The fasciola cinereum of the hippocampal tail as an interventional target in epilepsy

Epilepsy affects tens of millions worldwide, and approximately one-third of patients remain drug-resistant. For mesial temporal lobe epilepsy (TLE), standard surgical approaches ablate the anterior hippocampus and amygdala, yet roughly one-third of surgically treated patients do not achieve seizure freedom. This persistent seizure burden raises the hypothesis that seizure foci and critical propagation nodes may reside in posterior hippocampal structures that are typically spared, such as the fasciola cinereum (FC) in the posterior-medial hippocampal tail. The study aims to identify whether the FC is actively engaged in seizure activity and assess its potential as an interventional target in both mouse models of TLE and human patients.

Prior clinical literature establishes mesial temporal structures (anterior hippocampus and amygdala) as conventional surgical targets for drug-resistant TLE, via open resection or laser interstitial thermal therapy (LITT). Randomized and cohort studies show meaningful but incomplete seizure control after such procedures, with about one-third of patients not achieving adequate seizure freedom. Anatomical and histological works characterize the FC as a distinct hippocampal subregion: in rodents, it lies medially along the dorsal hippocampus, while in primates/humans it is located posteriorly in the hippocampal tail. Existing stereoelectroencephalography (sEEG) strategies rarely sample this posterior-medial tail region, potentially overlooking a relevant node. This context motivated targeted investigation of FC involvement in seizure initiation/propagation and its value as an interventional target.

Preclinical (mouse): - Seizure-activity labeling with scFLARE: Wild-type mice received AAVs encoding scFLARE and a reporter (mCherry or eGFP) in the hippocampus. Acute seizures were induced with kainic acid (intrahippocampal or intra-amygdalar). A closed-loop system detected seizures via hippocampal LFP and delivered blue light to activate scFLARE, permanently labeling neurons active during seizures. Control non-seizing mice received saline. scFLARE-labeled cells were examined histologically; PCP4 expression identified FC neurons. - Two-photon calcium imaging: In chronically epileptic PCP4-Cre mice, Cre-dependent jGCaMP8f was expressed in FC neurons. A GRIN lens enabled two-photon imaging. Simultaneous hippocampal LFP recorded interictal and ictal events. Calcium dynamics of individual FC neurons (n=80 cells across 3 mice) were aligned to interictal spikes, seizure onset, and seizure end. - Closed-loop optogenetic inhibition: In chronically epileptic PCP4-Cre mice, inhibitory opsin stGtACR2 (or control mCherry) was expressed in FC neurons. An optical fiber terminated just superior to FC; hippocampal LFP enabled real-time detection of seizure onset to trigger blue-light inhibition. Seizure durations with light-on vs light-off were quantified. Group sizes: stGtACR2 n=4 mice (2,075 seizures analyzed), control mCherry n=4 mice (1,627 seizures analyzed). Mixed-effects models and t-tests compared seizure durations. Clinical (human): - Patient cohort and sEEG: Six patients undergoing sEEG for seizure localization had targeted electrodes placed to sample the posterior-medial hippocampal tail containing the FC, with broader sampling across suspected networks (temporal, frontal, insular, occipital, thalamic). sEEG analyses identified epileptiform discharges, ictal spiking, and high-frequency oscillations (80–250 Hz). Patterns of FC involvement relative to anterior hippocampus were documented. - Therapeutic ablation case (Patient 6): Following prior LITT amygdalohippocampectomy with residual posterior hippocampal tail, sEEG localized seizures exclusively to the FC remnant (117/117 seizures). A repeat LITT targeted the residual FC using MRI-guided thermometry with safety thresholds to protect nearby thalamic structures. Outcomes were assessed at 18 months post-ablation. Statistics and validations: - Imaging and optogenetics included replication across animals; expression and targeting were histologically verified. Mixed-effects models compared seizure durations (light on vs off). Additional per-mouse seizure duration distributions were assessed with Mann–Whitney tests. Correlations between calcium activity and spiking were computed (mean r ± s.e.m.). sEEG interpretations were performed blinded to channel identities when assessing relative onsets separated by ≥50 ms.

- Mouse seizure-activity labeling: scFLARE robustly labeled PCP4+ neurons in the fasciola cinereum (FC) during seizures in both acute and chronic TLE models, while negligible labeling occurred in non-seizing controls. - FC neuronal dynamics: Two-photon calcium imaging in chronically epileptic mice showed FC neuronal activity increases aligned with interictal spikes and at seizure onset, and activity decreases at seizure end. Correlation between Ca2+ activity and spiking across cells: r = 0.43 ± 0.03 (s.e.m.). Recorded 80 FC neurons across 3 mice. - Optogenetic inhibition reduces seizure duration: Closed-loop light delivery in PCP4-Cre mice expressing stGtACR2 in FC significantly shortened seizures compared to light-off. Mixed-effect model: F(1, 1,010) = 51.47, P < 0.0001. Normalized seizure duration (light on vs off): 76 ± 3% for stGtACR2 (2,075 seizures from 4 mice) vs 98 ± 4% for mCherry controls (1,627 seizures from 4 mice), two-tailed t-test P = 0.0038. Controls showed no significant change (F(1, 778) = 0.1133, P = 0.74). Individual-mouse analyses corroborated reduced durations with light in stGtACR2 mice (Extended Data Fig. 5; all P ≤ 0.0001 except one P < 0.01), but not in controls (all P ≥ 0.12). - Human sEEG evidence of FC involvement: In six patients with epilepsy, the posterior-medial hippocampal tail containing the FC showed epileptiform discharges, ictal spiking, and high-frequency oscillations (80–250 Hz), consistent with participation in seizure networks. • Patient 1: FC involved in 7/7 seizures (100%). • Patient 2: FC involved in 15/18 seizures (83%); rapid spread from anterior hippocampus to FC; 3 seizures with AH involved without FC. • Patient 3: FC involved in 5/5 seizures (100%). • Patient 4: Neocortical STG onset with spread to right AH and right FC in 3/3 seizures (100%). • Patient 5: Occipital focal cortical dysplasia onset with rapid spread to left FC without spread to left anterior hippocampus in 15/15 seizures (100%). • Patient 6: Seizure onset exclusively in left FC remnant in 117/117 seizures (100%) after prior anterior mesial temporal LITT. - Therapeutic targeting in a human case: Repeat LITT lesioning of the residual FC in Patient 6 produced an 83% reduction in seizure frequency at 18 months (from ~2/month to ~1 every 3 months), despite prior successful ablation of anterior mesial temporal structures, demonstrating FC as a viable interventional target.

The study identifies the fasciola cinereum (FC) as a key node in seizure initiation and propagation across species. In mice, FC neurons are preferentially recruited during epileptiform events, and their closed-loop inhibition shortens seizure durations, indicating a mechanistic role in seizure dynamics rather than passive involvement. In humans, targeted sEEG sampling of the posterior-medial hippocampal tail reveals frequent FC participation, with diverse patterns including independent FC involvement, rapid spread from the anterior hippocampus to FC, and propagation from neocortical foci into FC. The consistent FC engagement, including strong independence from the anterior hippocampus in select cases, suggests that conventional anterior-focused approaches may leave a clinically relevant residual source in the posterior tail. A patient with seizure recurrence after standard anterior mesial temporal LITT showed exclusive FC onset and benefitted from subsequent FC-focused ablation, supporting the translational potential of FC targeting. These findings advocate for routine inclusion of the FC in sEEG evaluations for suspected TLE to better define the required extent of lesioning. For bitemporal epilepsy, FC-specific activity may guide responsive neurostimulation strategies, which currently do not cover posterior hippocampal tail regions. Overall, incorporating FC into diagnostic and therapeutic strategies could optimize outcomes for patients where conventional anterior mesial targeting is insufficient.

This work establishes the fasciola cinereum of the posterior hippocampal tail as an active and intervenable component of seizure networks in both mouse models and human epilepsy. Preclinically, FC neurons are highly active during seizures, and their targeted optogenetic inhibition reduces seizure duration. Clinically, FC participation was documented in all recorded patients, and targeted FC ablation in a case of post-anterior LITT seizure recurrence substantially reduced seizure burden. The main contributions are (1) identification of FC as a seizure-relevant node, (2) demonstration that FC inhibition can modulate seizures in vivo, and (3) translation to human targeting via sEEG and LITT. Future directions include prospective clinical studies systematically sampling and, when appropriate, lesioning the FC; development of neuromodulation strategies explicitly targeting the posterior hippocampal tail; and optimization of imaging and planning to safely access the FC while sparing neighboring thalamic structures.

- Human cohort size is small (n=6) and heterogeneous, reflecting complex clinical indications; findings need validation in larger, prospective studies. - Selection bias: patients with clear unilateral mesial TLE often proceed directly to ablation without sEEG and were underrepresented; results may not generalize to straightforward unilateral cases. - Interventional evidence in humans relies on a single FC LITT case with partial residual tissue post-ablation; seizure reduction, while substantial, was not complete. - Ethical and practical constraints currently preclude FC-only lesioning in first-line therapy given that many patients respond to conventional anterior mesial targeting. - Technical constraints: present RNS systems have limited leads and typical trajectories miss the posterior hippocampal tail; surgical access requires careful trajectory planning to avoid critical structures (e.g., thalamus). - Preclinical experiments, though replicated, involve limited animal numbers; generalizability across models and species requires further study.