Advances in Thalamic Neuromodulation for Epilepsy: From Mechanisms to Clinical Translation

Abstract

Thalamic neuromodulation has re-emerged as a central strategy for the treatment of drug-resistant epilepsy, driven by advances in stereo-electroencephalography, network neuroscience, and signal processing. This perspective paper summarizes and integrates the key scientific, technical, and clinical insights presented during the American Epilepsy Society (AES) 2025—Epilepsy Surgery Symposium on Advances in Thalamic Neuromodulation for Epilepsy: From Mechanisms of Action to Clinical Application. Beyond a summary of the individual presentations, this manuscript reflects the collective opinions of the panel regarding current evidence, unresolved controversies, and future directions. An unifying theme emerged: effective thalamic neuromodulation requires a shift from empiric, one-size-fits-all targeting toward patient-specific, hodology-informed strategies grounded in in situ electrophysiology and network-level biomarkers.

Introduction

The thalamus occupies a privileged position within large-scale epileptogenic networks, acting not only as a relay but as an active modulator of cortical excitability and synchronization.^1‐11 While anterior nucleus (ANT) stimulation has become an established therapy, ¹² growing clinical and experimental data suggest that the involvement of the thalamus during seizures and its propagation is heterogeneous across seizure types, nucleus-specific in some clinical scenarios, and strongly dependent on individual cortico-thalamic connectivity.^13‐16 The symposium convened leaders in clinical epilepsy surgery, systems neuroscience, and computational analysis to critically examine how emerging tools—particularly thalamic stereo-electroencephalography (SEEG)^9,17,18—are reshaping target selection, mechanistic understanding, and therapeutic strategies.

Scientific Rationale for Thalamic Target Selection

A central message of the symposium was that thalamic target selection should be driven by neurophysiology and connectivity rather than historical precedent alone. Current imaging and electrophysiological studies demonstrate early and sometimes asymmetric thalamic recruitment in many focal epilepsies, including posterior quadrant, temporal, and perisylvian syndromes.^9,19 These findings challenge the assumption that ANT is universally optimal and highlight the potential relevance of nuclei such as the pulvinar, centromedian, and ventral motor thalamus.

As highlighted by Damiani et al,^3,7 from a network-science standpoint, thalamic nuclei should be conceptualized as nodes with distinct cortical embedding. Selecting a stimulation target without understanding its patient-specific hodological context risks either undermodulating the epileptogenic network and inducing off-target effects. The group emphasized that diffusion imaging, ²⁰ functional connectivity, ²¹ and SEEG-derived effective connectivity metrics ²² provide a data-driven rationale for thalamic target selection in patients with medically and surgically intractable epilepsy.

Methods for Interrogating Thalamic Nuclei

Advances in SEEG methodology now permit safe and systematic sampling of thalamic structures alongside cortical regions. Modern implantation strategies allow simultaneous interrogation of cortical onset zones, propagation pathways, and deep network hubs, offering a unique opportunity to better understand thalamic involvement during ictogenesis.

According to Parvizi et al, ¹³ thalamic SEEG should not be viewed as an exotic extension of cortical monitoring but as a logical evolution of invasive electrophysiology. In keeping with this, the panel emphasized that thalamic SEEG is necessary but not sufficient for rational neuromodulation. Without rigorous hypotheses, careful temporal analysis, and integration with cortical data, deep recordings risk becoming descriptive rather than mechanistic. Thalamic SEEG should therefore be embedded within a structured network framework rather than deployed as a standalone diagnostic tool.

Signal Processing and Analytical Frameworks

Extracting mechanistic insight from thalamic recordings requires analytical tools capable of capturing directionality, timing, and frequency-specific interactions. Symposium discussions highlighted the importance of combining classical time–frequency analysis with modern causality, phase–amplitude coupling, and state–space approaches.

As discussed by Cash et al, ¹⁴ signal processing should be hypothesis-driven and clinically interpretable. While sophisticated models can reveal subtle thalamocortical interactions, their translational value depends on robustness and reproducibility at the single-patient level. A shared opinion was that biomarkers informing stimulation timing, frequency, and pattern may ultimately be more impactful than static anatomical targeting alone.

Cortical Network Effects of Thalamic Stimulation

Several presentations demonstrated that thalamic stimulation induces nucleus-specific and frequency-dependent changes in cortical connectivity. These effects extend beyond seizure suppression and include modulation of physiological rhythms and inter-regional coupling.

Specifically, Gregg et al ¹⁵ discussed that thalamic stimulation should be understood as a network intervention rather than focal therapy. The group emphasized that different stimulation paradigms—continuous, burst, or adaptive—may engage distinct mechanisms operating over multiple timescales, ranging from desynchronization to neuroplastic inhibition and network reorganization. This supports a move toward flexible, physiology-informed stimulation strategies tailored to individual network states. ¹⁵

Surgical Techniques and Early Clinical Outcomes

Practical aspects of thalamic electrode implantation, including trajectory planning, avoidance of vascular structures, and integration with cortical coverage, were discussed in detail. Early outcome data suggest that carefully selected patients may benefit from targeting non-ANT nuclei, particularly when guided by SEEG findings.

From a surgical standpoint, Gonzalez-Martinez et al highlighted the feasibility and safety of thalamic SEEG, concluding that neuromodulation has reached a level that justifies broader clinical adoption in specialized centers.^8,18,23‐25 However, the opinion was clear that thalamic targeting should remain hypothesis-driven and embedded within a comprehensive surgical strategy, rather than applied indiscriminately.

Ethical Considerations in Thalamic Neuromodulation

As thalamic neuromodulation expands beyond established targets and stimulation paradigms, ethical considerations become increasingly salient. Panel discussions underscored the responsibility of surgical teams to balance innovation with restraint, particularly when exploring off-label targets or adaptive stimulation strategies. Transparency in patient selection, informed consent, and expectation management were identified as essential safeguards.

A central ethical challenge relates to uncertainty. When neuromodulation is guided by emerging biomarkers or individualized network models, surgeons and neurologists must communicate both the scientific rationale and its limitations. The panel emphasized that ethical deployment does not preclude innovation, but rather demands rigorous hypothesis articulation, prospective data collection, and willingness to recalibrate strategies in response to outcomes.

Finally, equity of access was raised as an important consideration. Advanced thalamic SEEG and adaptive neuromodulation require specialized expertise and infrastructure, raising concerns about the concentration of innovation within select centers. Multicenter collaboration and standardized reporting were viewed as ethical imperatives to ensure that progress benefits the broader epilepsy community.

Panel Discussion: Consensus and Controversies

Across the panel, while there was strong agreement that thalamic neuromodulation represents a major inflection point in epilepsy surgery, the discussion also highlighted substantial areas of active controversy that currently limit widespread adoption and standardization.

A primary point of debate concerned whether thalamic activity represents a driver of epileptogenic networks or a secondary marker of propagation. Some panelists emphasized evidence of early thalamic recruitment and bidirectional cortico-thalamic coupling, supporting a causal role amenable to therapeutic modulation.^26,27 Others cautioned that, in certain syndromes, thalamic signals may reflect rapid entrainment rather than true network control, raising the risk of over-interpreting deep recordings without rigorous temporal and causal analysis.

A second controversy centered on target generalizability versus personalization. While ANT stimulation has Level I evidence and regulatory approval, the extension of neuromodulation to nuclei such as the pulvinar or centromedian remains largely supported by observational data and small series. Panelists differed on whether these nuclei should be explored broadly or reserved for narrowly defined electroclinical phenotypes. This tension reflects a deeper philosophical divide between syndrome-based algorithms and fully individualized, hodology-driven targeting.

Stimulation paradigms also generated debate. High-frequency, duty-cycle stimulation remains the clinical default, yet accumulating data suggest frequency-, pattern-, and state-dependent effects that may be better exploited through burst or adaptive strategies. Skeptics raised concerns regarding technical complexity, regulatory hurdles, and interpretability, while proponents argued that failure to move beyond static stimulation risks missing the true therapeutic potential of thalamic modulation.^3,27

Finally, there was disagreement regarding the evidentiary threshold required for clinical expansion. Some panelists advocated for cautious, protocolized adoption limited to specialized centers until prospective trials mature. Others argued that, given the morbidity of uncontrolled epilepsy and the relative safety of modern techniques, a more pragmatic expansion is justified when guided by strong physiological hypotheses.

Unresolved Questions and Points of Disagreement

Despite rapid progress, several fundamental questions remain unresolved. It is unclear which electrophysiological features reliably distinguish thalamic nuclei that are permissive to seizure spread from those that exert true network control—some panelists suggest that single pulse and repetitive thalamic stimulation during SEEG may resolve this question. The optimal balance between anatomical targeting and physiology-driven adaptation remains undefined, as does the extent to which thalamic SEEG findings can be generalized beyond expert centers. Moreover, whether adaptive or on-demand stimulation will prove superior to open-loop paradigms remains an open question that cannot be resolved without prospective, biomarker-driven trials.

Collectively, the panel agreed that transparency in reporting, standardized outcome measures, and multicenter collaborations are essential to resolving these debates.

Future Directions

The symposium concluded with a forward-looking discussion emphasizing prospective, multi-center studies integrating thalamic SEEG, advanced analytics, and adaptive neuromodulation. The collective opinion was that the next phase of progress will depend on closing the loop between mechanism and therapy—using in situ thalamic physiology not only to characterize seizure networks, but to actively guide personalized intervention.

A central challenge identified by the panel is the misalignment between rapidly evolving neuromodulation science and existing clinical trial and regulatory frameworks. Current paradigms favor fixed targets, static stimulation parameters, and narrowly defined endpoints, limiting the evaluation of adaptive and physiology-driven strategies. The authors collectively call for trial designs that permit controlled parameter evolution, incorporate mechanistic biomarkers as secondary endpoints, and recognize patient-specific targeting as a feature rather than a confound.

A specific constraint highlighted during the discussion concerns adaptive and closed-loop deep brain simulation (DBS). Present U.S. Food and Drug Administration investigational device exceptions and regulatory pathways are largely optimized for deterministic stimulation paradigms, where parameters are fixed a priori and modified infrequently. This structure complicates the testing of adaptive DBS approaches that rely on real-time biomarkers, dynamic parameter adjustment, or state-dependent stimulation logic. As a result, scientifically justified adaptive strategies are often forced into static frameworks that obscure their true mechanism of action and therapeutic value.

The panel emphasized that addressing this gap does not require reduced rigor, but rather updated regulatory constructs that explicitly accommodate algorithmic adaptation, predefined physiological decision rules, and transparent safety boundaries. Without such evolution, regulatory inertia risks reinforcing legacy stimulation models and slowing the translation of adaptive, biomarker-targeted neuromodulation innovations.

Conclusion

This symposium underscored a maturation of the field toward mechanistically informed, network-based thalamic neuromodulation. At the same time, it challenged the epilepsy community to critically reassess current regulatory, clinical, and conceptual paradigms that favor uniform targets and static stimulation strategies.

From a surgical perspective, thalamic neuromodulation represents not a departure from epilepsy surgery, but its logical extension. As resective and disconnective procedures reach their limits in complex network epilepsies, the ability to interrogate and modulate deep network nodes expands the surgical armamentarium. The authors jointly argue that continued reliance on generic neuromodulation risks constraining progress in individualized therapeutics. Instead, thalamic SEEG-guided, physiology-informed, and adaptive approaches should be viewed as integral components of modern epilepsy surgery.

Moving forward, progress will depend on embracing controlled innovation—balancing rigor with pragmatism—and on designing trials that allow neuromodulation strategies to evolve alongside our understanding of epileptic networks. In doing so, epilepsy surgery can continue its tradition of combining anatomical precision with physiological insight, ensuring that neuromodulation remains firmly grounded within the surgical discipline it seeks to advance.