Autonomic Network Signatures of Interictal Cardiorespiratory Dysfunction in Patients With Temporal Lobe Epilepsy

Figures

• 

  Figure 1. Abnormal Cardiorespiratory Function, Autonomic Network Connectivity, and Structural Volume in TLE (A) Comparison of the breathing rate (BR), respiratory rate variability (RRV), and high-frequency and low-frequency heart rate variability (HF-HRV/LF-HRV) between temporal lobe epilepsy (TLE) and healthy controls (HC). Two-sample t tests (2-tailed) were used for statistical comparisons. *p < 0.05, **p < 0.01, ***p < 0.001, false discovery rate (FDR)-corrected across 4 cardiorespiratory measures. (B) Regional differences in functional connectivity (FC) node strength of the central autonomic network between TLE and controls (p < 0.05, FDR-corrected across regions; 2-sample t tests, 2-tailed). (C) Regional differences in structural volume between TLE and controls (p < 0.05, FDR-corrected across regions; 2-sample t tests, 2-tailed). The FC, cardiorespiratory, and brain volume estimates were adjusted for age and sex. The brain volume estimates were not adjusted for total intracranial volume.

• 

  Figure 2. Reduced Functional Connectivity Strength Within the Central Autonomic Network Mediates Abnormal Interictal Cardiorespiratory Function in TLE (A) Relationship between the functional connectivity (FC) strength of a node in the left cingulate gyrus with breathing rate (BR), respiratory rate variability (RRV), and low-frequency heart rate variability (LF-HRV) in temporal lobe epilepsy (TLE) compared with healthy controls (HC). (B) Regional mediation effects of the FC node strength of the central autonomic network on differences in BR, RRV, and LF-HRV between TLE and controls (p < 0.05, false discovery rate [FDR]–corrected across regions; bootstrapping tests). The absolute values of the bootstrap z-scores are shown for ease of comparison. (C) Regional associations between the FC node strength of the central autonomic network with BR, RRV, and LF-HRV in TLE (p < 0.05, FDR-corrected across regions; 2-tailed Spearman correlation tests). The FC and cardiorespiratory estimates were adjusted for age and sex.

• 

  Figure 3. Multivariate Associations Between Clinical Disease Parameters, Interictal Cardiorespiratory Function, and the Central Autonomic Network in TLE (A) Loading weights of the canonical correlation analysis (CCA) mode that maximized the correlation between the functional connectivity (FC) strength of the central autonomic network and the combined set of cardiorespiratory measures and clinical parameters in temporal lobe epilepsy (r = 0.63; p = 0.015, familywise error rate [FWER]–corrected across 4 modes; permutation test). Contribution of the FC node strength of the central autonomic regions to the CCA mode (|z| > 1.96; bootstrap z-scores; left brain plot). Contribution of the cardiorespiratory measures and clinical parameters to the CCA mode (correlation loadings; right bar plot). (B) Loading weights of the CCA mode that maximized the correlation between the structural volume of the central autonomic network and the clinical parameters in temporal lobe epilepsy (r = 0.61; p = 0.018, FWER-corrected across 5 modes; permutation test). Contribution of the volume of the central autonomic regions to the CCA mode (|z| > 1.96; bootstrap z-scores; left brain plot). Contribution of the clinical…

• 

  Figure 4. Postsurgical Changes in the Functional Connectivity Strength of the Central Autonomic Network Impact Cardiorespiratory Function in TLE Relationship of presurgical-to-postsurgical alterations in the mean functional connectivity (FC) strength of the central autonomic regions with presurgical-to-postsurgical changes in breathing rate (BR), respiratory rate variability (RRV), and low-frequency heart rate variability (LF-HRV). The FC and cardiorespiratory measures were adjusted for age and sex. Partial Spearman correlations (2-tailed) were used to test for a significant association between the FC and the cardiorespiratory measures while controlling for type of surgery (anterior temporal lobectomy or amygdalohippocampectomy) and time between surgery and postsurgical scan. Bootstrapping (n = 1 × 105) was used to derive 95% CIs for the correlation values, and p-values were false discovery rate (FDR)–corrected across 3 cardiorespiratory measures. TLE = temporal lobe epilepsy.