Combination of Averaged Bregma-Interaural and Electrophysiology-Guided Technique Improves Subthalamic Nucleus Targeting Accuracy in Rats

Synthesis

This study presents and validates a novel electrophysiologically guided method (BITE) for implanting electrodes into the subthalamic nucleus (STN) in rats. It directly compares this method to two widely used stereotaxic approaches based on anatomical landmarks, one using bregma alone and the other using both bregma and interaural references. The results show a substantial increase in targeting accuracy: about five-fold improvement over the bregma-based method and two-fold over the bregma + interaural approach.

The work was reviewed by two Reviewers. They both agree that the method is innovative, addresses an important technical challenge, and is likely to improve experimental yield while reducing the number of animals used. They found the study to be well designed, statistically rigorous, and clearly described overall, with enough detail for replication. However, they identified several points that require clarification or additional analysis to strengthen the manuscript.

1) The Reviewers noted that the electrophysiological approach could be influenced by differences in neuronal excitability or firing patterns across genotypes or disease models (e.g. Parkinson's). This could lead an experimenter to reposition electrodes to obtain "typical" neuronal responses, potentially introducing bias. This limitation should be explicitly discussed.

2) The success rates reported for the bregma (17%) and bregma + interaural (40%) approaches appear low, but it is unclear how success was defined. The reviewers asked whether success refers to hitting an exact set of coordinates or simply being located within the STN. This needs to be clarified to contextualize the reported success rates.

3) The procedure for verifying electrode tip positions requires more explanation. In the Standardized protocol for electrode implantation in STN (points 1.4 and 1.5), the STN is described as a point defined by three coordinates rather than as a broader region. However, in practice, simply being within the STN may be sufficient for most experimental purposes without necessarily reaching the exact coordinate point. The difference in success rates between the two methods is striking, but depending on how "failure" is defined, many of these placements might still be considered acceptable in practice. This should be expanded upon.

4) Similarly, this raises the question of how "precise location" is determined histologically. The description provided is not entirely clear. Is verification based on the number of 10 μm sections examined and their comparison with images from Rat Brain in Stereotaxic Coordinates? On p. 7, it is stated that "targeting accuracy of both approaches was histologically verified through hematoxylin and eosin (H&E) staining to confirm electrode tip position," but this does not explain how the exact position was determined. The text also mentions that "we verified the exact electrode implantation sites using H&E staining and quantified the AP deviation by measuring the AP distance between the actual implantation site and the intended implantation site (3.7 mm posterior to bregma)." However, since bregma is not visible in histological slices, how is the "intended implantation site" inferred? Is a reference atlas or morphological mapping procedure used to estimate coordinates in tissue sections? If so, this needs to be described in detail. The subsection "Verification of electrode tip position" (p. 12) requires a much clearer explanation of the methodology.

5) The study claims that the operator followed established protocols without a learning curve, but this should be supported by data. Reviewers suggested showing that performance did not improve over time, for example with a logistic regression of success over trial order or a chi-square test comparing early vs late attempts.

6) The methods should describe how neural spike rates were calculated and how spectrograms were obtained, as this information is currently missing.

7) Visualization of data would be a welcome addition. For example, Table 1, Table 2, and animal weights could all be presented as violin plots, which would give us a much better understanding of the distribution of the data.

8) Figure 5 refers to the Zona Incerta (ZI). Figure 6 refers to the IC. Only IC is labelled in anatomical figures. Labels should correspond to exact outlined regions of interest in the images. In the text body, some introduction should be given to each region, explaining their relationship to one another, and relevance to the study.

9) Figure 6 only requires 1 example from STN and ZI so that the difference is immediately apparent to readers.

10) Section 1.6 mentions DVbregma and DVinteraural but does not explain how they were measured or used. Reviewers asked for a clear description of what is being measured and how these coordinates are compared or converted.

11) Section 1.3: "Skull drilling process generates high-frequency vibrations in the skull, which causes minor to significant bregma shift." The following claim is not substantiated by citation, nor data. An elaboration on how this happens and how this is known would be a welcome addition, especially for other surgeons in the field.

12) Section 1.3 (p. 10): The text states, "Such shift may significantly compromise the accuracy of the STN coordinates if skull drilling is executed after determining them in step 1.4 ('calculation of STN coordinates based on bregma')." while the calculation of STN coordinates based on bregma is actually described in step 1.5, not 1.4.

13) Page 19: The reference to Rat Brain in Stereotaxic Coordinates should be properly cited.

14) The subsection Verification of electrode tip position (p. 12) requires a much clearer explanation of the methodology.

15) Results, Experiment 3: It should be specified whether the dorsal and ventral sites were explicitly targeted or were simply reached while aiming for the standard coordinates while aiming for the standard three-coordinate target (determined by either the bregma- or bregma-interaural-based method). Clarification is needed to avoid ambiguity.

Author Response

Dear Reviewers, We sincerely appreciate the reviewers' time and insightful comments, which have significantly strengthened our manuscript. Their feedbacks have enhanced the clarity, scientific rigour, and overall impact of our work. In response to the reviewers' key concerns, we have conducted a major revision of the original manuscript. Below, we provide detailed point-by-point responses to each comment along with corresponding revisions in the manuscript.

1) The Reviewers noted that the electrophysiological approach could be influenced by differences in neuronal excitability or firing patterns across genotypes or disease models (e.g. Parkinson's). This could lead an experimenter to reposition electrodes to obtain "typical" neuronal responses, potentially introducing bias. This limitation should be explicitly discussed.

Reply: We appreciate the reviewers' comment and agree that this experimenter bias during electrode positioning may happen. To mitigate this bias, we have discussed this limitation in the revised manuscript, "since this study was performed in healthy rats, our approach may carry the risk of bias if electrode repositioning is guided by an assessment of 'typical' STN firing patterns in disease models or other genotypes, where STN activity may differ. To minimize this risk, we recommend that researchers first characterize the STN electrophysiological signatures in the specific disease model or genotype of interest before applying BITE for precise targeting. In addition, detailed characterization of STN firing features, including firing rate, amplitude, and burst duration, is warranted. Such physiological parameters may provide useful reference points for subjectively assessing accurate STN targeting." 2) The success rates reported for the bregma (17%) and bregma + interaural (40%) approaches appear low, but it is unclear how success was defined. The reviewers asked whether success refers to hitting an exact set of coordinates or simply being located within the STN. This needs to be clarified to contextualize the reported success rates.

Reply: We thank the reviewers for this clarification. To avoid ambiguity, we have revised the manuscript to state: "An implantation of the electrode was considered a "success" if the electrode tip was located anywhere within the anatomical boundaries of the STN, regardless of whether it matched the intended coordinates." 3) The procedure for verifying electrode tip positions requires more explanation. In the Standardized protocol for electrode implantation in STN (points 1.4 and 1.5), the STN is described as a point defined by three coordinates rather than as a broader region. However, in practice, simply being within the STN may be sufficient for most experimental purposes without necessarily reaching the exact coordinate point. The difference in success rates between the two methods is striking, but depending on how "failure" is defined, many of these placements might still be considered acceptable in practice. This should be expanded upon.

Reply: We thank the reviewers' comment, and we agree that "in practice, simply being within the STN may be sufficient for most experimental purposes without necessarily reaching the exact coordinate point." Thus, we revised the manuscript to state "An implantation of the electrode was considered a "success" if the electrode tip was located anywhere within the anatomical boundaries of the STN, regardless of whether it matched the intended coordinates".

4) Similarly, this raises the question of how "precise location" is determined histologically. The description provided is not entirely clear. Is verification based on the number of 10 μm sections examined and their comparison with images from Rat Brain in Stereotaxic Coordinates? Reply: We thank the reviewers for this comment. To avoid ambiguity, we have revised the manuscript to state: "The electrode tip position was confirmed through histological analysis. Following perfusion with phosphate-buffered saline and 4% paraformaldehyde, the brain was extracted, embedded in paraffin, and coronally sectioned at a thickness of 10 µm. Every section containing the visible electrode track was collected and stained with H&E. The precise location of the electrode tip was determined by aligning the histological sections with the corresponding sections in the Paxinos and Watson rat brain atlas (George Paxinos, 2007). This alignment was performed using some anatomical landmarks visible in both the H&E-stained sections and the atlas, including the third ventricles and the shape of hippocampus, optic fiber. Additionally, the H&E stained STN is in dark purple with densely packed neurons and is clearly delineated from the surrounding zona incerta and internal capsule (see Figure 5). These features facilitate the precise determination of the electrode tip location within STN when implantation is successful. An implantation of the electrode was considered a "success" if the electrode tip was located anywhere within the anatomical boundaries of the STN, regardless of whether it matched the intended coordinates. The AP coordinate of the electrode tip was obtained from the atlas (George Paxinos, 2007). In experiment 2, this AP coordinate was used to calculate the AP deviation as the absolute difference between the actual and intended coordinates." 5) On p. 7, it is stated that "targeting accuracy of both approaches was histologically verified through hematoxylin and eosin (H&E) staining to confirm electrode tip position," but this does not explain how the exact position was determined.

Reply: We agreed on this point. To avoid ambiguity, we added the explanation on the "Verification of electrode tip position" section that "...Every section containing the visible electrode track was collected and stained with H&E. The precise location of the electrode tip was determined by aligning the histological sections with the corresponding sections in the Paxinos and Watson rat brain atlas (George Paxinos, 2007). This alignment was performed using some anatomical landmarks visible in both the H&E-stained sections and the atlas, including the third ventricles and the shape of hippocampus, optic fiber. Additionally, the H&E stained STN is in dark purple with densely packed neurons and is clearly delineated from the surrounding zona incerta and internal capsule (see Figure 5). These features facilitate the precise determination of the electrode tip location when implantation within the STN is successful. An implantation of the electrode was considered a "success" if the electrode tip was located anywhere within the anatomical boundaries of the STN, regardless of whether it matched the intended coordinates ..." 6) The text also mentions that "we verified the exact electrode implantation sites using H&E staining and quantified the AP deviation by measuring the AP distance between the actual implantation site and the intended implantation site (3.7 mm posterior to bregma)." However, since bregma is not visible in histological slices, how is the "intended implantation site" inferred? Is a reference atlas or morphological mapping procedure used to estimate coordinates in tissue sections? If so, this needs to be described in detail. The subsection "Verification of electrode tip position" (p. 12) requires a much clearer explanation of the methodology.

Reply: Thank you for this comment. The intended implantation site was defined as 3.7 mm posterior to bregma. Bregma was visible and identified on the skull prior to electrode implantation, and the electrode was positioned accordingly (details please refer to the point 1.2 and 1.3 ). To clarify this point, we revised the manuscript to state "To achieve this goal, we verified the exact electrode implantation sites using H&E staining and quantified the AP deviation by measuring the AP distance between the actual implantation site and the intended implantation site (3.7 mm posterior to bregma). The actual implantation site in AP axis was determined by reference to the Paxinos and Watson atlas(George Paxinos, 2007)." Additionally, we revised the "Verification of electrode tip position" to "The electrode tip position was confirmed through histological analysis. Following perfusion with phosphate-buffered saline and 4% paraformaldehyde, the brain was extracted, embedded in paraffin, and coronally sectioned at a thickness of 10 µm. Every section containing the visible electrode track was collected and stained with H&E. The precise location of the electrode tip was determined by aligning the histological sections with the corresponding sections in the Paxinos and Watson rat brain atlas (George Paxinos, 2007). This alignment was performed using some anatomical landmarks visible in both the H&E-stained sections and the atlas, including the third ventricles and the shape of hippocampus, optic fiber. Additionally, the H&E stained STN is in dark purple with densely packed neurons and is clearly delineated from the surrounding zona incerta and internal capsule (see Figure 5). These features facilitate the precise determination of the electrode tip location within STN when implantation is successful. An implantation of the electrode was considered a "success" if the electrode tip was located anywhere within the anatomical boundaries of the STN, regardless of whether it matched the intended coordinates. The AP coordinate of the electrode tip was obtained from the atlas (George Paxinos, 2007). In experiment 2, this AP coordinate was used to calculate the AP deviation as the absolute difference between the actual and intended coordinates." 7) The study claims that the operator followed established protocols without a learning curve, but this should be supported by data. Reviewers suggested showing that performance did not improve over time, for example with a logistic regression of success over trial order or a chi-square test comparing early vs late attempts.

Reply: Thank you for this comment. To be more rigorous, we revised the manuscript to state: "Statistical analysis Success rate dynamics: To assess potential learning curve effects, we compared early versus later success rates for bregma-based approach and averaged bregma-interaural based approach. For the bregma-based method (2022), success rates were compared between the first and second six months. For the averaged bregma-interaural method (2023), we made the same comparison. Before analysis, expected cell frequencies in all contingency tables were checked: Fisher's exact test was used when any expected count was <5, otherwise Pearson's chi-square test was applied.

Results Success rate dynamics There was no significant difference in the success rate of the bregma-based approach between the first and second six-month periods. (0% vs 33%; p=0.206; Fisher's exact test). Similarly, the success rate of the averaged bregma-interaural approach did not differ significantly between the first and second six-month periods (44% vs 30%; p=0.449; Fisher's exact test)." 8) The methods should describe how neural spike rates were calculated and how spectrograms were obtained, as this information is currently missing.

Reply: Thank you for this important comment. To clarify, we revised the manuscript to state "Power Spectral Density We aimed to investigate the STN oscillations. To achieve this, we analyzed STN oscillations using spectral analysis. Raw data were collected at 30 kHz. Power spectral density was computed via short-time Fourier transforms with standard parameters (Hanning window, 50% overlap). This approach follows the methodology detailed in Deng et al (Deng et al., 2025).

Firing rate calculation Raw signals were band-pass filtered (300-3000 Hz) to isolate spikes. Spike detection was performed using a threshold based on the median absolute deviation, followed by template matching to refine spike selection. Spike times were used to calculate the spike count and firing rate (spikes per second)." 9) Visualization of data would be a welcome addition. For example, Table 1, Table 2, and animal weights could all be presented as violin plots, which would give us a much better understanding of the distribution of the data.

Reply: Thank you for this comment. To improve visualization, we have added violin plots to the manuscript (see Figure 6).

10) Figure 5 refers to the Zona Incerta (ZI). Figure 6 refers to the IC. Only IC is labelled in anatomical figures. Labels should correspond to exact outlined regions of interest in the images. In the text body, some introduction should be given to each region, explaining their relationship to one another, and relevance to the study.

Replay: We thank you for this important comment. To clarify, we revised the corresponding figures (see Figure 5 and Figure 7) to make it more clearer to the readers.

11) Figure 6 only requires 1 example from STN and ZI so that the difference is immediately apparent to readers.

Replay: Initially, we intended to present a single example from the STN and ZI to illustrate this result. However, we decided it is more informative to show all the data collected in this study, which not only ensures transparency for the readers but also fully addresses any doubts regarding the histology results.

12) Section 1.6 mentions DVbregma and DVinteraural but does not explain how they were measured or used. Reviewers asked for a clear description of what is being measured and how these coordinates are compared or converted.

Reply: Thank you for this comment. To improve clarity, we revised the manuscript to state "Using a micromanipulator and surgical microscope, we adjusted the electrode position until the tip was precisely aligned with the ear bar tip (see figure. 4A-C), the scales on the DV, ML, and AP arms were then recorded as AP', ML', and DV', representing the coordinates of the interaural zero. The STN coordinates based on inter-aural zero were measured as 5.3 mm anterior (APinteraural = AP'+ 5.3), 2.4 mm lateral for the right STN (MLinteraural = ML'- 2.4), and 1.65 mm superior (DVinteraural = DV'+ 1.65) to interaural zero, respectively (George Paxinos, 2007).

1.5 Calculations of STN coordinates based on bregma Using a surgical microscope, we carefully aligned the electrode as close as possible to the skull surface at bregma and recorded the scales as AP', ML' and DV' on the respective arms. According to the atlas (George Paxinos, 2007) and the previous experiences in targeting STN (Baunez and Robbins, 1997; Baunez et al., 2001; Winstanley et al., 2005; Siegle et al., 2017; Pelloux et al., 2018), the STN coordinates relative to bregma were determined as 3.7 mm posterior (APbregma=AP'- 3.7), 2.4 mm lateral for the right STN (MLbregma=ML'- 2.4), and 8.35mm ventral (DVbregma=DV'- 8.35).

1.6 Calculations of STN coordinates based on averaged bregma-interaural based approach.

We calculated the absolute values of difference between bregma based and interaural based STN coordinates as |APbregma - APinteraural|, |MLbregma - MLinteraural|, and |DVbregma - DVinteraural|. The difference thresholds were set to |APbregma - APinteraural| {less than or equal to} 0.5 mm, |MLbregma - MLinteraural| {less than or equal to} 0.3 mm, and |DVbregma - DVinteraural| {less than or equal to} 0.3 mm. If the calculated absolute values of difference were within the difference thresholds, we averaged the bregma-based and interaural-zero based coordinates to obtain the final STN coordinates: APbregma-interaural = (APbregma - APinteraural)/2, MLbregma-interaural = (MLbregma - MLinteraural)/2 and DVbregma-interaural = (DVbregma - DVinteraural)/2. If any differences exceeded the difference thresholds, we used the bregma-based coordinates for AP, ML or DV. " 13) Section 1.3: "Skull drilling process generates high-frequency vibrations in the skull, which causes minor to significant bregma shift." The following claim is not substantiated by citation, nor data. An elaboration on how this happens and how this is known would be a welcome addition, especially for other surgeons in the field.

Reply: Thank you for this comment. To enhance the clarity, we revised the manuscript to state, "The skull drilling process generates high-frequency vibrations and mechanical forces. We hypothesize that these forces could cause minor shift of the skull, potentially leading to a bregma shift. Although we have not yet quantified this phenomenon in this setup, we occasionally observed indications of such shifts during surgery. Thus, we recommend that operators perform the critical step of coordinate calculation (Step 1.5) after any major drilling to preemptively mitigate this potential source of error. Future work is needed to quantify this phenomenon." 14) Section 1.3 (p. 10): The text states, "Such shift may significantly compromise the accuracy of the STN coordinates if skull drilling is executed after determining them in step 1.4 ('calculation of STN coordinates based on bregma')." while the calculation of STN coordinates based on bregma is actually described in step 1.5, not 1.4.

Reply: Thank you for pointing out this mistake. We have revised this statement to "The skull drilling process generates high-frequency vibrations and mechanical forces. We hypothesize that these forces could cause minor shift of the skull, potentially leading to a bregma shift. Although we have not yet quantified this phenomenon, we occasionally observed indications of such shifts during surgery. Thus, we recommend that operators perform the critical step of coordinate calculation (Step 1.5) after any major drilling to preemptively mitigate this potential source of error. Future work is needed to quantify this phenomenon." in the manuscript.

15) Page 19: The reference to Rat Brain in Stereotaxic Coordinates should be properly cited.

Reply: Thank you for this comment. We revised this citation to "George Paxinos CW (2007) The Rat Brain in Stereotaxic Coordinates: Academic Press." 16) The subsection Verification of electrode tip position (p. 12) requires a much clearer explanation of the methodology.

Reply: Thank you for this comment. To enhance the clarity, we rewrote this section, "The electrode tip position was confirmed through histological analysis. Following perfusion with phosphate-buffered saline and 4% paraformaldehyde, the brain was extracted, embedded in paraffin, and coronally sectioned at a thickness of 10 µm. Every section containing the visible electrode track was collected and stained with H&E. The precise location of the electrode tip was determined by aligning the histological sections with the corresponding sections in the Paxinos and Watson rat brain atlas (George Paxinos, 2007). This alignment was performed using some anatomical landmarks visible in both the H&E-stained sections and the atlas, including the third ventricles and the shape of hippocampus, optic fiber. Additionally, the H&E stained STN is in dark purple with densely packed neurons and is clearly delineated from the surrounding zona incerta and internal capsule (see Figure 5). These features facilitate the precise determination of the electrode tip location within STN when implantation is successful. An implantation of the electrode was considered a "success" if the electrode tip was located anywhere within the anatomical boundaries of the STN, regardless of whether it matched the intended coordinates. The AP coordinate of the electrode tip was obtained from the atlas (George Paxinos, 2007). In experiment 2, this AP coordinate was used to calculate the AP deviation as the absolute difference between the actual and intended coordinates." 17) Results, Experiment 3: It should be specified whether the dorsal and ventral sites were explicitly targeted or were simply reached while aiming for the standard coordinates while aiming for the standard three-coordinate target (determined by either the bregma- or bregma-interaural-based method). Clarification is needed to avoid ambiguity.

Reply: Thank you for this comment. The electrodes were intended for general STN implantation, and our electrophysiology-based method for determining dorsoventral position (Step 1.7) is biased toward the dorsal STN. Specifically, we revised the manuscript to "...Electrophysiology plays a crucial role in BITE. In this study, the implanted electrode may typically pass through ZI before reaching STN. ZI is characterized electrophysiologically by near-silent neural activity, whereas the STN is distinguished by a burst-firing pattern with high-amplitude (0.5-1 μV), fast-bursty, and irregular spikes(Fluri et al., 2015; Deng et al., 2025). By halting the implantation immediately upon detecting signals indicative of the STN, we found that the electrode tip was positioned in dSTN in 50% of the rats, demonstrating the potential feasibility of using BITE for precise targeting in dSTN. Practically, electrodes were intended to be implanted into the STN in this study, and our electrophysiology-based method for determining the DV position of the electrode (step 1.7) was inclined toward the dorsal STN, as validated previously. This finding is noteworthy, as the dSTN is frequently studied as a key structure in the treatment of Parkinson's disease(Herzog et al., 2004). Moving forward, the ventral STN could be targeted by advancing the electrode when the STN-like firing pattern disappears, which could also theoretically allow for targeting the middle STN. This approach provides greater flexibility for STN targeting. However, although promising in theory, further studies are needed to confirm this strategy. ..."