Article Figures & Tables
Figures
- Figure 1
Differences in mean rSUV by region of interest. Regions of interest included the left and right sides of (A) the entire anterior cingulate cortex (violet); (B) the ACC pregenual (green), subgenual (red) and supracallosal (blue) subregions; and (C) the insula (yellow). (D) Regions of interest chosen a priori for hypothesis testing are highlighted in red boxes. We ran linear regressions for each region of interest (including covariates), finding significant differences between patients with major depressive disorder (n = 12) and controls (n = 23) in the left subgenual ACC (p = 0.012) when partial volume correction was used. ACC = anterior cingulate cortex; rSUV = relative standardized uptake value.
- Figure 2
The relationship between mean rSUV from the insula and HAM-D scores. A linear regression including age as a nonlinear factor (LOESS) and sex as covariates found that [18F]FEPPA activity (represented by rSUV) in (A) the left insula (p = 0.039) and (B) the right insula (p = 0.037) significantly predicted HAM-D scores in patients with major depressive disorder (n = 12). [18F]FEPPA = [18F]N-2-(fluoroethoxl) benzyl-N-(4phenoxypyridin-3-yl) acetamide; HAM-D = 17-item Hamilton Depression Rating Scale; LOESS = locally estimated scatterplot smoothing; rSUV = relative standardized uptake value.
- Figure 3
Differences in functional connectivity between the sgACC and the insula. A linear regression (including age and sex as covariates) found that functional connectivity (represented by Pearson r-to-z scores) did not differ significantly between patients with major depressive disorder (n = 12) and controls (n = 19). We found a borderline significant difference in functional connectivity between the right sgACC and the right insula (p = 0.053), and between the left sgACC and the right insula (p = 0.059). sgACC = subgenual anterior cingulate cortex.
- Figure 4
The relationship between mean rSUV and functional connectivity. We ran nonlinear regressions that included PCA-derived functional connectivity factor and age as nonlinear factors (LOESS), and linear covariates of no interest, to assess the impact of sgACC–insula functional connectivity (represented by Pearson r-to-z scores) on [18F]FEPPA activity (represented by rSUV). (A) We used all 6 pair-wise connectivity variables between the bilateral sgACC (red) and the insula (yellow) to create a single principal factor. (B) In the full sample (n = 31), increasing functional connectivity was associated with a sinusoidal-like change in [18F]FEPPA activity (p = 0.044). (C) In patients with MDD (n = 12), increasing functional connectivity was associated with first a decrease and then an increase in [18F]FEPPA activity (p = 0.015). (D) In controls (n = 19), we again found a sinusoidal relationship between connectivity and activity (p = 0.035). [18F]FEPPA = [18F]N-2-(fluoroethoxl) benzyl-N-(4phenoxypyridin-3-yl) acetamide; LOESS = locally estimated scatterplot smoothing; MDD = major depressive disorder; PCA = principal component analysis; rSUV = relative standardized uptake value; sgACC = subgenual anterior cingulate cortex.
- Figure 5
The relationship between microglial activity and structural connectivity. We assessed the impact of white matter integrity of the cingulum bundle on microglial activity in patients with major depressive disorder (n = 12). We ran linear regressions that included age as a nonlinear covariate and sex, TSPO status and diffusion tensor imaging metrics of white matter integrity as linear predictors. (A) Increasing axial diffusivity was associated with increasing microglial activity (p = 0.023). (B) Increasing fractional anisotropy was not significantly associated with microglial activity (p = 0.84). (C) Increasing mean diffusivity was not significantly associated with microglial activity (p = 0.07). (D) Increasing radial diffusivity was not significantly associated with microglial activity (p = 0.17). rSUV = relative standardized uptake value; sgACC = subgenual anterior cingulate cortex; TSPO = translocator protein.
Tables
Variable* Depression
n = 12Control
n = 23Significant difference between groups Sex, n 8 female, 4 male 11 female, 12 male No Age, yr 54.9 ± 4.5 60.3 ± 8.5 No Level of education, yr 17.1 ± 4.0 NA NA TSPO status, n 8 HAB, 4 MAB 12 HAB, 11 MAB No Injected dose [18F]FEPPA, MBq 222.7 ± 50.0 204.6 ± 49.9 No Montreal Cognitive Assessment score† 26.5 ± 2.1 NA NA HAM-D score 18.9 ± 7.1 NA NA Suicidality, % 66.7 NA NA Psychosis, % 8.3 NA NA Age of onset, yr‡ 30.4 ± 13.8 NA NA [18F]FEPPA = [18F]N-2-(fluoroethoxl)benzyl-N-(4phenoxypyridin-3-yl) acetamide; HAB = high-affinity binder; HAM-D = 17-item Hamilton Depression Rating Scale; MAB = mixed-affinity binder; NA = not applicable; TSPO = translocator protein.
↵* All means include ± standard deviation.
↵† Missing data from 1 participant.
↵‡ Missing data from 2 participants.
Figure Outcome Difference ± SD (controls − patients with MDD) t (1 − α/2g, n − k) p value Power 1 rSUV, left sgACC −0.165 ± 0.062 −2.684 0.012 0.986 1 rSUV, right sgACC −0.077 ± 0.071 −1.086 0.29 0.934 1 rSUV, left insula −0.091 ± 0.064 −1.409 0.17 0.739 1 rSUV, right insula −0.100 ± 0.069 −1.438 0.16 0.814 3 Connectivity, left sgACC–left insula 0.057 ± 0.078 0.725 0.47 0.407 3 Connectivity, right sgACC–right insula 0.136 ± 0.067 2.024 0.053 0.838 3 Connectivity, left sgACC–right insula 0.139 ± 0.070 1.973 0.059 0.792 3 Connectivity, right sgACC–left insula 0.092 ± 0.072 1.288 0.21 0.796 MDD = major depressive disorder; rSUV = relative standard uptake value; SD = standard deviation; sgACC = subgenual anterior cingulate cortex; TSPO = translocator protein.
↵* Group differences are presented after adjustment for age, sex and TSPO in a regression model. Bonferroni adjustments were applied to p values for simultaneous parameter estimates. α = 0.05, g indicates the number of parameters to be estimated jointly, k indicates the number of variables in the model and n − k represents degrees of freedom.
Figure Outcome Predictor (mean) Change ± SE t p value† Power 2A HAM-D rSUV, left insula (1.37760) −36.675 ± 14.670 −2.500 0.039 0.880 2B HAM-D rSUV, right insula (1.30117) −30.151 ± 11.871 −2.540 0.037 0.840 5A rSUV, left sgACC Axial diffusivity of the left cingulum (0.00091) 3.282 ± 1.042 3.150 0.023 0.683 5B rSUV, left sgACC Fractional anisotropy of the left cingulum (0.30021) −0.154 ± 0.735 −0.210 0.84 0.419 5C rSUV, left sgACC Mean diffusivity of the left cingulum (0.00068) 2.110 ± 0.982 2.150 0.07 0.839 5D rSUV, left sgACC Radial diffusivity of the left cingulum (0.00057) 1.238 ± 0.809 1.530 0.17 0.833 HAM-D = 17-item Hamilton Depression Rating Scale; rSUV = relative standard uptake value; SE = standard error; sgACC = subgenual anterior cingulate cortex.
↵* Represented are changes in microglial activity of the left sgACC associated with mean increase in diffusion tensor imaging scalars of the cingulum bundle, and changes in functional connectivity of the sgACC associated with mean increase in insular microglial activity.
↵† p value associated with the absolute value of t.
Figure Outcome Independent variable Population Nonlinear association F df p value Power 4B Mean rSUV, left sgACC PCA Pearson r-to-z scores All 3.77 1.755 0.044 0.938 4C Mean rSUV, left sgACC PCA Pearson r-to-z scores Patients with MDD 12.30 0.944 0.015 0.744 4D Mean rSUV, left sgACC PCA Pearson r-to-z scores Controls 4.46 2.266 0.035 0.815 df = degrees of freedom; MDD = major depressive disorder; PCA = principal component analysis; rSUV = relative standard uptake value; sgACC = subgenual anterior cingulate cortex.
↵* Nonlinear association between the left sgACC mean rSUV and Pearson’s r-to-z scores (PCA-derived) for each population.
In this issue
Article tools
Related Articles
Cited By...
- No citing articles found.