Normal Olfactory Functional Connectivity Despite Lifelong Absence of Olfactory Experiences

Abstract

Congenital blindness is associated with atypical morphology and functional connectivity within and from visual cortical regions; changes that are hypothesized to originate from a lifelong absence of visual input and could be regarded as a general (re) organization principle of sensory cortices. Challenging this is the fact that individuals with congenital anosmia (lifelong olfactory sensory loss) display little to no morphological changes in the primary olfactory cortex. To determine whether olfactory input from birth is essential to establish and maintain normal functional connectivity in olfactory processing regions, akin to the visual system, we assessed differences in functional connectivity within the olfactory cortex between individuals with congenital anosmia (n = 33) and matched controls (n = 33). Specifically, we assessed differences in connectivity between core olfactory processing regions as well as differences in regional homogeneity and homotopic connectivity within the primary olfactory cortex. In contrast to congenital blindness, none of the analyses indicated atypical connectivity in individuals with congenital anosmia. In fact, post-hoc Bayesian analysis provided support for an absence of group differences. These results suggest that a lifelong absence of olfactory experience has a limited impact on the functional connectivity in the olfactory cortex, a finding that indicates a clear difference between sensory modalities in how sensory cortical regions develop.

Keywords: anosmia; homotopic connectivity; regional homogeneity; resting-state; sensory loss.

Figure 1

Definition of ROI and their mean BOLD signal amplitudes. (A) Core olfactory processing regions: spherical ROIs in bilateral piriform cortex (Pir; center coordinates [−22 0–14] [22 2–12]), OFC (center coordinates [−24 30–10] [28 34–12]), and AI (center coordinates [−35 11 1] [35 17 1]). (B) Primary olfactory subregions that together comprise the primary olfactory ROI (Zhou et al. 2019): anterior olfactory nucleus (AON), olfactory tubercle (TUB), frontal piriform cortex (PirF), and temporal piriform cortex (PirT). (C) Boxplot of mean amplitude of preprocessed, but not denoised, BOLD signals in the six core olfactory processing ROIs from (A) and two nonolfactory reference ROIs where only small susceptibility artifacts is to be expected: left postcentral gyrus (PCen L; center coordinates [−44–27 52]; Supplementary Fig. S1A) and right lateral occipital cortex (LOC R; center coordinates [45–74 4]; Supplementary Fig. S1A). The borders of the boxes indicate the first and third quartile, the whiskers stretch to the furthest data points within 1.5 interquartile range above/below the boxes, the black line indicates the second quartile (median); a.u. = arbitrary units. (D) Boxplot of mean amplitude of preprocessed, but not denoised, BOLD signals in the eight primary olfactory subregions from (B). A marked decrease in signal strength in the AON is visible. The borders of the boxes indicate the first and third quartile, the whiskers stretch to the furthest data points within 1.5 interquartile range above/below the boxes, the black line indicates the second quartile (median); a.u. = arbitrary units.

Figure 2

Correlation matrices for the olfactory network. No significant group differences in connectivity within the olfactory network outlined in Figure 1A at a P < 0.05, uncorrected, statistical threshold. Pir = piriform cortex; L = left and R = right hemisphere. Color bar denotes r-values.

Figure 3

Connectivity results for the olfactory cortex. (A) Primary olfactory ROI, for which the analyses were done, are marked in blue. The yellow dashed boxes mark the part of the cortex that is magnified in (B) and (C). (B) Regional homogeneity in both groups (color bar indicate t-values). No significant group differences at a FWE corrected statistical threshold of P < 0.05; a single voxel indicated lower regional homogeneity in ICA as compared to control at a liberal statistical threshold of P < 0.01, uncorrected for multiple comparisons. (C) Voxel-mirrored homotopic connectivity in both groups (color bar indicate t-values). No significant group differences at a family wise error corrected statistical threshold of P < 0.05 or at the liberal statistical threshold of P < 0.01, uncorrected for multiple comparisons.