classification as dendrite or spine
- Manually trace filaments in a single plane YZ, XZ or XY plane at a time
- Manually draw filaments and connections on the plane
- Plane may be manually be moved along a given axis manually
- Plane may be automatically moved along the axis as the tracing is performed
Retain various sets of filament creation parameters - save and remember filament object creation process parameters to use identical settings for analysis of another data set
Dendrites and spines
- Line display– displays the center path of the tracing a fixed number of pixels in size
- Cylinder display - displays a cylinder of a fixed diameter along the center path of the tracing
- Cone display
- Displays a cylinder of varying diameter that corresponds to the measured diameter of the structure being traced
- The diameter may also be scaled to a fraction of the actual diameter of the structure
- Can choose to show or hide independently
- Can be colored any desired color
Points – Displayed as spheres
- Dendrite Beginning Point – The initial start point for a tracing
- Dendrite Branch Points – The point at which a dendrite branches into to more dendrites
- Dendrite Terminal Points – The point at which a dendrite ends
- Spine Attachment Points – The intersection point between a dendrite and a spine
- Spine Branch Points – The point at which a spine branches to two spines
- Spine Terminal Points – The point at which a spine ends
- Can choose to hide or show each independently
- Can be colored any desired color
Superimpose a filament object with any other object in Imaris including the original image or surfaces
Visual and interactive with the mouse
- Point Mode – selects one of the point objects defined above or selects the smallest piece of a segment that FilamentTracer measures
- Segment – Selects the piece of the tracing between two segments of the same class (dendrite or spine)
- Between dendrite beginning point and dendrite branch point
- Between dendrite branch points
- Between dendrite branch point and dendrite terminal point
- Between spine attachment point and spine terminal point
- Between spine attachment point and spine branch point
- Between spine branch point and spine terminal point
- Branch – Selects the piece of the tracing between the closest connection point and the terminal points of the same class (spine or dendrite)
- Between spine attachment or dendrite branch point to the dendrite terminal point for dendrites
- Between the spine attachment point or spine branching point to the spine terminal point for spines
- Automatically extend dendrite Selection to highlight attached Spines associated with that dendrite
- Relative branch length - selects short parts of the Filament defined by the ratio of the branch length to trunk radius
- Absolute branch length- selects all branches shorter than specified length
- Loops - finds and selects circular loops closed Filaments
Current selection based
- Path – finds and selects the shortest path between two currently selected points
- Invert – chooses all currently unselected segments or point and selects them
- Clear – clears currently selected objects from being selected (does not delete them)
- Pick and single or combination of object statistics in the statistics tab and the associated objects are selected in the visual view
- Selection is completely interactive and real-time
Assign as dendrite – Selected segments are assigned as dendrites
Assign as spine – Selected segments are assigned as spines
Assign as beginning point – Selected point is assigned as a beginning point
Center – centers the selected segments based on the intensity of the original data which is extremely useful for segments drawn using the AutoDepth or manual methods
- Calculates the correct diameter of the selected segments based on the image
- User must enter the largest and smallest diameter and a contrast ratio
- Adjusted filament segments in post-process by selecting an input channel for filament diameter's re-calculation.
Smooth – Removes big jumps in the tracing of the selected area which may be due to missing intensity values and results in a more uniform line
Delete – Deletes the selected components from the tracing
Duplicate – Takes the selected components and copies them to a new filament object where they can be analyzed separately
Join – Joins two unconnected filament points or segments by creating a new segment between them.
Smooth – same as described for selected object above but on entire filament object
Center– same as described for selected object above but on entire filament object
Diameter– same as described for selected object above but on entire filament object
Split - Splits unconnected parts (not touching in 3D space) into separate filament objects
Merge – Combines separate filament objects listed in the Surpass into one filament object
Export – Exports the tracing as a Neuron *.hoc file for analysis is the simulation program Neuron
- Area - Sum of the generated surfaces of a frustum (truncated cone)
- Length - The sum of the length of all edges which compose a dendrite
- Mean Diameter - The mean diameter within a dendrite
- Volume - The sum of volume of all edges (cones) which compose a dendrite
- Branching Angle - The angle directly at a branch point
- Branching Angle B - The angle of the whole dendritic segment
- Orientation Angle - Dendrite angle related to the x-axis of the data set
- Dendrite No. Spines - Number of spines branching off a dendrite
- Spine Density - Number of spines per length of the dendrite.
- Straightness h = the distance between two branch points. The dendrite straightness is h per length of the dendrite.
- Resistance - This value is proportional to the length and indirect proportional to the cross-section area. A variable diameter is considered. Multiply this value with an electrical material constant to obtain an electrical resistor (composed from small cones)
- Area - Sum of the frustums of the truncated cones plus area of the hemisphere at the spine end minus area of the hemisphere at the spine beginning
- Length - Sum of the length of the edges in a spine segment plus the radius of the sphere at the spine and minus the sphere at the spine beginning
- Volume - The sum of volume of all truncated cones plus volume of hemisphere at the spine end minus volume of the hemisphere at the spine beginning.
- Mean Diameter - Mean spine vertex diameter including the attachment point.
- Min Diameter - Min spine vertex diameter including the attachment point.
- Attachment Point Diameter - Diameter of a vertex, where the spine touches at the dendrite.
- Terminal Point Diameter - Diameter of a vertex at the end of a spine segment.
-Terminal Pt. Volume - Volume is calculated as a volume of the sphere with the diameter of a spine terminal point vertex
-Neck Volume -Calculated by subtracting the volume of the spine-ending hemisphere from the spine volume.
This video tutorial provides an overview of automatic dendrite and spine detection.
This video tutorial provides an overview of AutoPath, AutoDepth and Manual filament tracing methods.
Movie: Semi Automatic and Manual Filament Tracing Methods Tutorial movie (7:26)
Download the Movie
This movie shows the sequential analysis of the mitochondrial network. Courtesy of Arnaud Chevrollier
Click to view the movie
This movie shows the Zebrafish hindbrain from 1.5 to 2 days post fertilization. The arrows show new vessels sprouting into brain.
Click to view the movie
This 3-D rotational view shows zebrafish head vasculature of 3 days post fertilization.
Click to view the movie
FilamentTracer: Quantification of Spine Alterations
Dr. Shen, Toda, et al, University of South Carolina
Changes in cellular neuroplasticity may underlie behavioral changes after chronic and acute cocaine exposure. Dendrite spine morphology is thought to regulate synaptic strength therefore playing a critical role in synaptic plasticity. >>>
High-Resolution Imaging in Zebrafish
Prof. R. Klemke and colleagues, University of California at San Diego
Metastasis, the major cause of death in cancer patients, is a highly dynamic process that occurs in multiple steps, which include disruption of cell-cell adhesion, migration of cells away from the primary tumor and intravasation into the vasculature. >>>