The post-mortem material was gathered from the brain of a 61-year-old woman who died from a pluri-metastatic colorectal cancer with no signs of neurological, neurodegenerative or psychiatric diseases. After a post-mortem delay of 5.5 h, the brain was perfused through the internal carotids and the vertebral arteries with 4 L of cold saline solution (NaCL 0.9% to which 20 units/mL of heparin was added), followed by 6 L of a cold fixative solution containing 4% paraformaldehyde (PFA) diluted in phosphate buffer (PB, 0.1 M, pH 7.4) and 4 L of 4% PFA to which 0.3% of glutaraldehyde was added. We then performed a MRI scan of the head with a 3T Philips Achieva (TE: 17 ms; TR:4000 ms; slice thickness: 500 μm; duration 8 h, see Supplementary Material) before brain extraction. The extracted brain weighed 1,328 g (Figures 1A–F). The brainstem was dissected out and post-fixed by immersion in 4% PFA for 4 days. Tissue samples were obtained from the Human Anatomy Laboratory at Université Laval and kept in the Human Brain Bank of the CERVO research Center. Both Institutions required informed consent before donation of tissues. The Ethics Committee at Université Laval approved the brain collecting procedures, as well as the storage and handling of post-mortem human brain tissues.

The brainstem was dissected and cut along a sagittal plane at 1 mm from the midline (Figure 1G). From the right hemi-brainstem that contained the midline, four different blocks were cut along the ponto-mesencephalic junction plane (PMJ), a line that connects the inferior aspect of the quadrigeminal plate (qpc) posteriorly with the foramen caecum (Fc) of the interpeduncular fossa (ipf) anteriorly (Figure 1H). The PMJ is depicted by a red line on the different plates. Its exact position can be found on Figure 1H and on plate 10 (0.00 mm). We estimate that there exists a difference of 10 degrees between the PMJ plane and the ac-pc plane (Figure 1H) and 20 degrees between the PMJ plane and the one used in the atlas of Paxinos (Paxinos et al., 2020). Each of the blocks were then cut at 4°C into 50 μm-thick sections using a vibratome (Leica VT1200S). The sections were all collected serially in phosphate buffer saline (PBS, 0.1 M, pH 7.4). Pairs of adjacent sections were selected out of 22 sections and then processed for Cresyl Violet and Luxol Fast Blue, respectively, providing a mean interval of 1.1 mm between each plate. Because some sections were damaged during the cutting process, in some cases, we decided to choose the next undamaged section instead of the adjacent one, for a maximum of 400 μm interval between Cresyl Violet and Luxol Fast Blue stained sections. Because losses of tissue inevitably occurred between each block, in addition to the known section thickness and intervals, we referred to the MRI in order to confirm the exact position of each section, relative to the PMJ (Figure 1I).

Sections intended for Cresyl Violet staining were first mounted on gelatin-coated microscope slides, air dried at room temperature and incubated overnight in 95% ethanol at 56°C. The slides were then rinsed 15 times in distilled water and immersed for 3 min at room temperature in a pre-heated and filtered solution of Cresyl Violet acetate (catalog no. C5042, Sigma, St-Louis, MO, USA) dissolved at 0.1% in distilled water. Sections were then rinsed in distilled water followed by 10 times in ethanol 95%, two times in a solution of ethanol/acid acetic 0.5%, five times in ethanol 95%. The rinses in ethanol 95% and ethanol/acid acetic were repeated until the desired staining intensity was obtained. Sections were dehydrated in ethanol and coverslipped with Permount (Permount Mounting Medium, catalog no. SP15, Thermo Fisher Scientific, Waltman, MA, USA).

Sections intended for Luxol Fast Blue staining were first mounted on gelatin-coated microscope slides, air dried at room temperature and rinsed 10 times in ethanol 95%. They were then incubated overnight at 56°C in Luxol Fast Blue (Solvent Blue 38, catalog no. S3382, Sigma, St-Louis, MO, USA) dissolved at 0.1% in ethanol. They were rinsed once in ethanol 95%, 10 times in lithium carbonate (0.01%) and 10 times in ethanol 70%. When needed, rinses in lithium carbonate (0.01–0.1%) and ethanol were repeated until the desired staining intensity was obtained. Sections were dehydrated in ethanol and coverslipped with Permount. To increase contrast, five sections stained for Luxol Fast Blue had to be also stained with Cresyl Violet (Plates 9, 10, 26, 31, 37).

Eight additional equally-spaced sections taken trough the midbrain and the pons (between +4.88 mm and −10.59 mm) were stained for choline acetyltransferase (ChAT), the enzyme responsible for the synthesis of acetylcholine (plates A-H). After three rinses in PBS, the sections were placed for 20 min at room temperature in hydrogen peroxide (3% H₂O₂ diluted in ethanol) to eliminate endogenous peroxidase activity. The free-floating sections were then placed in sodium borohydride (0.5% diluted in PBS) for 30 min. After three rinses in PBS, the sections were preincubated for 1 h at room temperature in a PBS solution containing 2% normal rabbit serum and 0.5% Triton X-100, and incubated 48 h at 4°C in the same solution to which the goat anti-ChAT antibody was added (catalog no. AB144P, EMD Milipore Corporation, Billerica, MA, 1:20). The sections were then rinsed, reincubated for 1 h at room temperature with a rabbit anti-goat biotinylated IgG (catalog # BA-5000; Vector Labs, Burlingame, CA, USA; 1:200). After three rinses in PBS, the sections were reincubated for 1 h at room temperature in 2% avidin-biotin-peroxidase complex (catalog # PK-4000; Vector Labs). The bound peroxidase was revealed by placing the sections in a medium containing 0.05% 3,3′-diaminobenzidine tetrahydrochloride (DAB, catalog #D5637; Sigma) and 0.005% H₂O₂ in 0.05 M Tris buffer, pH 7.6. The reaction was stopped after 8 min by several washes in Tris buffer and PBS. Immunostained sections were mounted on gelatine-coated slides, dehydrated in ethanol and coverslipped with Permount.

Cresyl Violet, Luxol Fast Blue and ChAT-stained sections were digitalized at 1200 dpi (pixel resolution of 1 μm) using a slide scanner (TISSUEscopeTM 4000, Huron Technologies, Waterloo, Ontario, Canada) equipped with a 10X objective.

Chemical fixation of the brain and section processing inevitably lead to shrinkage. In order to minimize shrinkage, blocks were cut at 4°C with a vibratome, precluding the need of cryoprotection in sucrose solution and freezing. A comparison of the size of the red nucleus as it appears on post-mortem brain MRI images and on Cresyl Violet stained sections, has revealed the shrinkage to be of the order of 4%. No pre-mortem MRI was available to assess shrinkage that might have been caused by the perfusion. It should be reminded that this atlas is based on a single brain of a 61-year-old woman and that existing inter-individual variations should carefully be taken into account using provided MRI (see Supplementary Material). Three different magnifications had to be used to provide plates presenting histological sections of adequate size (plates 1–6, plates 7–29, and plates 30–45). Therefore, readers should pay close attention to individual scale bars when comparisons are made between plates. Segmentation of brainstem nuclei and fiber tracts from Cresyl Violet and Luxol Fast Blue stained sections were performed following careful examination of histological sections, so as to avoid arbitrary delineation. Only structures that could be easily delineated in our preparations are identified and dashed lines are used when brainstem regions don't show clear histological boundaries. For example, the cuneiform nucleus (CnF) and the tegmental part of pontine reticular nucleus (PnTn) are broadly delineated with dashed lines because their boundaries with surrounding structures could not be clearly established. Therefore, the CnF, as defined in the present study, might comprise a portion of the mesencephalic reticular formation, whereas the PnTn could include parts of the isthmic reticular formation, the retroisthmic nucleus, the retrorubral field and the ventrolateral tegmental nucleus, as defined by Paxinos (Paxinos et al., 2020). Our segmentation was confirmed with the help of other human brainstem atlases (Olszewski and Baxter, 1954; Schaltenbrand and Wahren, 1977; Afshar et al., 1978; Paxinos and Huang, 1995; Naidich et al., 2007; Paxinos et al., 2020). Readers might refer to other human brainstem atlases for more extensive delineation of brainstem subnuclei and additional plates, in different sectioning planes. The nomenclature and abbreviations used in the present study are largely based on those of Paxinos and Watson (1982). For brainstem nuclei that didn't show clear subdivisions, we used single abbreviations to identify structures composed of several subnuclei. For example, the abbreviation Sp5 for spinal trigeminal nucleus, as delineated in the present atlas, includes the oral (Sp5O), interpolar (Sp5I), and caudal (Sp5C) parts of the nucleus, and the abbreviation VC for ventral cochlear nucleus includes the anterior (VCA) and posterior (VCP) parts of the nucleus that are delineated in other atlases. Likewise, the abbreviation PAG for periaqueductal gray refers to all PAG columns. A complete list of abbreviations and equivalent Latin names, as published in Terminologia Anatomica (Terminology, 1998) is also provided. The asterisks added to some abbreviations indicate that these abbreviations are slightly different from those used in the human brainstem atlas of Paxinos (Paxinos et al., 2020) or point to structures that are not identified in that atlas.