Structural connectivity of the olfaction network and its relation to aging and olfactory function
Poster No:
2461
Submission Type:
Abstract Submission
Authors:
XIN LI1, Jonas Olofsson2, Jonas Persson3
Institutions:
1Karolinska Institutet, Stockholm, AB, 2Stockholm University, Stockholm, AB, 3Örebro universitet, Örebro, AB
First Author:
Co-Author(s):
Introduction:
Olfactory function declines with age, significantly impacting the quality of everyday life. Emerging research has unveiled a compelling link between olfaction dysfunction and Alzheimer's disease. Olfactory performance has been related to specific brain regions, including the hippocampus (HC) and parahippocampal gyrus (PhG). However, how the primary olfaction cortex plays a role in olfaction is still unknown. The knowledge gap is partly due to the neural mechanism of the olfaction may lie in the connections among multiple brain regions. Taking a data-driven approach, a previous study parcellated the primary olfactory cortex into four distinct subregions and characterized their unique whole-brain functional connectivity profiles, unveiling the olfactory system's complexity. Currently, no study has explored the structural connectivity foundation of the olfaction system and its relationship with aging.
Methods:
This study included 137 healthy adults without dementia (25-85 years), who underwent MRI scanning and olfaction tests, including smell identification and smell threshold (an index of smell sensitivity). Whole-brain Anatomically Constrained Tractography was performed on the preprocessed diffusion-weighted images in MRtrix 3.0. The primary olfaction cortex was parcellated into 4 subregions, anterior olfactory nucleus (AON), olfactory tubercle (TUB), and frontal and temporal piriform cortices (PriF and PriT), using the atlas from Zhou et al., 2019. Structural connectomes were constructed using tractograms and parcellation images. Connectivity strength was quantified using log-transformed fiber bundle capacity (FBC). After a two-step thresholding, 75 connections were included in subsequent analyses.
We first investigated the age effects on the FBC of each connection within the connectome. To investigate the associations between the olfaction network with individual olfaction performance, we performed canonical regression using generalized additive models (GAMs). These models underwent hyperparameter tuning, model training, and testing through nested 2-fold cross-validation, which was repeated 100 times. The resulting feature weights were extracted to discern the significance of each connection to olfaction and cognitive performance.
We first investigated the age effects on the FBC of each connection within the connectome. To investigate the associations between the olfaction network with individual olfaction performance, we performed canonical regression using generalized additive models (GAMs). These models underwent hyperparameter tuning, model training, and testing through nested 2-fold cross-validation, which was repeated 100 times. The resulting feature weights were extracted to discern the significance of each connection to olfaction and cognitive performance.
Results:
All 4 regions of the primary olfaction cortex are connected to several subcortical nuclei, including the HC, caudate (Cd), and putamen. Moreover, the subregions have unique connectivity patterns to the rest of the brain (Fig. 1). Importantly, only a small set of connections within the olfaction network demonstrated the age effects, including connections between AON to HC and PhG, between TUB to the HC, and Cd, and between PriT to insula (Pperm<0.05). We then examined if the connectivity strength (FBC) of the olfaction network (75 connections) can predict individual olfaction and other cognitive performance. Results show that the prediction models were significant for odor identification (mean Pearson's r = 0.25, mean MAE = 1.7 of range 0-12, Pperm < 0.01) and episodic memory (mean Pearson's r = 0.22, mean MAE = 7.1 of range 0-72, Pperm < 0.05), but not significant for other task domains. The weights of each connection in the prediction model of odor identification and episodic memory are shown in Fig. 2.
Conclusions:
The study is the first to reconstruct the brain connectome of the entire olfactory network. Using a wide age range, the study demonstrated age effects on connections between the olfaction cortex and key regions including the HC, Cd, and insula. Through quantifying the connectivity strength of the olfaction network, we were able to partially unveil the structural underpinnings of olfactory function. Notably, our findings demonstrated shared connectivity features that contribute to both odor identification and episodic memory. In summary, the study shed light on the neural mechanisms of olfactory function, providing novel insight into the intricate relationship among aging, olfaction, and episodic memory.
Modeling and Analysis Methods:
Classification and Predictive Modeling
Connectivity (eg. functional, effective, structural) 2
Diffusion MRI Modeling and Analysis
Novel Imaging Acquisition Methods:
Diffusion MRI
Perception, Attention and Motor Behavior:
Chemical Senses: Olfaction, Taste 1
Keywords:
Aging
Multivariate
Smell
White Matter
Other - connectome
1|2Indicates the priority used for review
Provide references using author date format
Smith, R.E., Tournier, J.D., Calamante, F., & Connelly, A. (2012). Anatomically-constrained tractography: improved diffusion MRI streamlines tractography through effective use of anatomical information. Neuroimage, vol. 62, pp. 1924-1938.
Tournier, J.D., Calamante, F., Gadian, D.G., & Connelly, A. (2004). Direct estimation of the fiber orientation density function from diffusion-weighted MRI data using spherical deconvolution. Neuroimage, vol. 23, pp. 1176-1185,
Zhou, G., Lane, G., Cooper, S. L., Kahnt, T., & Zelano, C. (2019). Characterizing functional pathways of the human olfactory system. Elife, vol. 8, article e47177.