Brain MAPS: an automated, accurate and robust brain extraction technique using a template library

Abstract

Whole brain extraction is an important pre-processing step in neuroimage analysis. Manual or semi-automated brain delineations are labour-intensive and thus not desirable in large studies, meaning that automated techniques are preferable. The accuracy and robustness of automated methods are crucial because human expertise may be required to correct any suboptimal results, which can be very time consuming. We compared the accuracy of four automated brain extraction methods: Brain Extraction Tool (BET), Brain Surface Extractor (BSE), Hybrid Watershed Algorithm (HWA) and a Multi-Atlas Propagation and Segmentation (MAPS) technique we have previously developed for hippocampal segmentation. The four methods were applied to extract whole brains from 682 1.5T and 157 3T T(1)-weighted MR baseline images from the Alzheimer's Disease Neuroimaging Initiative database. Semi-automated brain segmentations with manual editing and checking were used as the gold-standard to compare with the results. The median Jaccard index of MAPS was higher than HWA, BET and BSE in 1.5T and 3T scans (p<0.05, all tests), and the 1st to 99th centile range of the Jaccard index of MAPS was smaller than HWA, BET and BSE in 1.5T and 3T scans ( p<0.05, all tests). HWA and MAPS were found to be best at including all brain tissues (median false negative rate ≤0.010% for 1.5T scans and ≤0.019% for 3T scans, both methods). The median Jaccard index of MAPS were similar in both 1.5T and 3T scans, whereas those of BET, BSE and HWA were higher in 1.5T scans than 3T scans (p<0.05, all tests). We found that the diagnostic group had a small effect on the median Jaccard index of all four methods. In conclusion, MAPS had relatively high accuracy and low variability compared to HWA, BET and BSE in MR scans with and without atrophy.

Figure 1

The flowchart of MAPS. Please refer section 2.4.1 for the description of each processing step.

Figure 2

MAPS parameter section: the figure shows the average Jaccard index of ’undilated MAPS-brains’ using different numbers of best-matched atlases and label fusion techniques in a subset of 10 images.

Figure 3

Visual demonstration of MAPS. The sub-figures show the intermediate results of MAPS as described in Section 2.4.1 and Fig. 1.

Figure 4

Examples of whole brain extraction results of MAPS, BET, BSE and HWA of a 1.5T scan (ADNI subject ID: 126 S 0680). While all techniques had some errors in including non-brain (e.g. dura) voxels in some areas – the amount varied between methods (arrows).

Figure 5

Examples of whole brain extraction results of MAPS, BET, BSE and HWA of a 1.5T scan after thresholding using 60% of the mean intensity of the semi-automated whole brain segmentation (ADNI subject ID: 126 S 0680).

Figure 6

Examples of whole brain extraction results of MAPS, BET, BSE and HWA of a 3T scan (ADNI subject ID: 037 S 1225).

Figure 7

Examples of whole brain extraction results of MAPS, BET, BSE and HWA of a 3T scan after thresholding using 60% of the mean intensity of the semi-automated whole brain segmentation (ADNI subject ID: 037 S 1225).

Figure 8

Mean false positive maps of MAPS, BET, BSE and HWA from the segmentations of our whole dataset (682 1.5T and 157 3T scans). The colour maps show the average number of false positive counts (represented by the scales) in each projection plane.

Figure 9

Mean false negative maps of MAPS, BET, BSE and HWA from the segmentations of our whole dataset (682 1.5T and 157 3T scans). The colour maps show the average number of false negative counts (represented by the scales) in each projection plane. Note the differences in scale bar when comparing across these techniques; the scale bar for MAPS and HWA extend only to 0.6 whereas BET and BSE extend to 10.

Figure 10

Errors in a semi-automated segmentation. Extra dura and tentorial tissues were included in the segmentation (pointed by the white arrows).

Figure 11

Bland-Altman plot showing the agreement between brain atrophy measurement (as a percentage of the baseline brain volume) using KN-BSI calculated from semi-automated segmentations in baseline scans and propagated segmentations in 12-month follow-up scans (automated KN-BSI), and from ‘undilated MAPS-brains’ in baseline and 12-month follow-up scans (MAPS KN-BSI).