Investigation of faulty CO2 chemosensing brain regions in epileptic patients with resting-state fMRI

2039 

Abstract Submission 

Nooshin Jafari Fesharaki¹, Johnson Hampson¹, Samden Lhatoo¹, JungHwan Kim¹

¹University of Texas Health Science Center at Houston, Houston, TX

Nooshin Jafari Fesharaki  
University of Texas Health Science Center at Houston
Houston, TX

Johnson Hampson  
University of Texas Health Science Center at Houston
Houston, TX

Samden Lhatoo  
University of Texas Health Science Center at Houston
Houston, TX

JungHwan Kim  
University of Texas Health Science Center at Houston
Houston, TX

Sudden Unexpected Death in Epilepsy (SUDEP) stands as the most prevalent cause of mortality in patients with epilepsy (PWE), with ~3000 deaths/year in the United States alone. While there is currently no therapeutic intervention for SUDEP prevention, PWE with generalized convulsive seizures (GCS) are presumably at higher risk. GCS events induce significant hypoxia and hypercapnia, causing damage to cardiorespiratory control sites.^1-3 Therefore, one hypothesis is that the sudden death in PWE may result from a disrupted carbon dioxide (CO₂) chemosensing mechanism. Fortunately, it is feasible to identify change in CO₂ chemosensitivity using hypercapnic ventilatory responses (HCVR)⁴ that quantifies an increase in ventilation with rising end-tidal CO₂. Using blood oxygenation level dependent (BOLD) functional Magnetic Resonance Imaging (fMRI), cerebrovascular reactivity (CVR) can be obtained to assess the vascular response to CO₂ changes in cardiorespiratory control nuclei in the epileptic brainstem.⁵ However this method may not be reliable due to physical discomfort of the CO₂ inhalation for subjects. Resting-state fMRI (rs-fMRI) can be an alternative option to yield CVR with regular breathing^6-8 based on spontaneous fluctuations in intravascular pressure of CO₂. Here, we explore CO₂ chemosensory abnormality in PWE with GCS using rs-fMRI CVR.

Structural and 10-min rs-fMRI imaging at 3T were conducted for two groups: 1) 25 health subjects (12F&13M, mean age = 32.8±0.9 years) and 2) five PWEs (4F&1M, mean age = 27±3.4 years) with GCS frequency greater than 10 times per year. The rs-fMRI data underwent correction for head motions and was co-registeration to its structural volume in the standard space. The normalized rs-fMRI time series was then linearly detrended, spatially smoothed (FWHM=6mm), and temporally band-passed filtered into the low-frequency band of 0.01–0.1 Hz. A reference BOLD signal was created by spatially averaging skull-stripped brain-masked time series and used as a regressor in a general linear model to estimate CVR for each voxel.⁷ CVR maps were then masked to cerebellum, brain stem, and subcortical structures (e.g., thalamus, amygdala, caudate, putamen, and hippocampus) and compared between the two groups.

Group comparison shows multiple clusters with significant CVR mean increases (p<0.05) between two groups. The largest clusters were found in the cerebellum (right: lobules VII, VIII, VI, and X, left: lobule IX), caudate nuclei, midbrain (periaqueductal gray), and left thalamus. Further analysis corrected for multiple comparisons using the family-wise-error-correction method showed significant (corrected p <0.001, alpha <0.05, and minimum cluster size: 60) CVR increases in the left caudate nuclei in PWE with GCS (>10) compared to controls, Fig.1. Whole-brain group comparison demonstrated that the CVR mean of patients was bilaterally greater in cingulate gyrus compared to controls, Fig.2 (corrected p <0.001, alpha <0.1, and minimum cluster size: 100).

Increase in rs-fMRI CVR mean was observed in the cingulate cortex and caudate nuclei in PWE with GCS (>10). Consistent with previous animal studies showing increased CO₂-activated cells within these areas,⁹ our findings align with the importance of the human cingulate cortex in respiratory control according to previous electrophysiological analyses.¹⁰ This study demonstrated the identification of faulty CO₂ chemosensing areas with rs-fMRI CVR, offering a less demanding alternative to CO2 inhalation experiments. Future research should explore high-resolution BOLD rs-fMRI for more detailed insights into chemosensitivity areas.

Task-Independent and Resting-State Analysis ¹

Subcortical Structures

BOLD fMRI ²

Cerebrovascular Disease

Cortex

Epilepsy

FUNCTIONAL MRI

Sub-Cortical

Other - cerebrovascular reactivity