Magnetic resonance diffusion tensor imaging for the pedunculopontine nucleus: proof of concept and histological correlation

Abstract

The pedunculopontine nucleus (PPN) has been proposed as target for deep brain stimulation (DBS) in patients with postural instability and gait disorders due to its involvement in muscle tonus adjustments and control of locomotion. However, it is a deep-seated brainstem nucleus without clear imaging or electrophysiological markers. Some studies suggested that diffusion tensor imaging (DTI) may help guiding electrode placement in the PPN by showing the surrounding fiber bundles, but none have provided a direct histological correlation. We investigated DTI fractional anisotropy (FA) maps from in vivo and in situ post-mortem magnetic resonance images (MRI) compared to histological evaluations for improving PPN targeting in humans. A post-mortem brain was scanned in a clinical 3T MR system in situ. Thereafter, the brain was processed with a special method ideally suited for cytoarchitectonic analyses. Also, nine volunteers had in vivo brain scanning using the same MRI protocol. Images from volunteers were compared to those obtained in the post-mortem study. FA values of the volunteers were obtained from PPN, inferior colliculus, cerebellar crossing fibers and medial lemniscus using histological data and atlas information. FA values in the PPN were significantly lower than in the surrounding white matter region and higher than in areas with predominantly gray matter. In Nissl-stained histologic sections, the PPN extended for more than 10 mm in the rostro-caudal axis being closely attached to the lateral parabrachial nucleus. Our DTI analyses and the spatial correlation with histological findings proposed a location for PPN that matched the position assigned to this nucleus in the literature. Coregistration of neuroimaging and cytoarchitectonic features can add value to help establishing functional architectonics of the PPN and facilitate neurosurgical targeting of this extended nucleus.

Keywords: Deep brain stimulation; Diffusion tensor imaging; Histology; Pedunculopontine nucleus.

Fig. 1

Workflow of post-mortem and live controls data management

Fig. 2

Workflow of the digital processing of the post-mortem specimen: magnetic resonance images, high field histological images and block-face pictures

Fig. 3

a Olszewsky & Baxter Histological Atlas showing brainstem nuclei at the level of the pons: locus coeruleus (gray); nucleus cuneiformis (pink); PPN (red) nucleus parabrachialis lateralis (yellow); inferior colliculus (light purple); nucleus sagulum (dark blue); medial lemniscus (light blue); lateral lemniscus (dark purple) and cerebellar crossing fibers (green). Pictures of the brainstem in lightfield (b) and darkfield (c), axial flair (d) and FA colored map (e)

Fig. 4

3D reconstruction of the PPN, red nucleus (RN), locus coeruleus (LC), aqueduct (AQ), medial lemniscus fibers (ML) and cerebellar crossing fibers (CCF) passing around the nuclei. On the bottom right side, the brainstem indicating the plane of section (PS)

Fig. 5

Comparison between FA values of the inferior colliculus (IC), pedunculopontine nucleus (PPN), cerebellar crossing fibers (CCF) and medial lemniscus (ML): mean FA values: IC = 0.61; PPN = 0.67; CCF = 0.77; ML = 0.82

Fig. 6

Plates 1–2 representing brainstem with pedunculopontine area. Nissl-stained (430 μm thick) showing dark and light field microscopy, and the mask with PPN region. The histological set of images was oriented in the axial plane, parallel to h1 and the B–F line, as defined by Afshar (Afshar et al. 1978). The h1 line emerges in an acute angle (15°) at the crossing point of the projection of the AC-PC line and the projection of the ventricular floor line (VLF) and runs in a rostral direction. Stereotactic coordinates of the PPN before registration to MRI are given relatively to this system, as considered to be more reliable than the system based on PC, due to the variability of the mesencephalic angle. Coordinates are given in millimeters (mm) Positive values for X (lateral) are given to the right PPN as positive Y (anteroposterior) values are anterior to B point (point where the VLF meets a perpendicular line originated from Fastigium) and positive Z (rostrocaudal) values are rostral to B point. Relative to PC, rostral PPN is situated at 8.2 mm below PC (−8.2 mm) and caudal PPN at −14.2 mm for both sides. For comparison purposes, rostrocaudal values are also displayed in relation to PC. 15,430 μm slides contain the PPN, but 8 slides in 860 μm were chosen to demonstrate the histological localizations. They are named Plate 1 to Plate 8, increasing in number in the caudal-rostral direction. Plates 3–4, 5–6 and 7–8 are shown in Figs. 6, 7 and 8, respectively. The reference system adopted and overall location of plates within the brainstem is depicted in Fig. 9

Fig. 7

Plates 3–4. See Fig. 2 caption for details

Fig. 8

Plates 5–6. See Fig. 2 caption for details

Fig. 9

Plates 7–8. See Fig. 2 caption for details

Fig. 10

Reference System used for Figs. 2, 6, 7 and 8. See Fig. 2 caption for details