Alternative Methods to Enhance the Axial Resolution of Total Internal Reflection Fluorescence–Structured Illumination Microscopy

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

In this paper, the authors present a new method for the enhancement of TIRF-SIM’s axial resolution by utilizing the quantitative information of the distance between fluorophores and the surface within the evanescent field. This is an interesting study. The paper can be accepted after the following issues addressed.

1. In Lines 70-73, the authors need to define the necessary and sufficient conditions for the generation of evanescent waves and provide appropriate references to support their statements.
2. In Lines 74-76, the authors also need to explain why the penetration depth of the evanescent wave is defined in this way, and provide the original source of this definition.
3. Please explain how Equation (2) is derived, and provide the corresponding steps or derivation.
4. The authors should label the incident light and the critical angle in Figure 1(a).
5. For the evanescent field, the maximum light intensity is located at δ = 0, and it decays exponentially with distance from this point. Given this, how can one determine whether the emission from a fluorophore is stronger than the intensity at δ = 0? Furthermore, how can this information be reliably used to estimate the spatial distribution and axial position of the fluorophores?
6. In addition to the textual explanations in the main body, it is recommended to add textual labels to the three different colored curves in Figures 2(a–c) for clarity.
7. The use of silica microspheres for calibration needs to be explained more clearly, specifying what exactly they are being used to calibrate.
8. The line 179 contains a mistake; it should refer to Figures 3c-d instead of Figures 4c-d.

Author Response

Comments 1:  In Lines 70-73, the authors need to define the necessary and sufficient conditions for the generation of evanescent waves and provide appropriate references to support their statements.

 

Response 1:  Thank you for the insightful comment. The generation of an evanescent wave requires two essential conditions:

• The incident light must travel from a medium with a higher refractive index to one with a lower refractive index.
• The incident angle must be equal to or greater than the critical angle for total internal reflection. In response to the comment, we have carefully added these conditions to the revised manuscript.

 

“[ When an incident beam travels from a medium with high refractive index (n1) to a medium with low refractive index (n2), total internal reflection occurs if the incident angle exceeds the critical angle, which is defined as θc = arcsin(n2/n1). This results in the generation of an evanescent wave (Figure 1a) . The evanescent wave excitation intensity at a distance z from the interface can be ex-pressed as follows:

 

where θ is the incident beam angle and I(0) is the intensity at the interface. The evanescent wave penetration depth d(θ)  is the distance from the interface at which the intensity decays to 1/e of I(0) [17]. ]”

 

Comments 2:  In Lines 74-76, the authors also need to explain why the penetration depth of the evanescent wave is defined in this way, and provide the original source of this definition.

 

Response 2: Thank you for the insightful comment. We sincerely apologize for not citing the source of the penetration depth formula in the original manuscript. This equation is derived from the article titled "Mapping fluorophore distributions in three dimensions by quantitative multiple angle-total internal reflection fluorescence microscopy." We have now added the corresponding reference in the revised version of the manuscript.

 

”[The evanescent wave penetration depth d(θ)  is the distance from the interface at which the intensity decays to 1/e of I(0) [17]. ]”

 

Comments 3:  Please explain how Equation (2) is derived, and provide the corresponding steps or derivation.

 

Response 3: Thank you for the insightful comment. We sincerely apologize for not clearly indicating the source of Equation (2) in the manuscript. This equation is derived from Reference [17]. Equation (2) represents a general formula for calculating fluorescence intensity. It describes the measured fluorescence intensity at a given (x, y) pixel in the image. Due to the dependence of the decay constant d on the incidence angle, the measured fluorescence intensity can be modeled to estimate the axial position of fluorophores. In this model, D_{x, y}(z) represents the fluorophore distribution along the z-axis at a fixed (x, y) position.

 

 

 

Comments 4:  The authors should label the incident light and the critical angle in Figure 1(a).

 

Response 4: Thank you for the insightful comment. We have added the indication of the incidence angle in Figure 1a .

 

 

Figure 1. Incident Schematic of lighting and reconstruction. (a) Incident light illuminating the sample at an angle greater than the critical angle, generating an evanescent field above the coverslip. (b) Relationship between the penetration depth and incident angle within 80°. (c) Flowchart of the reconstruction process. Scale bar: 10μm.

 

 

Comments 5:  For the evanescent field, the maximum light intensity is located at δ = 0, and it decays exponentially with distance from this point. Given this, how can one determine whether the emission from a fluorophore is stronger than the intensity at δ = 0? Furthermore, how can this information be reliably used to estimate the spatial distribution and axial position of the fluorophores?

 

Response 5: Thank you very much for your valuable comment. According to Equation (2), the fluorescence intensity reaches its maximum when δ = 0. In our microsphere experiment, the grayscale value at δ = 0 is indeed higher than at other positions, which supports this observation.

In this method, we exploit the fact that the decay constant d depends on the incidence angle, and estimate the axial position of fluorophores based on the measured fluorescence intensity. Building on this, we adopted the assumption of fluorophore sparsity and uniform distribution within the sample, following the approach of Sundd et al. (Reference 18 in the manuscript), who treated the fluorophores (fluorescent dyes) as being evenly distributed throughout the specimen.

 

 

Comments 6: In addition to the textual explanations in the main body, it is recommended to add textual labels to the three different colored curves in Figures 2(a–c) for clarity.

 

Response 6: Thank you for the insightful comment. We sincerely apologize for the inconvenience caused by the unclear labeling of Figures 2(a–c) in the original manuscript. We have added the indication of the incidence angle in Figure 2(a-c).  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 2. Microtubule simulation results. (a-c) Reconstruction results under a SNR of 30, 20, and 10. (d) RMSE of the reconstruction results from the ground truth under different SNR and depth values.

 

 

Comments 7: The use of silica microspheres for calibration needs to be explained more clearly, specifying what exactly they are being used to calibrate.

 

Response 7: Thank you for the insightful comment. We sincerely apologize for not providing a detailed explanation of the role of the microspheres. In the experimental validation section, we used microspheres with a known radius, whose outer surfaces were uniformly labeled with fluorescent dyes. This allowed us to calculate the fluorophore distribution on the outer shell of the sphere using the Pythagorean theorem. We then validated our proposed method by estimating the fluorophore distribution from the TIRF image and comparing the estimated result with the ground truth (Figure 3c, d). The method of using microspheres for calibration and validation has also been employed in MA-TIRF (MAIM). 

 

“[ The silica sphere, with its well-defined axial geometry, serves as a standard sample for calibration. As shown in Figure 3a, the sphere radius is 4.86±0.47μm. Based on this known radius, the axial distance of fluorophores located on the outer shell of the sphere relative to the coverslip surface was calculated using trigonometry. The computed axial distances were then compared with the true depths to further validate the model. ]”

 

 

Comments 8: The line 179 contains a mistake; it should refer to Figures 3c-d instead of Figures 4c-d.

 

Response 8: Thank you for the insightful comment. We sincerely apologize for overlooking this mistake. It has now been corrected in the revised manuscript:

 

“[ As shown in the Figures 3c-d, the reconstructed average height profile agrees well with the ground truth. ]”

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

I have read your manuscript and it read well. Some figures need improving ad they are not very clear and quite small. Figure 1 might benefit from been split into multiple figures to make the difference between the different setting more clear to the reader, as it stand the various images are too small. I have some questions about your conclusion as they di not match the rest of the articles, therefore some part is missing overall. You stated that your sample demonstrated broad compatibility with  standard fluorescent labeled samples, chemically fixed cells, cryo-preserved tissue sections  live cell preparation with fluorescent tags. As you are  not presenting any example of this samples, wither you remove this phrase from your conclusion completely, or you need to provide experimental samples to support your statement.

Author Response

Comments 1:  Figure 1 might benefit from been split into multiple figures to make the difference between the different setting more clear to the reader, as it stand the various images are too small.

 

Response 1: We sincerely apologize for the inconvenience caused by the layout of Figure 1. We have carefully considered your valuable suggestion and attempted to split Figure 1 into two separate parts in the revised manuscript. However, after multiple trials, the split layout did not produce a satisfactory result. We apologize again for not being able to implement your suggestion effectively.

To improve the overall readability of the figures, we have made some adjustments: we added a marker for the critical angle of total internal reflection in Figure 1, and we have also added clearer labels in Figure 2.

Thank you once again for your insightful comments — they are truly important for improving the quality of our work.

 

Comments 2:  I have some questions about your conclusion as they did not match the rest of the articles, therefore some part is missing overall. You stated that your sample demonstrated broad compatibility with standard fluorescent labeled samples, chemically fixed cells, cryo-preserved tissue sections live cell preparation with fluorescent tags. As you are not presenting any example of this samples, wither you remove this phrase from your conclusion completely, or you need to provide experimental samples to support your statement.

 

Response 2: Thank you very much for your thoughtful suggestion — this is indeed an important point. We sincerely apologize for the inappropriate statement in the conclusion section. We discussed the applicability of our method to live-cell imaging but failed to provide corresponding experimental evidence to support that claim, which may have caused confusion. We deeply regret this oversight.

In fact, our method relies solely on TIRF raw images, and the entire reconstruction process is performed under the 2D TIRF-SIM mode. Therefore, we believe that the temporal resolution of our method is comparable to that of SIM. Given SIM’s established high temporal resolution, we assumed that our method would also be suitable for live-cell imaging.

Nevertheless, your comment is absolutely valid. It was indeed a mistake on our part to assert this compatibility without direct experimental validation. We sincerely apologize.

Due to current experimental constraints, we are not able to conduct live-cell imaging experiments at this time. However, we fully recognize the importance of this aspect and will make it a key focus in our future work. We have edited the corresponding statements in the conclusion section of the manuscript accordingly.

 

“[ In this paper, we present an alternative method for achieving 3D super-resolution imaging. Unlike previous methods, our approach does not require a complex optical system or high-power lasers. We achieved axial super-resolution based on TIRF-SIM without the need for a multi-angle illumination system in MA-TIRF, thus avoiding errors caused by angle or mode switching. We conducted calibration experiments on silica sphere with known axial depths. The deviation between the axial distribution reconstructed by this method and the true distribution of the microspheres < 25 nm. Additionally, we calculated the RMSE of the axial depths of synthetic microtubules under different SNR. These experiments confirmed the accuracy of the proposed method. Although the axial resolution varied slightly under different imaging conditions—such as changes in emitter brightness, fluorophore density, and sample depth—it remained consistently within 25 nm across a range of signal-to-noise ratios (SNRs) and axial positions. These results demonstrate the robustness of the method under practical conditions and highlight its potential for high-precision axial localization in complex biological samples. Because of its simple reconstruction process, our method can be combined with existing commercial TIRF SIM microscopes. Beyond the resolution enhancement, the proposed method preserves the intrinsic advantages of TIRFM through an evanescent field. A shallower excitation depth of the evanescent field implies that a larger out-of-focus background is removed during the imaging process. Fewer incident angles results in a higher reconstruction speed, which also leads to lower photobleaching and phototoxicity and higher temporal resolution. The samples demonstrated broad compatibility with standard fluorescently labeled specimens, including, but not limited to, chemically fixed cells and cryo-preserved tissue sections.

The proposed method prioritizes simplicity over theoretical resolution, making it particularly suitable for 3D super-resolution in commercial TIRF microscopes. However, due to experimental limitations, we were unable to perform axial localization experiments on live-cell samples. In addition, owing to the limitations of the traditional SIM algorithm used for fast reconstruction, the current reconstruction speed is primarily constrained by the SIM reconstruction rate. In future work, we aim to improve the proposed method by incorporating faster lateral super-resolution techniques and conducting reconstruction experiments on live-cell samples. These will serve as key directions for further development. We hope that the method presented in this paper can provide a different perspective on TIR-based 3D super-resolution. ]”

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

This manuscript proposes an approach for enhancing axial resolution in total internal reflection fluorescence structured illumination microscopy (TIRF-SIM), leveraging quantitative intensity-distance relationships within the evanescent field to reconstruct axial depth information. By combining lateral super-resolution capabilities inherent to TIRF-SIM with computational reconstruction, the authors achieve an improved three-dimensional imaging resolution.

 

While the presented technique appears well-executed from a practical standpoint, the manuscript falls short in several critical aspects necessary for publication in Photonics, which prioritizes substantial advances in photonic principles, theoretical rigor, and methodological innovation in optical imaging.

 

Major Concerns:

1. The manuscript directly maps fluorescence signal intensity to axial height without adequately accounting for distortions caused by uneven fluorophore distributions within the sample. Consequently, intensity variations due to differences in labeling density could be mistakenly interpreted as changes in height within real biological specimens, significantly compromising the accuracy of axial localization.
2. The authors' modeling of the evanescent field is overly simplified, without addressing the impact of practical factors such as optical stray light, background noise, and multiple scattering on the achievable axial resolution.
3. The approach proposed by the authors is conceptually similar to established methods such as multi-angle interference microscopy (MAIM) or variable-angle total internal reflection fluorescence microscopy (VA-TIRF), but the manuscript fails to clearly articulate the unique advantages of their method over prior work.
4. The authors did not employ standard samples with known axial structures or independent measurement techniques to validate the claimed improvement in axial resolution, resulting in a lack of direct experimental evidence to support their assertion of achieving 50 nm axial resolution.
5. The authors utilized a pre-trained Weka classifier for structure segmentation but did not specify the training data or the generalization capability of the classifier. Inaccurate classification or inappropriate threshold selection could directly affect the accuracy of the axial reconstruction results.
6. The manuscript lacks direct comparative experiments with advanced methods; the authors did not benchmark their results against established high-performance axial imaging techniques, such as 3D-SIM or advanced VA-TIRF, making it difficult to assess the relative advantages or limitations of their approach compared to the current state-of-the-art.
7. In summary, the primary contribution of this work lies in proposing a practical approach to enhancing axial resolution in TIRF-SIM by leveraging intensity-based depth reconstruction. However, the theoretical foundation is overly simplified, the innovation beyond existing techniques such as MAIM and VA-TIRF is limited, and the experimental validation lacks the rigor necessary to convincingly support the claimed performance improvements. Furthermore, without direct comparisons to advanced axial imaging methods, it remains difficult to assess the true advantages of the proposed approach.Given these considerations, I believe that the overall level of theoretical advancement and methodological innovation is currently insufficient to meet the publication standards of Photonics. Substantial improvements in theoretical modeling, experimental validation, and comparative analysis would be necessary to make this work suitable for consideration in a high-quality optics journal.

Author Response

Comments 1:  The manuscript directly maps fluorescence signal intensity to axial height without adequately accounting for distortions caused by uneven fluorophore distributions within the sample. Consequently, intensity variations due to differences in labeling density could be mistakenly interpreted as changes in height within real biological specimens, significantly compromising the accuracy of axial localization.

 

Response 1: Thank you very much for your insightful comments. You are absolutely right—non-uniform fluorophore distributions are often present in real biological samples, and this can indeed affect the accuracy of axial localization. Incorporating the fluorophore distribution into the model without compromising computational efficiency will be an important focus of our future work.

However, in the present study, our goal was to develop a broadly applicable and straightforward axial localization method. To this end, we made the simplifying assumption that the fluorophores are uniformly distributed. This approach follows the precedent set by Sundd et al. in their study titled "Quantitative dynamic footprinting microscopy reveals mechanisms of neutrophil rolling," published in Nature Methods in 2010. In this work, the authors introduced quantitative dynamic footprinting (qDF) microscopy based on the principle of total internal reflection fluorescence (TIRF). By measuring the decay of evanescent wave intensity with depth, combined with microfluidic and magnetic tweezer setups, they quantitatively analyzed the contact interface between rolling neutrophils and the substrate.

Specifically, to facilitate mathematical modeling and derivation, Sundd et al. simplified the complex real-world scenario to a two-layer model consisting of the glass coverslip and the cytoplasm, assuming a uniform distribution of fluorophores within the cytoplasm. During experimental validation, they compared axial localization results using both GFP expression and DiI staining. They found consistent distance measurements at the front and rear ends of the cells, which indirectly confirmed the reliability of the method based on the uniform distribution assumption. The validation results can be found in Supplementary Figure 12 of their paper.

We sincerely apologize for not citing this article in the original manuscript. Thank you once again for your valuable suggestion — it is truly important to our work. We have revised the relevant statements in the manuscript and added the corresponding reference accordingly.

 

“[ where Dx, y(z) is the axial distribution of fluorophore. Under the simple two-layer model, this distribution is assumed to be spatially uniform, i.e., Dx, y(z) = D [18].

18. Sundd, P., Gutierrez, E., Pospieszalska, M. et al. Quantitative dynamic footprinting microscopy reveals mechanisms of neutrophil rolling. Nat Methods 7, 2010, 821–824. ]”

 

 

 

 

Comments 2: The authors' modeling of the evanescent field is overly simplified, without addressing the impact of practical factors such as optical stray light, background noise, and multiple scattering on the achievable axial resolution.

 

Response 2: Thank you very much for your valuable comments. We sincerely apologize for the confusion caused by our unclear wording, and we have added the relevant references in the revised manuscript for clarification. It is true that our current model does not explicitly account for optical stray light, background noise, or multiple scattering, and the reasons are as follows:

Our method is based on 2D TIRF-SIM, where the penetration depth of TIRF is typically around 200 nm. Within this shallow imaging range, multiple scattering is generally negligible.

Optical stray light is mainly a concern in label-free imaging. In fluorescently labeled samples, this effect is usually absent. To support this, we refer to the work of Fu et al. published in PNAS, in which axial super-resolution was achieved using multi-angle TIRF microscopy (MA-TIRF) combined with sequential imaging and photobleaching. By subtracting images taken at different incident angles, they measured the diameters of fluorescence “rings,” and then used the known radius of microspheres along with the Pythagorean theorem to calculate the penetration depth. This work also indirectly demonstrated that within the limited penetration depth of TIRF, multiple scattering is negligible and does not cause significant errors.

The limited penetration depth of TIRF naturally restricts the excitation of fluorophores to a narrow axial range, which helps suppress background noise. Furthermore, before performing axial localization, we apply a binary mask extracted from the SIM image to filter out background fluorescence (Figure 1c).

Once again, we appreciate your helpful suggestions. While our model does not explicitly address all potential factors affecting signal-to-noise ratio, we did assess the method’s robustness under various noise levels in the simulation (Section 3.1). Specifically, we added Gaussian noise to simulate TIRF images with different SNRs (10, 15, 20, 25, and 30), and evaluated the axial localization error. As shown in Figure 2, our method maintains acceptable prediction accuracy even at SNR = 10.

 

Reference：Y. Fu, P.W. Winter, R. Rojas et al. Axial superresolution via multiangle TIRF microscopy with sequential imaging and photobleaching. Proc. Natl. Acad. Sci. U.S.A. 2016, 16, 4368-4373.

 

 

Comments 3:  The approach proposed by the authors is conceptually similar to established methods such as multi-angle interference microscopy (MAIM) or variable-angle total internal reflection fluorescence microscopy (VA-TIRF), but the manuscript fails to clearly articulate the unique advantages of their method over prior work.

 

Response 3: Thank you for the insightful comment. We sincerely apologize for not clearly explaining the advantages of our method over VA-TIRF in the Introduction. In summary, our approach can be applied on any commercial TIRF microscope and involves a simple workflow that only requires capturing TIRF SIM images without switching the incident angle. This avoids potential errors introduced during mode switching. We have revised the manuscript accordingly.

 

”[ Despite its popularity, SIM's relatively low spatial resolution compared to other SRM techniques is its biggest drawback. Nevertheless, a recent work demonstrated that by combining sparsity and continuity features in fluorescence imaging, SIM can achieve a lateral spatial resolution of 60 nm and achieve fast super-resolution imaging at 564 Hz in live cell [11]. However, the improvements in axial resolution enhancement still lag be-hind existing methods in terms of lateral resolution.

In our previous works based on variable-angle total internal reflection fluorescence microscopy (VA-TIRF), we explored the axial distribution of fluorophores by acquiring TIRF images at different incident angles. Based on VA-TIRF, we developed multi-angle interference microscopy (MAIM) and demonstrated that by obtaining a TIRF image stack at 20 TIRF excitation angles, the axial resolution of SIM can be enhanced to 20 nm. Hence, a 3D reconstruction of fluorophores can be achieved [12,13]. Despite improved axial resolution, the complex multi-angle illumination system may affect both the temporal resolution and operational efficiency [14-16]. ]”

 

 

Comments 4:  The authors did not employ standard samples with known axial structures or independent measurement techniques to validate the claimed improvement in axial resolution, resulting in a lack of direct experimental evidence to support their assertion of achieving 50 nm axial resolution.

 

Response 4: Thank you for the insightful comment. We sincerely apologize for not providing a detailed explanation of the role of the microspheres. In the experimental validation section, we used microspheres with a known radius, whose outer surfaces were uniformly labeled with fluorescent dyes. This allowed us to calculate the fluorophore distribution on the outer shell of the sphere using the Pythagorean theorem. We then validated our proposed method by estimating the fluorophore distribution from the TIRF image and comparing the estimated result with the ground truth (Figure 3c, d). The method of using microspheres for calibration and validation has also been employed in MA-TIRF (MAIM). 

 

“[ Next, we performed an experimental calibration using an Alexa Fluor 488-labeled silica sphere deposited centrally on a coverslip surface and immersed in a refrac-tive-index-matched medium. The silica sphere, with its well-defined axial geometry, serves as a standard sample for calibration. As shown in Figure 3a, the sphere radius is 4.86±0.47μm. Based on this known radius, the axial distance of fluorophores located on the outer shell of the sphere relative to the coverslip surface was calculated using trigo-nometry. The computed axial distances were then compared with the true depths to further validate the model. The system we used was a typical TIRFM, which was inverted on a vibration-isolation stage. The system was based on a Nikon-Ti microscope using three polarized excitation beams of 488, 561, and 639 nm and collimated using a coupler (Nikon LU4A) and collimation lens (CL) [12]. TIRF-SIM imaging was performed using a 100X/NA1.49 Nikon objective lens to acquire nine raw SIM images. ]”

 

 

Comments 5:  The authors utilized a pre-trained Weka classifier for structure segmentation but did not specify the training data or the generalization capability of the classifier. Inaccurate classification or inappropriate threshold selection could directly affect the accuracy of the axial reconstruction results.

 

Response 5: Thank you very much for your valuable suggestion. We sincerely apologize for the unclear expression that may have caused misunderstanding. Since Weka is a machine learning framework, it can be pretrained or trained on a single image. The reason this is feasible is that Weka’s feature maps are not generated by CNNs but are instead derived using native Fiji methods. For example, Gaussian-blurred versions of the input image with different σ parameters serve as different feature maps, representing features across various directions and scales.

In our work, we use these Gaussian-blurred images with different σ values as feature maps and apply Sobel filters to extract edge features, etc., thereby avoiding issues related to the generalization ability of trainable models across different organelles. These feature maps are then fed into classifiers such as decision trees, KNN, clustering, or regression algorithms to produce the final pixel classification results. We have corrected the original description in the manuscript accordingly.

 

(s)

“[To avoid potential artifacts arising from out-of-focus background signals that may confound the axial reconstruction by introducing spurious intensity values, TIRF-SIM was used as a pre-processing step before reconstruction, as shown in Figure 1c. Since SIM image may distort the intensity information during the reconstruction process, TIRF image is needed to obtain axial information. Trainable Weka Segmentation (TWS) was used for structure extraction [18]. TWS utilizes Fiji’s built-in feature-extraction modules to generate multi-layer features, including Gaussian blur filters with varying kernel sizes and Sobel filters. The resulting feature maps are then fed into a classifier. These features are con-catenated into high-dimensional vectors to train the classifier, ultimately yielding a binary mask. ]”

 

 

Comments 6:  The manuscript lacks direct comparative experiments with advanced methods; the authors did not benchmark their results against established high-performance axial imaging techniques, such as 3D-SIM or advanced VA-TIRF, making it difficult to assess the relative advantages or limitations of their approach compared to the current state-of-the-art.

 

Response 6:

Thank you very much for your valuable suggestion. We sincerely apologize for not benchmarking our results against 3D-SIM or VA-TIRF in the manuscript.

In fact, prior to this work, we conducted two studies related to axial localization based on VA-TIRF: MAIM and STARII. These previous efforts provided a foundation and technical insight that motivated the development of the simplified single-angle approach presented in this manuscript:

 

12. Chen, Y.; Liu, W.; Zhang, Z. et al. Multi-color live-cell super-resolution volume imaging with multi-angle interference microscopy. Nat Commun. 2018, 9, 4818.
13. Liu, W.; Kuang, C.; Yuan, Y.; Zhang, Z.; Chen, Y.; Han, Y.; Xu, L.; Zhang, M.; Zhang, Y.; Xu, Y.; Liu, X. Simultaneous Two-Angle Axial Ratiometry for Fast Live and LongTerm Three-Dimensional Super-Resolution Fluorescence Imaging. J Phys Chem Lett. 2019,  10, 7811-7816.

 

Building upon our previous work, we introduced improvements to the VA-TIRF approach by simplifying the modeling process, thereby enabling axial localization using a single incident angle. To ensure the feasibility of our proposed method, we conducted calibration experiments using fluorescent microspheres with known axial structures.

In alignment with the methodology described in STARII, we calculated the deviation between the reconstructed axial depth and the actual depth. The following is a detailed description of the calibration experiments performed in STARII using fluorescent beads:

Figure 2. Experimental calibration of the STARII theory and STARII and STARII-SIM images of microtubules. (a) Schematics of a silica microsphere whose surface was labeled with a monolayer of Alexa Fluor 488 dye deposited on a coverslip. (b) Representative conventional TIRFM images (top), wide field image (bottom left), and reconstructed STARII image (bottom right). (c) Experimental data (orange dots) and theoretical depth profile (gray line) from the center of the sphere (top). Localization deviation (residues) between the experimental and theoretical data (bottom).

As mentioned in the STARII paper:
“The average deviation was <30 nm, most <20 nm, over a depth range from 0 to 300 nm, which was slightly lower than the simulated analysis, possibly reflecting the physical roughness and labeling non-uniformity of the sphere as well as the experimental localization ability of our method (Figure 2c, bottom).”
  This is consistent with the validation procedure described in Section 3.2.1 of our manuscript. Using the same type of fluorescent microspheres, we calculated the average deviation between the reconstructed and true axial depths. Our results show that the average deviation is as follows:

 

Figure 3. Experimental calibration of the silica sphere. (a) Schematic of the sphere with radius R and depth δ. (b) Reconstruction results representing a maximum depth of 220 nm. (c) Comparison between the ground truth and the reconstruction results. (d) Axial resolution obtained from the calibration. Scale bar in (b): 5μm.

 

 

Comments 7: In summary, the primary contribution of this work lies in proposing a practical approach to enhancing axial resolution in TIRF-SIM by leveraging intensity-based depth reconstruction. However, the theoretical foundation is overly simplified, the innovation beyond existing techniques such as MAIM and VA-TIRF is limited, and the experimental validation lacks the rigor necessary to convincingly support the claimed performance improvements. Furthermore, without direct comparisons to advanced axial imaging methods, it remains difficult to assess the true advantages of the proposed approach. Given these considerations, I believe that the overall level of theoretical advancement and methodological innovation is currently insufficient to meet the publication standards of Photonics. Substantial improvements in theoretical modeling, experimental validation, and comparative analysis would be necessary to make this work suitable for consideration in a high-quality optics journal.

 

Response 7:

Thank you very much for your insightful suggestions. Building upon our previous work, MAIM, we have improved the multi-angle illumination strategy. The primary aim of this study is to develop a simpler and more practical method for axial localization. Previously, we implemented both a 20-angle MAIM reconstruction approach and a 2-angle STARII method. Based on these experiences, we believe that a single-angle illumination scheme offers greater potential for broader adoption on commercial microscope systems due to its simplicity and compatibility.

In the experimental validation phase, we conducted both fluorescent microsphere calibration and synthetic microtubule simulations. Specifically, we followed the validation strategy used in MAIM to calculate the root-mean-square error of synthetic microtubules under various signal-to-noise ratios. In addition, inspired by the STARII method, we evaluated the axial deviation of microspheres at different positions to demonstrate the effectiveness of our approach.

Therefore, this method places greater emphasis on the advantages in terms of speed and cost. In future work, we will further improve the reconstruction accuracy on the basis of high-speed reconstruction. We hope to share these results in a future submission to Photonics.

Author Response File: Author Response.pdf

Round 2

Reviewer 3 Report

Comments and Suggestions for Authors

The authors have revised the paper according to the comments one by one. This version is ready for publication now.