Interictal waking and sleep electrophysiological properties of the thalamus in focal epilepsies

Abstract

Epilepsy is a cortico-subcortical network disease. Thalamo-cortical relationships in focal epilepsies, studied by stereoelectroencephalography in complex patients during pre-surgical evaluation, might help refine epilepsy surgery prognostic indicators and patient-specific treatments (i.e. thalamic deep brain stimulation). To this aim, we studied interictal thalamic traces, during rest and sleep recordings, in a cohort of 121 patients, delving into thalamo-cortical connectivity, hyperexcitability biomarkers and their correlation with treatment outcome. We retrospectively gathered stereoelectroencephalography recordings and clinical variables from patients who underwent stereoelectroencephalography with mainly a posterior-thalamic implantation, aiming at the pulvinar. Interictal recordings during rest and sleep were analysed to detect spikes and fast ripples automatically. Functional connectivity between the thalamus and other brain regions (involved or non-involved in the epileptogenic network) was examined using linear regression analysis. Higher thalamic hyperexcitability biomarker rates during sleep were linked to unfavourable surgical outcomes (Engel Class III/IV) compared to favourable outcomes (Engel Class I/II) (spikes: N = 117, P = 0.009, effect size = 0.25; fast ripples: N = 17, P = 0.036, effect size = 0.52). Thalamo-cortical functional connectivity analysis revealed heightened thalamic strength, particularly in the beta (P < 0.001, effect size = 0.38) and gamma (P = 0.012, effect size = 0.24) bands during sleep, among patients with poor surgical outcomes, especially with non-involved networks. Conversely, during rest, lower hyperexcitability biomarkers (spikes r = -0.2, P = 0.048; fast ripples r = -0.52, P = 0.045) and lower values of thalamic strength (delta band r = -0.28, P = 0.025; broadband r = -0.23, P = 0.01) were observed in patients with longer epilepsy duration. Furthermore, thalamic strength values during rest were lower in patients of older age (broadband r = -0.19, P = 0.045). These findings confirmed the important role of the thalamus in focal epilepsy. According to this exploratory group-level study, thalamic recordings could potentially improve pre-surgical assessment and help identify patients who may have a less severe outcome. Additionally, diminished thalamic activity and connectivity associated with epilepsy duration and age prompt speculation on the role of thalamo-cortical interactions in ageing-related physiological and pathological processes.

Keywords: epileptogenic networks; intracranial EEG; surgical outcome; thalamic hyperexcitability; thalamo-cortical functional connectivity.

Graphical Abstract

Figure 1

Thalamic implantation and signal recordings. (A) MRI and CT co-registration showing a SEEG electrode (dotted line) whose distal contacts (green and blue dots) are implanted at the level of the left pulvinar. (B) A SEEG signal was recorded at the level of the two distal contacts of the electrode implanted in the pulvinar, in a sleep recording, recognizable for slow wave and spindle activity. The presence of interictal epileptiform activity, a spike, is observed (arrow). (C) Raw signal, whitened signal and time-frequency visualization of a Fast Ripple in the pulvinar (green circle). CT, computed tomography.

Figure 2

Thalamic interictal biomarkers associated with hyperexcitability and clinical characteristics. (A) Association between surgical outcome and spikes [N = 117, mean estimate = −0.032, 95% CI (−0.065; −0.006), Wilcoxon test: P = 0.009, effect size = 0.25] and (B) HFO [N = 17, mean estimate = −0.075, 95% CI (−0.200; −0.006), Wilcoxon test: P = 0.036, effect size = 0.52] in sleep recordings. (C) HFO rate according to Engel score [N = 17, r = 0.44, 95% CI (−0.07; 0.77), P = 0.075]. (D) Association between epilepsy duration and spikes [N = 121, r = −0.20, 95% CI (−0.37; −0.022), P = 0.026) and (E) HFOs (N = 15, r = −0.52, 95% CI (−0.83; 0.02), P = 0.045] in rest recordings.

Figure 3

Thalamo-cortical interictal FC and clinical characteristics. (A) Association between surgical outcome and FC in beta (N = 115, mean estimate = −0.007, 95% CI (−0.011; −0.004), Wilcoxon test: P < 0.001, effect size = 0.38) and (B) gamma band [N = 116, mean estimate = −0.001, 95% CI (−0.002; −0.0001), Wilcoxon test: P = 0.012, effect size = 0.24] in sleep recordings. (C) FC in beta band varies according to Engel score [N = 115, r = 0.267, 95% CI (0.073; 0.441), P = 0.007]. (D) Association between epilepsy duration and thalamo-cortical FC in delta band [N = 120, r = −0.208, 95% CI (−0.377; −0.024), P = 0.025] and (E) broadband [N = 120, r = −0.234, 95% CI (−0.401; −0.052), P = 0.01] in rest recordings. (F) Association between age at SEEG and thalamo-cortical FC in broadband [N = 120, r = −0.186, 95% CI (−0.357; −0.003), P = 0.045] in rest recordings.