Introduction

The hippocampus is a critical structure for learning and memory, but it is difficult to image given its small size and complex internal architecture. Most hippocampal studies in typical childhood/adolescent development focus on whole/subfield volume or external shape analysis on 3D T1-weighted images^1,2. Diffusion tensor imaging (DTI) can provide complementary measures on hippocampal microstructure, but very few studies have focused on its development. The whole hippocampus has shown lower MD and higher FA with age over 8-13 years in healthy children³. However, the accuracy of the diffusion parameters is limited and subfield analysis is not possible since DTI is usually acquired with very low spatial resolution (voxel sizes of > 2x2x2 = 8 mm³). A recently published protocol enables the acquisition of high resolution (voxel size = 1 mm³) DTI of the hippocampus in a reasonable 5.5 minutes at 3T that yields excellent visualization of internal structures and segments while minimizing partial volume effects⁴. The purpose of this study was to use high resolution DTI to evaluate age-related changes of the hippocampus head, body and tail in healthy children to young adults.

Methods

High resolution 1 mm isotropic DTI was acquired in 40 healthy volunteers (6-20 years; 19 males) on a Siemens Prisma 3T using a recently published method⁴: scan-time 5:18 minutes for 20 1 mm slices, no gap, single shot 2D EPI (GRAPPA R2; 6/8 PPF; A/P phase encode), FOV 220 x 216 mm², matrix 220 x 216, BW 1420 Hz/px, 1x1x1 mm³ with no interpolation, TE 72 ms, TR 2800 ms, diffusion-time 29 ms, b 500 s/mm² with 10 monopolar gradient directions and 10 averages, and 12 b 0 s/mm², aligned to the long axis of the hippocampus on 3D T1-weighted MPRAGE (0.85x0.85x0.85 mm³). Processing included denoising, Gibbs-ringing, motion, and distortion correction, and tensor parameter estimation (MRtrix3). Regions-of-interest were traced manually on mean DWI axial slices using ITK-SNAP to yield left and right whole-structure volume, FA and MD of the hippocampus. FA and MD were also measured in the head, body and tail of the left hippocampus segmented manually on a single central mean DWI slice (Figure 1)⁵. Paired t-tests revealed left/right differences in whole-hippocampal volume but not diffusion, and unpaired t-tests demonstrated no male/female differences. Therefore, age-related change was examined in males/females together for whole volumes (left/right separated), whole FA or MD (left/right averaged), and in the head, body, and tail for FA or MD (segmented in the left only; right not analyzed yet). Linear, quadratic and exponential models were assessed using Akaike information criterion (SPSS). Tests with p < .05 were considered significant.

Results

Major hippocampal anatomy, including head digitations and the stratum lacunosum moleculare, are well-delineated on mean DWI over the full age range of 6-20 years and visual analysis of the color-coded MD maps suggest age-related diffusion changes (Figure 2). Right whole-hippocampal volumes of 2368 +/- 480 mm³ were greater than left of 2230 +/- 499 mm³ (t = 3.37, p = .002) over the entire group, and linear fits showed a steeper positive slope of whole hippocampal volume versus age in the left (Figure 3). For whole-hippocampal diffusion (left/right averaged), quadratic fits demonstrated lower MD (Figure 4A) and higher FA (Figure 5A) with age from 6-20 years, albeit by small amounts (-4% MD, +9% FA). However, these diffusion changes were not constant across the entire structure. Quadratic relationships were identified only in the head and tail with the steepest age-related MD reduction in the head (Figure 4B) then the tail (Figure 4D), but no age-related MD change in the body (Figure 4C). Similarly, the head had the only significant positive correlation (quadratic) of FA with age (Figure 5B-D).

Discussion and Conclusions

High resolution 1 mm isotropic DTI revealed developmental changes in the healthy hippocampus over 6 to 20 years of age. The whole structure findings replicated those found with low spatial resolution³, but it was shown here that age associations varied regionally with the largest diffusion changes found in the hippocampal head relative to the body and tail. This finding is corroborated by a large sample anatomical MRI study which reported greater expansion with age in the head over the first two decades of life¹. Previous ex vivo work suggests that myelination and neuron density increases occur throughout childhood and into adulthood with a regional-specific pattern that may be related to higher rates of cell maturation and connections in the head^6,7. Overall, high resolution DTI enables the delineation of hippocampal substructures and demonstrates regional specificity of microstructural neurodevelopment in childhood and adolescence.

Acknowledgements

Operating grant was provided by the Canadian Institutes of Health Research. Author KS acknowledges a scholarship award from the Natural Sciences and Engineering Research Council of Canada (Alexander Graham Bell Canada Graduate Scholarship-Doctoral, CGS D).

References

1. Lynch KM, Shi Y, Toga AW, Clark KA. Hippocampal shape maturation in childhood and adolescence. Cereb Cortex. 2018;[Epub ahead of print]:1-15.

2. Gogtay N, Nugent TF, Herman DH, et al. Dynamic mapping of normal human hippocampal development. Hippocampus. 2006;16(8):664-672.

3. Mah A, Geeraert B, Lebel C. Detailing neuroanatomical development in late childhood and early adolescence using NODDI. Leemans A, ed. PLoS One. 2017;12(8):e0182340.

4. Treit S, Steve T, Gross DW, Beaulieu C. High resolution in-vivo diffusion imaging of the human hippocampus. Neuroimage. 2018;182:479-487.

5. Wisse LEM, Daugherty AM, Olsen RK, et al. A harmonized segmentation protocol for hippocampal and parahippocampal subregions: why do we need one and what are the key goals? Hippocampus. 2017;27(1):3-11.

6. Beaujoin J, Palomero-Gallagher N, Boumezbeur F, et al. Post-mortem inference of the human hippocampal connectivity and microstructure using ultra-high field diffusion MRI at 11.7 T. Brain Struct Funct. 2018;223:2157-2179.

7. Ábrahám H, Vincze A, Jewgenow I, et al. Myelination in the human hippocampal formation from midgestation to adulthood. Int J Dev Neurosci. 2010;28(5):401-410.