Facial Emotion Processing Changes in Functional Seizures after Neurobehavioral Therapy
Jerzy Szaflarski1, Jane Allendorfer1, Adam Goodman1, Rodolphe Nenert1, Tyler Gaston1, Leslie Perry Grayson1, Noah Philip2, Stephen Correia2, W. LaFrance2
1University of Alabama at Birmingham, 2Brown University
Objective:

To prospectively investigate effects of neurobehavioral therapy (NBT) on fMRI emotion processing in adults with psychogenic non-epileptic (functional) seizures (FS).

Background:

NBT reduces FS frequency and affects emotion processing in people with FS. We hypothesized that changes in FS frequency, mood symptoms and overall functioning with NBT would be associated with changes in emotion processing.

Design/Methods:
Adults with FS after traumatic brain injury (TBI+FS; n=47) underwent a 12-sessions of NBT. TBI+FS and TBI-only controls (n=57) completed two fMRIs and assessments of depression (BDI-II), anxiety (BAI), and Global Assessment of Functioning (GAF) about 12 weeks apart (V1; V2). Subjects performed the emotion faces task (EFT) during fMRI where they viewed happy, sad, fearful and neutral faces. Linear mixed effects assessed group-by-time interactions on emotion response (corrected p<0.05), fMRI signal was extracted, and correlations performed (Bonferroni-corrected p<0.0125).
Results:

53% of TBI+FS were seizure-free following NBT. There were significant group-by-time interactions (V1-to-V2 increases in TBI+FS and decreases in TBI-only) for fMRI response within right middle frontal gyrus (RMFG) to happy, sad, and fearful faces, and right inferior frontal gyrus (RIFG) to fearful faces. In combined sample, there were negative correlations between V2-V1 BDI-II changes and RMFG response changes to happy (rho=-0.30, p=0.003) and sad faces (rho=-0.27, p=0.008), and positive correlations between V2-V1 GAF changes and RMFG response changes to happy (rho=0.34, p<0.001) and sad faces (rho=0.31, p=0.001), and RIFG response changes to fearful faces (rho=0.26, p=0.008). There were non-significant trends in TBI+FS for relationship between percent FS reduction and RMFG response changes to happy (rho=-0.24, p=0.106) and sad faces (rho=-0.24, p=0.109). 

Conclusions:

We observed strong relationships in the combined group between behavioral changes and fMRI changes. The observed seizure reduction was not correlated with imaging changes suggesting that factors other than seizure reduction may be driving post-NBT changes in facial emotion processing.

10.1212/WNL.0000000000203236