Table 1

Hippocampal volumes and subfield papers: 7T MRI in vivo studies including subjects with neurodegenerative dementia

PaperImaging methodsParticipantsQuestionResults
Boutet et al 67 Whole brain: MP-RAGE acquisition; 176 slices; voxel size 0.9×0.9×0.9 mm
Hippocampus: 2D T2* GRE acquisition; 3×15 slices; voxel size 0.3×0.3×1.2 mm
Manual segmentation based on anatomical landmarks
NC=7
‘mild to moderate’ AD*=4
Can volume changes in distinct hippocampal layers be detected in vivo using 7T MRI (using imaging protocol distinguishing between layers richer/poorer in neuronal bodies)?Bilateral reduction of CA-SRLM and SUB-SP volumes (sig) in AD (Average difference −29% to −49%).
Trend towards reduction of left CA-SP in AD (non-sig).
Kerchner et al 35 Whole brain: T1-weighted 3T MRI
MTL: 2D T2*-weighted GRE acquisition; 100 slices in 15 subjects and 40 slices in 15 subjects; slice gap 0 mm in 15 subjects and 3 mm in 15 subjects; voxel size 0.195×0.195×2 mm; scan time 9.6 min
Manual segmentation based on anatomical landmarks
NC=16
‘Mild’ AD=14
Is tissue loss in hippocampal CA1 apical neuropil layer seen on in vivo MRI in mild AD?CA1 cell–body layer thickness and entire CA1 subfield thickness not sig diff between groups.
CA1-SRLM width decreased in AD (sig).
CA1 apical neuropil thickness decreased in AD (sig).
Left lateralised changes in AD
THV not sig diff between groups (measured at 3T).
Kerchner et al 25 MTL: T2-weighted FSE acquisition; 16–18 slices; voxel size 0.166×0.166×1.5 mm (interpolated reconstruction from 0.22×0.22×1.5 mm acquired voxels); scan time around 10 min
Manual and semiautomated segmentation based on anatomical landmarks
Mild AD=9Do hippocampal subfield widths correlate with episodic memory task performance?Delayed recall performance (DRP) correlated with CA-SRLM width r2=0.69, with CA1-SP width r2=0.5, with ERC width r2=0.62 (all sig).
Differences between groups left lateralised.
Kerchner et al 26 As per Kerchner et al 25 YNC=9
ONC=18
MCI§=15
AD=11
Do ERC and CA1-SRLM share early vulnerability to AD pathology? (Is atrophy proportional between the two structures?)Ratio of CA1-SRLM to ERC width 0.25±0.03 across all subjects.
Width of CA1-SRLM and ERC decreased as function of age and cognitive impairment.
CA1-SRLM width and episodic memory performance correlated in AD/ aMCI; not in controls.
ERC width and episodic memory performance correlated with AD (sig). Non-sig trend in aMCI.
ERC and CA1-SRLM widths correlated, and ERC and CA1-SP widths correlated (sig when THVs controlled for).
ERC and CA1-SRLM share vulnerability to age and AD-associated atrophy.
Kerchner et al 27 As per Kerchner et al 25 NC=14
MCI§=14
AD*=11
Is there a relationship between APOE e4, hippocampal subfield morphology and episodic memory?APOE e4 load associated with thinner CA1-SRLM (sig) and poorer episodic memory task performance (association not seen with other areas of cognitive examination). Association persisted after controlling for dementia severity.
APOE e4 not associated with changes in other hippo subfields or ERC.
Wisse et al 28 MTL: 3D T2-weighted TSE (whole brain) acquisition; voxel size 0.35×0.35×0.7 mm; scan time 10:15 min
Manual in-house segmentation approach
NC=29
MCI=16
AD*=9
Do hippo subfield and ERC volumes differ between NC, MCI and AD groups? Do hippo subfield and ERC volumes correlate with age in NCs?ERC, SUB, CA1, CA3, DG&CA4, THV volumes smaller in AD than NC (sig).
ERC, SUB, CA1, DG&CA4, THV volumes smaller in AD than MCI (sig).
MMSE correlated with ERC, SUB, CA1, DG&CA4 volumes (sig).
CA1, DG&CA4, THV volume loss correlated with increasing age (sig).
Wisse et al 43 Hippocampal formation: 7T 3D T2-weighted TSE (whole brain) acquisition; voxel size 0.35×0.35×0.7 mm; scan time 10:15 min
ICV: 3T T2-weighted fast field echo; voxel size 3×0.99×0.99 mm; scan time 2:48 min
Diffusion data: 3T single-shot spin echo planar imaging sequence (SENSE) (twice-refocused); voxel size 2.5×2.5×2.5 mm; scan time 5:32 min
Manual in-house segmentation approach
NC=17
MCI=15
mild AD*=10
Does ERC and hippocampal degeneration induce degeneration of associated WM tracts (fornix and parahippocampal cingulum)?Fornix FA lower in AD (sig).
CA1, DG&CA4, SUB volumes lower in MCI/ AD (sig).
Fornix FA associated with SUB volume β=0.53 (sig) in MCI/AD. Fornix FA not associated with ERC or other hippo subfield volumes in MCI/AD.
PHC FA not associated with ERC or other hippo subfield volumes in MCI/AD.
Hippo subfield atrophy not associated with reduction in whole brain WM FA.
  • AD diagnostic criteria used: *2, †criteria not specified, ‡4 .

  • MCI diagnostic criteria used: §2324 .

  • 3D, three dimensional; AD, Alzheimer’s disease; B0, static field strength; C[11] -PiB, Carbon 11 Pittsburgh B compound (binds to amyloid); CA, cornu ammonis; CSF, cerebrospinal fluid; DG, dentate gyrus; EOAD, early-onset AD; ERC, entorhinal cortex; FA, fractional anisotropy; FSE/TSE, fast/turbo spin echo (these terms are synonymous); GM, grey matter; GRE, gradient echo sequence; hippo, hippocampus; ICV, intracranial volume; LOAD, late-onset AD; MCI, mild cognitive impairment; MMSE, Mini-Mental State Examination; MP- RAGE, magnetisation-prepared rapid acquisition gradient echo; MTL, medial temporal lobe; NC, cognitively normal controls; ONC, older cognitively normal controls; PHC, parahippocampal cingulum; QSM, quantitative susceptibility mapping; rs-fMRI, resting state functional MRI; RF, radio frequency; ROI, region of interest; sig, significant (p<0.05); SP, stratum pyramidale; SPACE, sampling perfection with application optimised contrasts by using different flip angle evolutions; SRLM, strata radiatum, lacunosum and moleculare; SUB, subiculum; SWI, susceptibility-weighted imaging; TE, echo time; THV, total hippocampal volume; TMP, temporoparietal region; WM, white matter; YOC, younger cognitively normal controls.