Layer-specific fMRI of the human hippocampus in autobiographical memory

2323 

Abstract Submission 

Viktor Pfaffenrot¹, Antoine Bouyeure², Nikolai Axmacher², David Norris^3,1

¹Erwin L. Hahn Institute for Magnetic Resonance Imaging, University of Duisburg-Essen, Essen, North-Rhine Westphalia, Germany, ²Department of Neuropsychology, Ruhr-Universität Bochum, Bochum, North-Rhine Westphalia, Germany, ³Donders Institute for Brain, Cognition and Behaviour, Radboud University, Nijmegen, Gelderland, Netherlands

Viktor Pfaffenrot  
Erwin L. Hahn Institute for Magnetic Resonance Imaging, University of Duisburg-Essen
Essen, North-Rhine Westphalia, Germany

Antoine Bouyeure  
Department of Neuropsychology, Ruhr-Universität Bochum
Bochum, North-Rhine Westphalia, Germany

Nikolai Axmacher  
Department of Neuropsychology, Ruhr-Universität Bochum
Bochum, North-Rhine Westphalia, Germany

David Norris  
Donders Institute for Brain, Cognition and Behaviour, Radboud University|Erwin L. Hahn Institute for Magnetic Resonance Imaging, University of Duisburg-Essen
Nijmegen, Gelderland, Netherlands|Essen, North-Rhine Westphalia, Germany

Recent advances in ultra-high field (UHF) fMRI have helped to describe the distinct functional roles of hippocampal subfields in memory (e.g., [1]). Anatomical descriptions of hippocampal circuits have shown that hippocampal inputs and outputs are segregated into different laminae [2]. However, fMRI studies on the layer-specific organization of hippocampal subfields are still largely lacking (but see [3]). Indeed, the human hippocampus is a difficult target to study due to its complex folded structure, the susceptibility artifacts that affect MRI signals in this region, and the need to account for its vascular architecture to reliably interpret fMRI responses. In this work, we sought to robustly acquire laminar fMRI responses from hippocampal subfields at 7T during an autobiographical memory task and to assess subject-specific differences in these responses.

Two male healthy subjects were scanned on a 7T Siemens MAGNETOM-Terra MRI. Functional data were acquired using a GRE-BOLD 3D-EPI sequence (0.9 mm isotropic resolution). The same sequence with reversed-phase encoding was acquired for distortion correction. The functional task was an autobiographical memory (AM) paradigm adapted from [4, 1]). Three runs were acquired for each session and two sessions were acquired for each participant on two different days. In each run, subjects were asked to recall autobiographical episodic memories following the presentation of generic cues (e.g., 'party', 'pet', etc.). Once the subjects had found a memory related to the cue, they were tasked to imagine it with as much detail as possible during the remainder of the trial. Subjects were instructed to retrieve recent memories (no older than two years) in order to ensure hippocampal activation. The AM task was randomly interleaved with an mental arithmetic (MA) task where the subjects should solve a simple arithmetic operation. AM and MA trials each lasted for 17.6 seconds.
For anatomical reference, a T1-weighted MP2RAGE [5, 6] was acquired (0.75 mm isotropic resolution). Furthermore, a high in-plane resolution SWI image was acquired (0.3x0.3x0.6 mm resolution) as well as a TOF-MRA to investigate how hippocampal vasculature would influence the laminar profiles.
Hippocampus surfaces and subfields were extracted using HippUnfold [6]. Functional data were processed using an in-house developed pipeline based on ANTS [7] and SPM. GM signals were equidistantly sampled on 20 depth bins between the inner and outer surface boundary using a custom-written MATLAB script. In case of cornu ammonis 1 (CA1) and CA2, the sampling was extended by 10 depth bins into the stratum radiatum, lacunosum and moleculare (SRLM). Data was corrected for physiological noise using aCompCor [8].

Higher activation of the subiculum relatively to other subfields during the autobiographical memory condition compared to the mental arithmetic condition was found in both subjects, consistent with the idea that the subiculum plays an important role in memory retrieval (see [1, 9]). At the laminar level, layer profiles corrected for physiological noise with aCompCor showed high within-subject reproducibility between sessions (Fig. 1). Profiles from both subjects showed a stronger bias toward the inner surface of the hippocampus, with some notable differences between subjects that could be attributed to differences in venous vascularization (Fig. 2).

We have shown that robust and reproducible laminar profiles can be obtained for hippocampal subfields at the subject level. Inter-subject differences in vascular anatomy are possible factors influencing layer profiles. Future directions will include consideration of the functional organization of the hippocampus on its long axis along with subfield- and layer-specific analyses in the description of hippocampal function.

Long-Term Memory (Episodic and Semantic)

Activation (eg. BOLD task-fMRI)

Methods Development

Other Methods ²

BOLD fMRI ¹

Cortical Layers

FUNCTIONAL MRI

Memory

Other - Hippocampus