Pupillometry as a potential biomarker for predicting vagus nerve stimulation response in refractory epilepsy

Details

Supervisors

El Tahry, Riëm

Faculty

Faculté de pharmacie et des sciences biomédicales

Degree label

Master [120] en sciences biomédicales, à finalité spécialisée: sciences biomédicales cliniques

Abstract

enBackground: Neuromodulation techniques are considered for patients with refractory epilepsy who are not eligible for resective surgery. Vagus nerve stimulation (VNS) therapy consists of an implanted generator delivering electrical pulses to the vagus nerve with different stimulation parameters. Activation of the locus coeruleus (LC) in the brainstem is thought to mediate its therapeutic effect. However, approximately 30% of implanted patients do not respond to VNS. The pupil dilation response (PDR), an indicator of LC activation, shows variations with VNS parameters in its late component, while its early component may reflect the subjective perception of stimulation. Considering that PDR could serve as a potential marker of VNS activation, the main objective of this study was to evaluate the effect of different VNS frequencies on pupil size in responders (R) and non-responders (NR). Methods: Ten VNS-implanted patients (4 R and 6 NR) from Saint-Luc University Hospital were included. Four VNS frequencies (1, 10, 20, 30 Hz) were tested. Pupil dilation was recorded using an infrared camera, and perception scores were collected after each stimulation. We used linear mixed models to analyze changes in pupil size and perception scores, with successive post-stimulation time windows included as a fixed effect in the pupil model. Post hoc comparisons were performed using Tukey’s method. Results: PDR varied across time windows (p = 0.036) and was maximal 2–3 s after stimulation onset. The effect of stimulation frequency depended on response status (p < 0.001). NR showed a decrease in pupil size with increasing frequency. Specifically, compared to 1 Hz, NR exhibited the lowest PDR at 30 Hz (difference = -0.06 mm, p = 0.002), whereas R showed a peak at 10 Hz (difference = 0.06 mm, p = 0.002). Perception scores were influenced by frequency (p = 0.002), with no effect of response status. These findings indicate that pupil size varies according to VNS frequency and clinical response status. Subjective perception did not account for the observed group differences in PDR profiles. Conclusion: This study suggests that R and NR exhibit differential levels of LC engagement, supporting the role of PDR as a marker of VNS in patients with epilepsy. By reflecting LC activation, PDR could serve as a biomarker predicting VNS response and guiding VNS parameter optimization. However, in our study, the frequencies commonly used in clinical practice (20 – 30 Hz) did not elicit the strongest PDR. This discrepancy could be explained by different temporal dynamics of LC activation depending on stimulation frequency. Further studies should test combinations of various parameters to help identify optimal settings.