Ventral skull based approach to whole-brain extraction in mice

Abstract

Head fixation combined with cranial window, GRIN lens, or prism implantation are common techniques in systems neuroscience for real-time imaging of neural activity. A critical step in these experiments is the removal of these implants and subsequent histological processing to assess tissue integrity and verify opsin expression. However, removing implants from the dorsal surface of the skull often causes traumatic damage to the underlying cortex. This study introduces a novel technique for intact brain excision that removes dorsal implants while preserving cortical integrity. Instead of the conventional dorsal approach, which often risks cortical damage, the skull is resected from the ventral side, allowing the implant to remain in place for post-mortem analysis.

This approach:

• •Enables dorsal implants to remain embedded in the brain during fixation.
• •Improves reproducibility by standardizing the extraction method, reducing variability introduced by traditional dorsal implant removal.
• •Preserves tissue integrity that would otherwise be compromised by removing the implant from fresh brain tissue.

Graphical abstract

Specifications table

Subject area	Neuroscience
More specific subject area	Methods in Neuroscience
Name of your method	Ventral Excision of Entire Mouse Brain
Name and reference of original method	Excision of intact mouse brain
Resource availability	N/A

Background

The extraction of the mouse brain is a critical step for post-mortem histological analysis. Standard protocols typically excise the brain dorsally through the skull [1]. This approach, however, is complicated by the presence of cranial implants commonly used in neuroscience research, including cannulas, cranial windows, and gradient refractive index (GRIN) lenses [2,3]. Such implants must often be removed prior to brain extraction, which can result in cortical tissue damage.

To overcome this limitation, ventral excision techniques have been developed. In this approach, the brain is removed ventrally, allowing fixation in sucrose while remaining attached to a dorsally mounted head-plate. This method permits the preservation of implants in situ during fixation, thereby maintaining tissue integrity and preventing structural damage associated with premature removal. After fixation, the implant can be carefully removed with minimal disruption to surrounding tissue.

This strategy preserves neural architecture, facilitating high-resolution imaging and more accurate assessment of implant placement and tissue responses to experimental interventions. Furthermore, ventral excision can also be applied to non-implanted brains in contexts where intact dorsal structures are required, since it avoids blade contact with the dorsal surface during excision. This manuscript presents a ventral excision method for brain removal and validates the technique for extracting GRIN lens and cranial window implants from the primary visual cortex (V1) in mice.

Required Materials

• •Utility Scissors (Fine Science Tools, Item No. 37500-00)
• •Iris Scissors - ToughCut® (Fine Science Tools, Item No. 14058-09)
• •Hartmann Hemostat (Fine Science Tools, Item No. 13002-10)
• •Adson-Baby Hemostat (Fine Science Tools, Item No. 13013-14)
• •Student Vannas Spring Scissors (Fine Science Tools, Item No. 91500-09)
• •Heavy Operating Forceps (Fine Science Tools, Item No. 16060-11)
• •Semi-Micro Spatula (Fisherbrand, S50822)

Method details

• 1.Remove head from body (Fig. 1.A.) by cutting across the neck at the cervical spine after transcardial perfusion using Utility Scissors (Fine Science Tools, Item No. 37500-00).
• 2.Cut skin from base of neck on the ventral side along the midsagittal plane to the snout (Fig. 1.B.) using Iris Scissors (Fine Science Tools, 14058-09) and holding the head in place using Adson-Baby Hemostat (Fine Science Tools, 13013-14).
• 3.Starting from the existing incision on the ventral side, cut under the skin to expose the muscle in small sections using Iris Scissors while peeling back skin using Hartmann Hemostat (Fine Science Tools, 13002-10).
• 4.Sever masseter muscle along its circumference on the sagittal plane where it connects to the skull (Fig. 1.E.) using Iris Scissors.
• 5.Cut mandibular condyle on the transverse plane where the jaw connects to the skull (Fig. 1.F.) using Iris Scissors.
• 6.Cut lower jaw down the midsagittal plane from between the bones of the lower incisors to the muscle of the base of the neck (Fig. 1.G.) using Iris Scissors.
• 7.Remove each half of the lower jaw by gripping one incisor at a time using the Adson-Baby Hemostat and the skull using the Hartmann Hemostat, then pulling the lower jaw bone diagonally away from the skull.
• 8.Remove the tongue by cutting across the transverse plane where the muscle attaches to the skull using Iris Scissors.
• 9.
  Cut the skull on the midsagittal plane

  • A.
    Insert one blade of the Student Vannas Spring Scissors (Fine Science Tools, 91500-09) into the brain stem at the base of the skull and maneuver it so the sharp side of the blade makes contact with the occipital bone located ventrally on the skull.
  • B.
    Cut shallowly through the skull, being careful to cut only skull and muscle, and continue on the midsagittal plane through the palatine bone (Figure 1.H.) using the Student Vannas Spring Scissors.

    • i.
      Do not push the blade dorsally into the brain tissue.
  • 10.
    Starting from the existing incision on the ventral side, maneuver and peel back muscle and skin until the skull is visible along the midsagittal plane using Heavy Operating Forceps (Fine Science Tools, 16060-11).
• 11.
  Expose the brain

  • A.
    If there is no head plate:

    • i.
      Starting from the existing incision on the ventral side, resect the skull piece by piece using Student Vannas Spring Scissors to cut, Heavy Operating Forceps to remove bone fragments, and Adson-Baby Hemostat to hold the skull in place by the snout, until the circumference of the brain is exposed but the brain remains in the skull.
    • ii.
      Detach the brain matter from the skull using the Semi-Micro Spatula (Fisherbrand, S50822) to pull back the brain tissue along its circumference from the skull walls and sever any meninges using Student Vannas Spring Scissors.
    • iii
      Remove the entire brain from the skull carefully (Fig. 1.I.) using Semi-Micro Spatula.
    • iv.
      The brain is exposed and ready to be fixed for histology.
  • B.
    If head-plate is attached, leave it attached:

    • i.
      Starting from the existing incision on the ventral side, resect skull and muscle piece by piece using Student Vannas Spring Scissors to cut, Heavy Operating Forceps to remove bone fragments, and Adson-Baby Hemostat to hold the skull in place by the headplate, until the entire surface of the brain is exposed, aside from the head-plated section.
    • ii.
      The brain is exposed, while leaving the head-plate attached to the dorsal surface of the brain, and ready to be fixed for histology (Fig. 1.J.).

Fig. 1.

Ventral excision method of the entire mouse brain. The figure illustrates the stepwise approach to brain removal through the ventral aspect of the skull. After careful exposure and excision of the ventral cranial bones, the brain can be mobilized and removed in one piece. This method highlights the anatomical orientation and procedural steps required to achieve complete extraction while preserving overall structural integrity.

Histology and imaging procedure

All mice were anesthetized with 5 % isoflurane and ketamine and transcardially perfused with phosphate-buffered saline (PBS) followed by 4 % paraformaldehyde (PFA). Brains were extracted using the method described herein, post-fixed overnight in PFA at 4°C, rinsed three times in PBS, and cryoprotected in 30 % sucrose for 1–2 days until they sank, indicating complete cryoprotection [4,5]. Coronal sections (40 μm) of V1 were obtained using a sliding microtome (American Optical Company) frozen with dry ice and ethanol to ensure proper adhesion and sectioning. Sections were stained with DAPI, mounted onto glass slides, air-dried, and coverslipped with Vectashield mounting medium (Vector Laboratories). Imaging was performed on an Olympus MVX10 stereo microscope with a 2 × and 0.63 × objective lens, 1.8 × zoom, 0.5 numerical aperture, and 500 ms exposure under 455 nm illumination.

Curvature extraction pipeline

Images were converted to grayscale, Gaussian blurred (σ = 2.0), and segmented using Otsu thresholding with removal of small objects and holes (≥ 5,000 px) to isolate the brain tissue boundary [6,7]. The outer brain contour was extracted as the longest closed polyline. A Savitzky–Golay filter was applied only for visualization [8], while curvature calculations were derived exclusively from raw contours to prevent smoothing bias.

To reduce pixel-level noise before curvature estimation, contour coordinates were processed with a moving-average filter (window = 5 points) and lightly downsampled (every 2 points). Curvature κ was computed along the arc-length–parameterized boundary using the standard differential-geometry formulation:

$$\kappa \left(t\right)=\mid x^{\prime }\left(t\right)y^{″}\left(t\right)-y^{\prime }\left(t\right)x^{″}\left(t\right)\mid /\left(x^{\prime }\left(t\right)2+y^{\prime }\left(t\right)2\right)3/2$$

where r(t)=(x(t),y(t)) denotes the boundary parameterization. For each slice, curvature standard deviation and mean absolute deviation were extracted as quantitative measures of boundary irregularity.

Statistical analysis

Curvature-based cortical perimeter smoothness was quantified from coronal brain-slice TIFF images in ventral and dorsal extraction groups. Section-level curvature metrics were aggregated per mouse to avoid pseudoreplication, and group differences in curvature standard deviation (curvature SD) were assessed using the Mann–Whitney U test (two-tailed). All statistical analyses were performed in Python, and confidence intervals were computed from per-mouse summary values. The full analysis pipeline is available in the GitHub repository:

https://github.com/quintanad0610-spec/-Ventral-Skull-Based-Approach-to-Whole-Brain-Extraction-in-Mice.git

Method validation

The ventral skull-based brain removal technique was evaluated for extraction of GRIN lens (n = 5; 2 males; 3 ventral extractions) and cranial window implants (n = 6; 2 males; 3 ventral extractions) mice and compared with the conventional dorsal approach.(Figs. 2 and 4). To minimize observer bias, both authors independently performed brain extractions. To objectively assess tissue integrity at the slice margins, an automated image-based curvature analysis was applied across all samples (Methods; Table 1, Table 2, Figs. 3 and 5).

Fig. 2.

Representative images of V1 brain slices processed using ventral and dorsal brain extractions in mice with GRIN lens implant. Coronal sections from a mouse implanted with a GRIN lens into V1, illustrating the outcomes of ventral (A) and dorsal (B) excision approaches. These representative slices demonstrate preserved tissue architecture following ventral removal technique compared to dorsal approach. Scale bars = 500 µm.

Fig. 4.

Representative images of V1 brain slices processed using ventral and dorsal brain extractions in mice with cranial window implants. Coronal sections from mice with cranial window implant over V1, highlighting the results of ventral (A) and dorsal (B) excision methods. These examples demonstrate preservation of cortical and subcortical structures with the ventral removal strategy compared to the dorsal approach. Scale bars = 1 mm.

Table 1.

Summary of mean curvature SD values for each mouse, averaged across all segmented slices in the GRIN lenses comparison included in the analysis.

Mouse	Mean Curvature SD	Standard Deviation	95 % CI	Number of Slices Per Mouse
Ventral 1	0.38935	0.237414	0.242 – 0.536	10
Ventral 2	0.323155	0.031938	0.303 – 0.343	6
Ventral 3	0.411567	0.053540	0.378 – 0.445	9
Dorsal 1	0.716124	0.214368	0.583 – 0.849	10
Dorsal 2	0.832832	0.129929	0.752 – 0.914	9

Table 2.

Summary of mean curvature SD values for each mouse, averaged across all segmented slices in the cranial window implant comparison included in the analysis.

Mouse	Mean Curvature SD	Standard Deviation	95 % CI	Number of Slices Per Mouse
Ventral 1	0.462947	0.158482	0.365 – 0.561	9
Ventral 2	0.700904	0.273537	0.531 – 0.871	9
Ventral 3	0.428562	0.215484	0.295 – 0.562	9
Dorsal 1	0.954969	0.769361	0.478 – 1.432	9
Dorsal 2	1.060764	0.735720	0.604 – 1.518	9
Dorsal 3	0.836551	0.347335	0.621 – 1.052	9

Fig. 3.

Curvature analysis of brain slices following ventral versus dorsal excision in mice implanted with a GRIN lens. (A) Representative coronal slice obtained using the ventral skull-based removal approach. The blue outline indicates the algorithm-detected brain boundary used for curvature extraction. The corresponding plot to the right shows the curvature (κ) computed along the contour, with the shaded region representing ±1 SD around the mean. (B) Representative slice obtained using the dorsal removal method, showing a more irregular excision contour. The curvature profile on the right shows greater variability in κ along the boundary compared to the ventral example. Scale bar = 500 µm (C) Per-mouse comparison of mean curvature variability (curvature SD, κ). Each point represents one mouse, reflecting the average curvature SD across all slices for that animal. Ventral excision produced significantly smoother slice boundaries than dorsal excision (*p < 0.05, Mann–Whitney U test). Boxplots show medians and interquartile ranges; whiskers extend to 1.5 × IQR.

Fig. 5.

Cortical curvature analysis following ventral versus dorsal excision in mice with cranial window implants. (A) Representative cortical slice obtained using the ventral skull-based removal approach. The blue outline indicates the algorithm-detected cortical boundary used for curvature extraction. The plot to the right shows the curvature (κ) computed along the boundary, with the shaded region marking ±1 SD around the mean curvature. (B) Representative cortical slice following the dorsal removal approach. Compared with the ventral method, the dorsal excision produces a more irregular cortical edge, reflected by larger fluctuations in the curvature profile. Scale bar = 1 mm (C) Per-mouse comparison of curvature variability (curvature SD, κ). Each point represents one mouse, corresponding to the mean curvature SD across all slices for that animal. Ventral cortical removal produced significantly smoother boundaries than dorsal cortical removal (*p < 0.05, Mann–Whitney U test). Boxplots show medians and interquartile ranges; whiskers extend to 1.5 × IQR.

For GRIN lens implants, the ventral approach yielded substantially smoother slice boundaries than the dorsal method, with a lower mean curvature SD (0.375; 95 % CI: 0.262–0.488; n = 3 mice) compared with dorsal extraction (0.774; 95 % CI: 0.034–1.515; n = 2 mice, p= 0.0487, Mann–Whitney U test). For cranial window (cortical) implants, a similar pattern was observed: ventral excision produced lower curvature variability (mean curvature SD: 0.531; 95 % CI: 0.163–0.899; n = 3 mice) relative to the dorsal approach (0.951; 95 % CI: 0.673–1.229; n = 3 mice, p= 0.045, Mann–Whitney U test). These results consistently demonstrate that ventral skull-based removal preserves cortical and subcortical tissue geometry more effectively than the traditional dorsal method.

Limitations

This technique has several limitations. Its success relies heavily on the histologist’s technical expertise, which may influence reproducibility across laboratories with varying levels of experience. Additionally, although the ventral excision method improves preservation of cortical tissue adjacent to implants, it requires more time: approximately 15 minutes compared to the approximately 5 minutes needed for the classical dorsal extraction. This increased duration may limit its practicality in high-throughput settings or when rapid tissue processing is required.

Moreover, the method was demonstrated only in adult mice, and its feasibility for fetal mice or larger species such as rats or ferrets remains untested. All extractions were performed unblinded, introducing the possibility of operator bias in execution or assessment. Finally, curvature-based smoothness metrics provide an indirect surrogate for tissue integrity and should be interpreted in conjunction with conventional histological assessments. Despite these constraints, the ventral approach offers a practical, tissue-preserving alternative for brains bearing dorsal implants and may facilitate improved downstream imaging, electrophysiology, and anatomical analyses.

Related research article

D. B. MacManus, Excision of whole intact mouse brain, MethodsX, Volume 10 (2023) 102246, https://www.sciencedirect.com/science/article/pii/S2215016123002431

Ethics statements

All procedures in this study were conducted in accordance with the Animal Care and Use Committee of UC Berkeley and the National Institutes of Health guide for the care and use of laboratory animals (NIH Publications No. 8023, revised 1987).

CRediT author statement

Alexandra N. Buntin-Nakamura, BA: Conceptualization, Methodology, Investigation, Writing - Original Draft. Daniel Quintana, BA: Validation, Formal Analysis, Writing - Review and Editing, Visualization

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

Data will be made available on request.