A computational atlas of the hippocampal formation using ex vivo, ultra-high resolution MRI: Application to adaptive segmentation of in vivo MRI

Abstract

Automated analysis of MRI data of the subregions of the hippocampus requires computational atlases built at a higher resolution than those that are typically used in current neuroimaging studies. Here we describe the construction of a statistical atlas of the hippocampal formation at the subregion level using ultra-high resolution, ex vivo MRI. Fifteen autopsy samples were scanned at 0.13 mm isotropic resolution (on average) using customized hardware. The images were manually segmented into 13 different hippocampal substructures using a protocol specifically designed for this study; precise delineations were made possible by the extraordinary resolution of the scans. In addition to the subregions, manual annotations for neighboring structures (e.g., amygdala, cortex) were obtained from a separate dataset of in vivo, T1-weighted MRI scans of the whole brain (1mm resolution). The manual labels from the in vivo and ex vivo data were combined into a single computational atlas of the hippocampal formation with a novel atlas building algorithm based on Bayesian inference. The resulting atlas can be used to automatically segment the hippocampal subregions in structural MRI images, using an algorithm that can analyze multimodal data and adapt to variations in MRI contrast due to differences in acquisition hardware or pulse sequences. The applicability of the atlas, which we are releasing as part of FreeSurfer (version 6.0), is demonstrated with experiments on three different publicly available datasets with different types of MRI contrast. The results show that the atlas and companion segmentation method: 1) can segment T1 and T2 images, as well as their combination, 2) replicate findings on mild cognitive impairment based on high-resolution T2 data, and 3) can discriminate between Alzheimer's disease subjects and elderly controls with 88% accuracy in standard resolution (1mm) T1 data, significantly outperforming the atlas in FreeSurfer version 5.3 (86% accuracy) and classification based on whole hippocampal volume (82% accuracy).

Figure 1

sample sagittal (a), coronal (b) and axial (c) slices from the ex vivo data of Case 8. Sample sagittal (d), coronal (e) and axial (f) slices from the ex vivo MRI data of Case 14. In (e), two regions of the slice are zoomed in to better appreciate the level of resolution of the scan (0.1 mm). Note that the acquisition of Case 8 was carried out in a bag, whereas Case 14 was scanned in a tube.

Figure 2

Eight coronal slices from Case 14 and corresponding manual annotations. The slices are ordered from anterior to posterior. Sagittal and axial slices, as well as 3D renderings of the manual segmentation are shown in the supplementary material (Figure 15, Figure 16 and Figure 17).

Figure 3

In vivo dataset and comparison with ex vivo images. (a) Sagittal slice in vivo. (b) Corresponding manual delineation of brain structures; note that the hippocampus (in yellow) is labeled as a single entity. (c) Coronal slice in vivo. (d) Corresponding manual delineation. e) Close-up of the hippocampus (in yellow) on a sagittal slice in vivo. f) An approximately corresponding slice from Case 12 of the ex vivo dataset. (g) Close-up of the hippocampus on a coronal slice in vivo. (h) An approximately corresponding slice from Case 12 (ex vivo).

Figure 4

Illustration of the generative model of the manual labels for ex vivo (top) and in vivo (bottom) MRI.

Figure 5

corresponding coronal slices (from anterior to posterior) of the label probabilities derived from the proposed ex vivo (top two rows) and original in vivo (middle two rows) atlases, as well as the UPenn atlas (Yushkevich, et al., 2009) (bottom two rows). For the FreeSurfer atlases, the color at each voxel is a linear combination of the colors assigned to the substructures, weighted by the corresponding probabilities. For the UPenn atlas, the color corresponds to the label with highest probability at each location. The color legend for the hippocampal subregions is the same as in Figure 2. The color code for the surrounding structures is displayed in the figure – note that the in vivo atlas uses generic labels for the gray matter, white matter and cerebrospinal fluid structures. Sagittal and axial slices of the atlases are provided as part of the supplementary material (Figure 18 and Figure 19).

Figure 6

Illustration of the generative model of MRI images (monomodal data).

Figure 7

Sample images (T1 and T2 weighted) and manual segmentations of “subject 1” from (Winterburn, et al., 2013). (S1–S4) Sagittal slices, from medial to lateral. (C1–C9) Coronal slices, from anterior to posterior. (R1) 3D rendering of manual segmentation, anterior view. (R2) 3D rendering, inferior view.

Figure 8

Sample coronal slices of “subject 3” from (Winterburn, et al., 2013), from anterior (left) to posterior (right). Top row: T1 image. Second row: T2 image. Third row: segmentation computed with the T1 scan. Fourth row: segmentation computed with T2 scan. Fifth row: segmentation computed with T1 and T2 scans simultaneously, overlaid on the T2 images. Bottom row: manual segmentation from the original study. The red arrow marks a CSF pocket, the blue arrows mark the molecular layer, and the yellow arrow marks the medial digitation.

Figure 9

Sample sagittal slices of “subject 3” from (Winterburn, et al., 2013), from medial (left) to lateral (right). See caption of Figure 8 for an explanation of the different rows. The red arrow marks a CSF pocket, the blue arrows mark the molecular layer, and the yellow arrows mark segmentation errors in the whole hippocampal shape.

Figure 10

3D renderings of segmentations of the high-resolution T1/T2 data. (a) Manual segmentation from (Winterburn, et al., 2013), anterior view. (b) Automated segmentation using T1 and T2 volume simultaneously, anterior view. (c–d) Inferior view of (a–b).

Figure 11

(a) Coronal slice from T2 scan from ADNI, and close-up of the hippocampi. (b) Sagittal slice from a T2 scan from ADNI, overlaid on the corresponding T1 volume. This view illustrates the limited field of view of the T2 scans in ADNI. The in-plane resolution of the T2 scans is 0.4 mm, and the slice separation is 2 mm. The T1 scans are 1 mm isotropic.

Figure 12

Inputs (T1, T2) and segmentations for five representative cases of the ADNI T1/T2 dataset. The resolution of the T1 scans is 1 mm isotropic, whereas the T2 scans have 0.4 mm in-plane resolution (coronal) and 2 mm slice separation. (a) A cyst being segmented as hippocampal fissure. (b) A case with good contrast, well-segmented. (c) A case where the partial volume effect has misguided the segmentation, such that part of the lateral ventricle (marked by the arrow) is labeled as CA2/3. (d) A case with poor contrast; the internal segmentation of the hippocampus is largely determined by the prior. (e) A case in which the field of view of the T2 scan does not cover the whole hippocampus; the segmentation of the tail relies solely on the T1 data. All slices are coronal except for (e), which is sagittal. More slices are displayed in the supplementary material (Figure 22).

Figure 13

ROC curves for the AD discrimination task using a LDA classifier on the hippocampal subregion volumes estimated by our ex vivo atlas (FreeSurfer v6.0) and the in vivo atlas (FreeSurfer v5.3 and earlier), as well as for discrimination based on whole hippocampal volume as estimated by FreeSurfer (“aseg”) and by the ex vivo atlas (adding up the volumes of the subregions).

Figure 14

Automated segmentation of the hippocampal subregions of a sample case from the ADNI T1 dataset (T1-weighted, 1mm isotropic) using FreeSurfer automated segmentation (“aseg”), as well as the in vivo and ex vivo atlases. a) Slices of the segmentation. b) 3D renderings of their shape. The color map is the same as in Figure 2 and Figure 5.