JP5569892B2 - Automatic tracking system for moving particles in axons - Google Patents

Description

  The present invention relates to a method, an analysis apparatus, and a program for analyzing a particle path of intraneuronal transport of nerve cells.

  Nerve cells (hereinafter sometimes referred to as “neurons”) have axons that range from a few centimeters to sometimes 1 meter. Axons are guided by various axon guidance molecules and extend to form synapses with target cells, forming complex neural networks. Various proteins, intracellular organelles, and RNA required for axon formation and synapse formation are synthesized in the cell body and transported to the required site. On the contrary, the growth factor received at the tip of the axon is carried to the cell body as a signaling vesicle together with the receptor, and is involved in cellular activities such as transcriptional control (Non-patent Document 1). Such axonal transport is carried out mainly by motor molecules such as kinesin and dynein along microtubules. Transport from the cell body to the tip of the axon is called an anterograde axonal transport, and conversely, transport from the tip of the axon to the cell body is called a retrograde axonal transport (non-retrograde axonal transport). Patent Document 2). Impairment of axonal transport has been suggested to be associated with the development of various neurological diseases.For example, in Alzheimer's disease, axonal transport is impaired in early lesions, and microtubule-associated protein tau The effect of hyperphosphorylation of CRMP2 has been reported (Non-patent Document 3), and many of the causative genes of hereditary motor neuron disease are reported to be genes related to axonal transport ( Non-patent document 4). Therefore, intracellular axon transport is considered to be a vital phenomenon essential for the maintenance and functional expression of neurons. Axon transport analysis as an indicator of neurodegenerative diseases, especially tracking and quantitative analysis of axon transport particles Analysis is expected to make a significant contribution to the medical and medical fields for these diseases.

  Conventionally, in the research field of neuronal axonal transport, methods such as visualization of proteins related to axonal transport using GFP (green fluorescence protein) tags and tracer experiments using radioisotopes have been performed. In quantitative analysis of intracellular transport, axons are photographed with high magnification using a differential interference microscope, and particles passing through virtual lines are visually measured with emphasis on contrast (Non-Patent Document 5).

  However, visualization using a GFP tag is effective for tracking the transport of specific molecules, but is not suitable for the overall analysis of axonal transport. In addition, it is not possible to eliminate the possibility of an effect on transport, and furthermore, GFP labeling by normal gene transfer that is not homologous recombination has a problem that overexpression of the target molecule affects the axon transport. On the other hand, GFP labeling by homologous recombination requires enormous effort and cost. In labeling with a radioisotope, there is a problem that the use of a radioactive substance itself increases the risk of experiments and is not suitable for observation over time. In addition, observation with a differential interference microscope makes it very difficult to recognize the transported particles, so it requires a high level of skill, and requires a long time for observation and measurement per specimen. Therefore, there is a drawback that the obtained data has poor objectivity.

  On the other hand, in particle tracking technology using image processing, tracking technology for two-dimensionally distributed particles has been developed, and target particles are tracked two-dimensionally based on the two-dimensional freedom of movement of each particle. It is possible to do. However, within the axon of a nerve cell, the distribution and movement of transport particles do not have a planar spread due to the elongated shape of the axon, and the transport particles move along the axon, and the distribution also moves within the axon. Therefore, it is difficult to separate the motion between the granules. Therefore, it is difficult to analyze a transport route by applying a conventional tracking technique as it is based on two-dimensional information of particle movements for axon transport of nerve cells. In addition, the presence of many stationary particles around the axon and temporal random noise make the particle transport path analysis more difficult. Therefore, it is necessary to establish an analysis system that can automatically and quantitatively analyze the transport of nerve cells in the axon without being affected by interference between transport particles, particles around the axon, and random noise.

Zweifel LS et al., Nat. Rev. Neurosci., 2005, 6 (8): 615-25. Hirokawa N et al., Nat. Rev. Neurosci., 2005, 6 (3): 201-214. Stokin GB et al., Science, 2005, 307 (5713): 1282-8. Chevalier-Larsen E et al., Biochim. Biophys. Acta., 2006, 1762 11-12: 1098-1108. Goshima Y et al., J. Neurobiol., 1997, 33: 316-328.

  As described above, the particle path tracking method according to the prior art is not affected by interference between transport particles, particles around axons or random noise, and cannot realize highly accurate particle path analysis.

  Therefore, an object of the present invention is to provide a method for analyzing a novel axon transport particle path of nerve cells that is not affected by interference between transport particles, particles around axons or random noise, An apparatus for realizing the analysis method and a computer program therefor are provided.

  As a result of diligent research, the inventors of the present application visualized vesicle particles in nerve cells clearly and experimentally proved that the visualized vesicle particles are transported by an axonal transport mechanism. Furthermore, the inventors of the present application can suppress the influence of extra-axon stationary particles and random noise by extracting the axon region and limiting the processing target in the initial stage of analysis, and further, particle tracking In this process, based on the movement trajectory of the particles after the primary tracking, the possibility of mistracking is evaluated based on the mutual overlap of particles and the two-dimensional proximity state, and a stable tracking path is extracted. Accurate particle tracking is possible by setting an evaluation function that takes into account the continuity between particles and the continuity of the shape characteristics of particles, and integrating stable tracking paths so that the evaluation of each particle as a whole is optimized. The technique which develops was developed and it came to complete this invention.

  That is, the present invention comprises a step of visualizing vesicle particles passing through nerve cell axons, a step of obtaining images obtained by photographing the nerve cells over time, a step of inputting the obtained image frame to a computer, Extracting an axon region in an image frame; detecting a vesicle particle in the axon region; locally tracking the detected vesicle particle; extracting a stable tracking path; and integrating the stable tracking path. There is provided a method for analyzing axonal transport of a nerve cell, including a step of performing an arithmetic processing and a step of outputting a trajectory of a vesicle particle calculated by the arithmetic processing.

  Furthermore, the present invention provides a means for obtaining an image obtained by taking a time-lapse image of a nerve cell in which vesicle particles passing through an axon are visualized, a means for inputting the obtained image frame to a computer, and an axis in the image frame Means for extracting a cord region, means for detecting a vesicle particle in the axon region, means for computing local tracking of the detected vesicle particle, extraction of a stable tracking path and integration of the stable tracking path; And a neuron axon transport analyzing apparatus including a means for outputting a trajectory of a vesicle particle calculated by the arithmetic processing.

  For the first time, the present invention tracks clearly axon transport particles that are clearly visualized, and does not depend on interference between transport particles, particles around the axon, or random noise, and the particle path of nerve cell intra-axon transport Can be analyzed with high accuracy. Furthermore, the analysis system according to the present invention is expected to contribute to various fields such as drug development for the treatment of neurological diseases, screening of candidate substances, basic medical research, and biological research. In addition, by providing a visualization method of vesicle particles transported by microtubule-dependent axonal transport of nerve cells, it will greatly contribute to the research of axonal transport in general.

  The nerve cells to be analyzed for axonal transport according to the present invention are not particularly limited, but neurons that can be cultured in vitro are preferable from the viewpoint of simplicity of taking images of nerve cells. Examples of such neurons include neurons that secrete amino acid neurotransmitters such as glutamate, γ-aminobutyric acid, aspartate, and glycine, neurons that secrete peptide neurotransmitters such as vasopressin, somatostatin, and neurotensin, and And monoamine neurotransmitters such as noradrenaline, dopamine and serotonin, and neurons that secrete acetylcholine. Further, as will be described later, in the analysis method according to the present invention, the axon region of the nerve cell is extracted and the axon transport is analyzed. Therefore, the culture system of the nerve cell to be analyzed includes a type other than the nerve cell. Cells, for example, glial cells, oligodendrocytes, vascular epithelial cells, blood cells and the like, which are expected to be mixed when collecting nerve cells from nerve tissue, may be present.

  Cargo transported by axonal transport (Cargo) includes vesicles encased in lipid bilayers and protein complexes that are not encased in lipid bilayers. Vesicular particles are the detection target, and particularly preferably, the vesicle particles transported by microtubule-dependent axonal transport are the detection target. Whether the visualized vesicle particles are vesicle particles transported by microtubule-dependent axonal transport depends on axons such as nocodazole, a microtubule polymerization inhibitor, and sodium azide, an ATPase inhibitor. The addition of a transport inhibitor can be verified by inhibiting microtubule-dependent axonal transport. That is, when the transport of visualized vesicle particles decreases due to inhibition of microtubule-dependent axonal transport, the visualized vesicle particles are transported by microtubule-dependent axonal transport. It can be said that it is a particle.

  The step of visualizing vesicles according to the present invention is not particularly limited as long as it is carried out by visual observation or a method that can be visualized through an optical instrument such as a microscope, but from the viewpoint of detection sensitivity, the fluorescence of the lipid bilayer membrane It is preferably performed by labeling. Furthermore, staining with a carbocyanine fluorescent dye is particularly preferable from the viewpoint of staining efficiency. Staining with a carbocyanine fluorescent dye can be performed using a commercially available dye, and can be performed with CellTracker CM-DiI (Molecular Probe) as shown in the following examples, but is not limited thereto. It is not something.

  In the step and means for obtaining images obtained by photographing nerve cells according to the present invention over time, the photographing method is not particularly limited, but a confocal laser microscope with a built-in digital video camera normally used for observation of cell dynamics, etc. It is preferable to take a picture. In photographing over time, the frame rate as the photographing speed is not particularly limited, but is preferably about 1-10 frames / second, and particularly preferably 1-4 frames / second, from the viewpoint of the speed of particle axonal transport. In the step or means for inputting the obtained image frame to the computer, the input method is not particularly limited, and it can be input via a commercially available scanner or image processing software. If the image of the cell is obtained as a digital image, the input operation in this step becomes as easy as possible.

  In the process or means for extracting an axon region in an image frame according to the present invention, various methods can be applied. As an example, the axon region is a temporal moving average process of an image frame, Examples thereof include a binarization process for the moving average image, a labeling of each binarized area, and a method extracted by selecting the maximum area. More specifically, the axon region where particles move is emphasized by temporal moving average processing of image frames, and the background and axon part are separated from particle enhancement processing and binarization processing of image frames subjected to moving average processing. Then, each region in the binarized image frame is individually labeled, and by selecting the region with the largest area among the labeled regions, the noise region is deleted and the axon region is extracted Is possible. Other extraction methods can be applied to this step, and are not limited to the extraction method. In addition, the particle enhancement process applies a median filter, and further, a spike-shaped noise is eliminated by applying a circular particle enhancement process using a linear filter using a conical weight coefficient mask or a Gaussian filter. As a result, a clearer image frame of the axon region can be obtained.

  In the step and means for detecting vesicle particles in the axon region according to the present invention, the method for detecting vesicle particles is not particularly limited, but a detection method using a resolution filter is preferred. The separability filter obtains the degree of segregation between two regions and evaluates it to detect particles. Therefore, when the separability filter is used, the adverse effect of binarization processing is avoided, and the luminance value is reduced. Since detection of a low particle is enabled, it will become possible to detect as a separate particle also to some overlap of particles.

In an arbitrary region R, when the central region R ₁ and its peripheral region R ₂ are defined (see FIG. 1), δ _b ² is a luminance difference between the two regions in the following equation (3) for obtaining the separation η. And δ _T ^{2, which} is the denominator, evaluates the variance in each region. That is, Δ _b ² and Δ _T ² are defined by the following equations (4) and (5), respectively.

Therefore, when the degree of separation η takes a large value between the central region and the peripheral region and a small value between the peripheral regions, the central region is likely to be a particle. An expression of this relationship expressed by a mathematical formula is an evaluation formula H represented by the following formula (1). Accordingly, the separability filter, in the case of changing the area of the central region R ₁ of any region R in a plurality of stages, a small area of R ₁ to H of the following evaluation formula (1) takes the maximum value胞粒Ko It is preferable that the vesicle particles are detected by determining the area.

（但し、式(1)中のη₁₂、η₁₃、η₁₄、η₁₅は、それぞれ領域Rの中心領域R₁と周辺領域R₂
乃至R₅の各分離度を表す。）

(However, η ₁₂ , η ₁₃ , η ₁₄ , and η ₁₅ in the formula (1) are respectively the central region R ₁ and the peripheral region R _{2 of the} region R.
Each degree of separation of R ₅ is represented. )

  Subsequently, each component of the process and means for calculating the local tracking of the detected vesicle particle, the extraction of the stable tracking path, and the integration of the stable tracking path according to the present invention will be described below.

  According to the present invention, “local tracking” means that particles are tracked by matching each detected particle between frames of an image. Accordingly, the local tracking is performed by using the distance between the particles, the luminance, and the movement vector as indices as to which particle the particle detected in each image frame corresponds to in the frame at the next time. However, local tracking is an association based only on two pieces of continuous frame information. When the particle motion speed is remarkably high or when different particle movement paths are close to each other, the association becomes difficult and erroneous. The possibility of detection cannot be excluded. Therefore, for accurate path tracking, the following “stable tracking path extraction” and “stable tracking path integration” must be processed.

  According to the present invention, “stable tracking path extraction” means that a path with high possibility of false detection is removed from the paths obtained by the local tracking, and a stable tracking path with high confidence is extracted. Means that. Therefore, specifically, a partial tracking route is created using the tracking result obtained by the local tracking, and a portion that may be mistracked is removed by approaching or crossing particles of the partial tracking route. The operation is repeated for all routes, and all stable tracking paths are extracted.

  An example of the extraction of the stable tracking path is shown in FIG. In FIG. 2, the route indicated by the arrow is close to or intersects with another route, and the possibility of mistracking increases at such a location. Therefore, by determining based on the relationship with other routes, a stable tracking route is extracted by dividing the route into portions close to other routes and dividing the route into three partial tracking routes.

  According to the present invention, “integration of stable tracking paths” means that an evaluation function is set in consideration of continuity between stable tracking paths, continuity of shape characteristics of particles, etc., so that evaluation of each particle as a whole is optimal. In other words, it means integrating stable tracking paths. The processing method for integrating the stable tracking path is not particularly limited as long as it is a method capable of integrating the partial safety tracking path obtained by the extraction of the stable tracking path with suitability based on a certain evaluation criterion. It is preferably performed by an optimization algorithm suitable for handling combinatorial problems. Optimization algorithms include the steepest descent method, the simulated annealing method, the genetic algorithm, etc., but are executed according to the genetic algorithm in terms of preferentially searching for and selecting a more suitable route. It is preferable. When a genetic algorithm is used, specifically, it is possible to proceed with combinatorial optimization of the stable tracking path by genetic algorithm, with the stable tracking path as the gene and the evaluation function as the environmental fitness, but this is not limited to this. Is not to be done. The process flow of the genetic algorithm is shown in FIG.

  As shown in FIG. 3, when a genetic algorithm is applied, a simulated route (initial population) is obtained by recombining (integrating) partial stable tracking routes created by local tracking and stable tracking route extraction. ), Evaluation is performed using the evaluation function for the fitness of each simulated route, and repeated calculation processing is performed until an optimum vesicle particle path is obtained through selection and crossover or mutation operations.

In the case of using a genetic algorithm to integrate the stable tracking channel, said genetic algorithm, the evaluation function represented by the following formula (2) with the environment adaptability, stable tracking pathway genes, the stable tracking channel re It is preferable to search for the vesicle particle path so that the fitness of Equation (2) takes the minimum value using the combined simulated route as an individual.

（但し、式(2)では、n個の部分的追跡ルートから成る模擬ルートに対し、各部分追跡ルートの接点における傾きの差をH、X座標値の差をD、フレーム数の差をI、ルート全体の長さをLと定めて評価する。また、x,y,z及びwはそれぞれの評価に対する重みを表す。）

(However, in Equation (2), for a simulated route consisting of n partial tracking routes, the difference in slope at the point of contact of each partial tracking route is H, the difference in the X coordinate value is D, and the difference in the number of frames is I The length of the entire route is evaluated as L. Also, x, y, z, and w represent weights for the respective evaluations.)

  The definition of each parameter in the above equation (2) is further shown in FIG. Here, for each simulated route consisting of n partial tracking routes, H is the difference in slope, D is the difference in X coordinate values, I is the difference in the number of frames, and L is the difference in the number of frames. This is a parameter for evaluating the length. As described above, in the genetic algorithm using the evaluation function of Equation (2), it is determined that the smaller the value is, the better the route is.

  In the case of adopting a genetic algorithm, a more preferable simulated route is preferentially selected, and an optimal axon transport route is searched while repeating operations such as crossover and mutation. It is preferable to search for a route by one-point crossover.

  In the step and means for outputting the trajectory of the vesicle particle calculated by the arithmetic processing according to the present invention, the method for outputting the trajectory of the vesicle particle is not particularly limited. Can be easily output in the form of a graph or the like.

  FIG. 5 shows the overall processing steps of the processing method of the analysis method, analysis apparatus, and analysis program for nerve cell axonal transport according to the present invention.

  Hereinafter, the present invention will be described more specifically based on examples. However, the present invention is not limited to the following implementation.

Visualization of vesicle particles passing through nerve cell axons Culture Haiyowai day 7 dorsal root ganglia from chicken embryonic neural cells: (DRG D orsal R oot G anglion ) neurons, a glass bottom culture dishes coated with 100 [mu] g / ml of poly -L- lysine, F12 the basal medium, 10% fetal bovine serum (FBS: F etal B ovine S erum) and 10 ng / ml of nerve growth factor: in (NGF N eural G rowth F actor ) medium supplemented with at 37 ° C. Cultured.

2. Staining of vesicle particles After 8-14 hours from the start of the above culture, add 1 mg / ml CellTracker CM-DiI (Molecular Probe) to the culture, continue the culture for another 1 hour, and add Dil dye to DRG neurons. I took it in. Thereafter, the DRG neurons were washed 3 times with PBS. The washed DRG neurons were cultured at 37 ° C. for 6 hours or longer in a culture solution obtained by adding 10% FBS and 10 ng / ml NGF to L15 basal medium. After 6 hours from the start of the culture, the DRG neurons were observed with a confocal laser microscope, and it was confirmed that the fluorescently stained vesicles passed through the axons of the DRG neurons (FIG. 6).

3. Inhibition of axonal transport Furthermore, in order to confirm whether the DiI-stained vesicles are particles moving by axonal transport, nocodazole and sodium azide were used along microtubules. Were observed using a confocal laser microscope. Specifically, a confocal laser microscope was used to photograph 12 minutes of DiI-stained neurons at a frame rate of 2 frames / second, and microtubule polymerization was inhibited as an axonal transport inhibitor after 4 minutes from the start of imaging. The drug nocodazole was added at 30 μM, or the ATPase inhibitor sodium azide was added at 0.1%, and the course of subsequent axonal transport was observed. Inhibition of axonal transport is to set a virtual line that crosses the axon almost at right angles to the axon part about 20 μm away from the cell body, and count the number of particles that pass through the virtual line for 2 minutes, The number of particles was defined as an axonal transport activity index. The results of inhibition of axonal transport with nocodazole and sodium azide, respectively, are shown in FIG. 7. It was found that both inhibitors reduced axonal activity to nearly 0% compared to the activity before inhibition. (FIG. 7). Therefore, transport of particles visualized by DiI staining is reduced by microtubule polymerization inhibitors and ATPase inhibitors, so DiI staining stains vesicular particles transported by microtubule-dependent axonal transport. Confirmed to do.

Tracking of Axonal Transport Path of Vesicle Particles An axonal transport path of vesicle particles was tracked by the automatic tracking system for moving particles in the axon according to the present invention.

Detection of vesicle particles using a resolution filter The DIG neurons derived from chicken embryos stained with DiI in Example 1 above were photographed using a confocal microscope, and the images were converted into bitmap images. Image processing was performed. In this embodiment, a time-series image consisting of 240 frames is acquired and automatically analyzed with respect to one shooting data at a frame rate of 2 frames / second. In this example, three types of imaging data are used, (1) the number of detected particles when using a separability filter, (2) the number of detected particles by visual inspection, and (3) the number of detected particles by binarization processing. The particle detection method using the resolution filter was verified. The results are shown in Table 1 below.

  From the results shown in Table 1, it was confirmed that the particle detection method using the separation degree filter has a higher particle detection ability than the method using visual observation or binarization. In the detection stage, it is necessary to detect all possible particles even if it is noise or dust. Therefore, the present invention satisfies this condition, and the usefulness of the particle detection method using the resolution filter has been confirmed.

Comparison of tracking accuracy before and after introduction of genetic algorithm processing Next, the tracking accuracy of the tracking method using the genetic algorithm according to the present invention will be verified. As a verification method, the total number of routes was verified by comparing the tracking accuracy at the stage of applying the local tracking method with the tracking accuracy after introducing the genetic algorithm for the route with 87 routes. The fitness is evaluated using the evaluation function represented by the above formula (2), and based on the evaluation, the route for the next generation is selected, but in this embodiment, a certain generation has been reached using one-point crossover. Completion of calculation processing at the stage. The results are shown in Table 2, and the state before and after introduction is shown in FIG.

  From the results shown in Table 2, it was confirmed that the tracking accuracy was improved by applying the genetic algorithm. The improvement in the tracking accuracy is considered to be due to the fact that the mistracking due to the intersection and approach between particles that could not be overcome by the local tracking was improved by the optimal route selection by applying the genetic algorithm.

Comparison of path tracking accuracy by genetic algorithm processing and visual observation Finally, we verified whether particle tracking was performed in the same way as visual observation. In the conventional analysis method of axon transport, a virtual line is drawn on the axon located 10 μm from the point where the axon starts, and particles passing through the line are measured. The route of axon transport was followed. As for the automatic analysis, since the tracking result is displayed in a graph format as shown in FIG. 7, the X coordinate corresponding to the position of 10 μm is obtained from the cell body on the axon, and intersects the X coordinate on the graph. The number of routes was determined, and the number of routes was defined as particles that passed through the phantom line. Table 3 shows the results. From this experiment, it was confirmed that the analysis by genetic algorithm processing can be tracked to the same extent as visual observation.

  From the results in Table 3, in this example, almost the same results were obtained by the genetic algorithm processing and the path tracking of visual observation. Therefore, according to the automatic tracking system for moving particles in an axon according to the present invention, it was confirmed that tracking with the same accuracy as visual observation is possible.

FIG. 6 is a schematic diagram when a central region R1 and its peripheral regions R2-R5 are defined in an arbitrary region R by a separability filter. It is a graph which illustrates extraction of a stable tracking path. It is a flowchart of the process by the genetic algorithm used for this invention. It is a figure which shows the definition of each parameter of the evaluation formula of a genetic algorithm. It is a schematic diagram of the whole processing process of the processing method of the analysis method, analysis apparatus, and program for analysis based on this invention. It is a microscope picture of the DRG neuron derived from a chicken embryo by which the endoplasmic reticulum was fluorescently stained. It is a figure which shows the result of having inhibited axonal transport with the axonal transport inhibitor. It is a figure which shows the comparison of the tracking precision before and behind introduction of a genetic algorithm process.

Claims (19)

  A step of visualizing vesicle particles passing through a nerve cell axon, a step of obtaining an image of the nerve cell taken over time, a step of inputting the obtained image frame to a computer, an axon in the image frame Extracting a region, detecting vesicle particles in the axon region, computing local tracking of the detected vesicle particles, extraction of stable tracking paths and integration of stable tracking paths, and A method for analyzing axonal transport of nerve cells, comprising: outputting a trajectory of a vesicle particle calculated by the calculation process.   The method according to claim 1, wherein in the step of visualizing the vesicle particles passing through the nerve cell axons, the visualized vesicle particles are transported by microtubule-dependent axonal transport.   The method according to claim 1 or 2, wherein, in the step of visualizing the vesicle particle passing through the axon of the nerve cell, the particle is visualized by fluorescent labeling of a lipid bilayer membrane of the vesicle particle.   The method according to claim 3, wherein the fluorescent labeling of the lipid bilayer membrane is performed by carbocyanine fluorescent dye staining.   The method according to any one of claims 1 to 4, wherein in the step of detecting vesicle particles in the axon region, the vesicle particles are detected using a resolution filter. The separability filter, in the case of changing the area of the central region R ₁ of any region R in a plurality of stages, the area H of vesicular particles the area of R ₁ having the maximum value of the following evaluation formula (1) The method according to claim 5, wherein the vesicle particles are detected.

(However, η ₁₂ , η ₁₃ , η ₁₄ , and η ₁₅ in the formula (1) are respectively the central region R ₁ and the peripheral region R _{2 of the} region R.
Each degree of separation of R ₅ is represented. )   The step of calculating the local tracking of the detected vesicle particle, the extraction of the stable tracking path and the integration of the stable tracking path is calculated by an optimization algorithm of an evaluation function. The method according to any one of 1 to 6.   The method of claim 7, wherein the optimization algorithm is a genetic algorithm. The genetic algorithm uses the evaluation function represented by the following formula (2) as the environmental suitability, the stable tracking path is a gene, and the simulated route in which the stable tracking path is recombined is used as an individual to match the formula (2). 9. A method according to claim 8, wherein the vesicle particle pathway is sought so that the degree takes a minimum value.

(However, in Equation (2), for a simulated route consisting of n partial tracking routes, the difference in slope at the point of contact of each partial tracking route is H, the difference in the X coordinate value is D, and the difference in the number of frames is I The length of the entire route is evaluated as L. Also, x, y, z, and w represent weights for the respective evaluations.)   10. A method according to claim 8 or 9, wherein the genetic algorithm searches for the vesicle particle path by a single point crossover.   In the step of extracting an axon region in the image frame, the axon region is temporally moving average processing between image frames, binarization processing for the moving average image, and labeling of each binarized region And the method of any one of claims 1 to 10, extracted by selection of a maximum region.   Means for obtaining an image obtained by taking time-lapse images of nerve cells in which vesicle particles passing through axons are visualized, means for inputting the obtained image frame to a computer, means for extracting an axon region in the image frame Means for detecting vesicle particles in the axon region, means for computing local tracking of the detected vesicle particles, extraction of stable tracking paths and integration of stable tracking paths, and A device for analyzing axonal transport of nerve cells, comprising means for outputting the calculated trajectory of vesicle particles.   13. The apparatus of claim 12, wherein the means for detecting vesicle particles in the axon region is detected using a resolution filter. The separability filter, in the case of changing the area of the central region R ₁ of any region R in a plurality of stages, the area H of vesicular particles the area of R ₁ having the maximum value of the evaluation formula (1) The apparatus according to claim 13, wherein the vesicle particles are detected.   The step of calculating the local tracking of the detected vesicle particle, the extraction of the stable tracking path and the integration of the stable tracking path is calculated by an optimization algorithm of an evaluation function. The device according to any one of 12 to 14.   The apparatus of claim 15, wherein the optimization algorithm is a genetic algorithm. The genetic algorithm uses the evaluation function represented by the above formula (2) as the degree of environmental suitability, the stable tracking path is a gene, and the simulated route in which the stable tracking path is recombined is used as an individual. 17. An apparatus according to claim 16, wherein the vesicle particle path is sought so that the degree takes a minimum value.   The apparatus according to claim 16 or 17, wherein the genetic algorithm searches the vesicle particle path by a single point crossover.   In the means for extracting an axon region in the image frame, the axon region is temporally moving average processing between image frames, binarization processing for the moving average image, and labeling of each binarized region And an apparatus according to any one of claims 12 to 18, which is extracted by selection of a maximum area.

Images