The autonomic nervous system (ANS) comprises the sympathetic and parasympathetic nervous systems, which regulate visceromotor, neuroendocrine, pain perception, behavioral responses and organ activities (Wehrwein et al., 2016). Cardiovascular homeostasis maintained by the ANS is essential for survival (Mohanta et al., 2023; Silvani et al., 2016), and is mainly characterized by heart rate (HR) regulation, blood pressure modulation, and blood flow redistribution. The cortical areas that are involved in the central control of the ANS include the insular cortex, amygdala, hypothalamus, and periaqueductal gray matter, and together they form the neural network termed the central autonomic network (CAN) (Benarroch, 1993; Ruiz Vargas et al., 2016; Samuels, 2007). The role of the cerebral cortex within the CAN in regulating cardiovascular function requires further clarification.

Recent studies have progressively focused on the insula in the CAN in modulating cardiac activities (Mazzola et al., 2023; Oppenheimer and Cechetto, 2016). However, the respective contributions of insular lateralization and cytoarchitectonic subdivisions to specific cardiac outcomes remain controversial. Evidence suggests right insular damage is positively correlated with sympathetic dysfunction (Olga et al., 2022). Supporting this finding, de Morree et al. (2016) reported that right insular lobectomy significantly increased parasympathetic activity and induced bradyarrhythmia, whereas no changes were observed in the left insular resection. In contrast, Alice et al. (2023) found that atrophy in the left middle-to-posterior insula is associated with impaired parasympathetic activity and cardiac arrhythmias, while degeneration in the left ventral anterior insula correlates with reduced sympathetic tone. Low-frequency ultrasonic stimulation of the anterior insula in healthy individuals was found to alter heart rate variability (HRV), whereas stimulation of the posterior insula elicited no such effect (Legon et al., 2024). However, an electrical stimulation (E-Stim) study reported that both anterior and posterior insular regions participate in autonomic regulation, with anterior stimulation preferentially influencing parasympathetic tone and posterior stimulation affecting sympathetic function (Chouchou et al., 2019). Notably, clinical follow-up studies have identified two cases of sudden unexpected death in epilepsy (SUDEP) in patients with left insular damage. These individuals exhibited progressive alterations in HRV prior to death, suggesting that insular lobe lesions may constitute additional risk factors for SUDEP (Lacuey et al., 2016; Matusik et al., 2023).

Beyond its role as a critical upstream ANS region that requires further study, the insula elicits a spectrum of downstream cardiac effects (Nagai et al., 2010). E-stim studies report direct HR changes and impaired cardiac output (Chouchou et al., 2019; Sanchez-Larsen et al., 2021), while clinical investigations have primarily focused on its role in modulating HR variability (Lacuey et al., 2016; Danilin et al., 2022). In animal models, insular E-Stim induced seizures in 92.3% of rats, accompanied by typical autonomic changes—notably a decrease in HR. Among the rats that exhibited significant bradycardia, post-seizure mortality within 2 weeks reached 25% (Ming et al., 2020). Complementing these findings, non-invasive neuroimaging studies have extensively characterized the insula’s involvement in cardiac interoceptive accuracy and subjective cardiopulmonary sensation (Haruki and Ogawa, 2021; Hassanpour et al., 2018; Chong et al., 2017). Despite these advances, the precise functional organization of insular subregions underlying specific cardiac outcomes remains incompletely resolved.

The association between insular cortex function and HR dynamics has not been systematically investigated in large-scale studies. It remains unclear whether distinct insular subregions can elicit differential cardiac contractility or rhythm responses and how these responses are modulated by with E-Stim parameters. Clarifying these relationships may have important implications for preventing arrhythmias, sudden cardiac death, and enhancing the precision of epileptic focus localization prior to surgery. Direct cortical stimulation via intracranial electrodes offers superior temporal and spatial resolution over non-invasive methods like functional magnetic resonance imaging, thereby providing a unique capacity for causal validation of precise anatomical-functional relationships. The present study aimed to investigate the relationship between insular subregions and cardiac responses to direct insular E-stim, and to explore the roles of E-stim parameters and lateralization.

To the best of our knowledge, this is the first study to summarize the role of the middle short gyrus in insula in cardiac rhythm regulation in a large-scale SEEG-based study with 487 E-stims. In summary, 40.8% (199/487) of insula stimulations evoked a cardiac response, of which 19.5% (95/487) increased the HR and 21.3% (104/487) decreased the HR. The laterality of insular stimulation and current intensity were independently associated with both the incidence and directionality of evoked HR alterations. Compared to the distribution of the insula electrode contacts in the D-HR group, those in the E-HR group were predominantly distributed posterior to the insula under the same current intensity. Most notably, we pinpointed a specific functional subregion: E-Stim of the middle short gyrus robustly and specifically elicited significant cardiac deceleration.

The interaction of the vagus and sympathetic nerves is key to HR regulation and underlies the rhythmic contraction of the heart. While the classic literature on neurocardiology focused on the role of subcortical regions of the central ANS in the regulation of HR, more recent studies have suggested that the cerebral cortex can also regulate the ANS by acting on subcortical structures, including the insula (Benarroch, 2020), cingulate gyrus (Kim et al., 2019), hypothalamus (Henderson and Macefield, 2021), zona incerta (Messina and Overton, 2007), and frontal lobe (Harnod et al., 2009), to affect the HR. A few studies have reported the characteristics of HR changes after E-stim of the insula cortex. Oppenheimer and Cechetto (1990) found earlier in rats that E-stim of posterior insula cortical areas affected HR. In addition, a link between insula neural activity and HR has also been observed by non-invasive techniques (Legon et al., 2023, 2024). Owing to advancements in SEEG technology, direct observation of HR changes induced by E-stim in the human insula has become feasible in recent years. A retrospective study analyzing 100 E-stims performed on patients with epilepsy found that 90% of stimulation-induced clinical sensations, and 47% induced changes in HR (Chouchou et al., 2019). However, these findings cannot exclude potential confounding effects of stimulation-evoked paresthesia and current intensity on cardiac responses, particularly given our demonstration of significant current intensity-dependent HR modulation. The distribution and prevalence of E-Stim sites associated with HR variability are consistent with the results of the present study. Sanchez-Larsen et al. (2021) eliminated sensory confounders and reported predominant HR deceleration following insular stimulation (n = 100); however, their study neither identified potential tachycardia-inducing stimulation sites nor accounted for the significant interindividual variability inherent due to real-time HR monitoring analysis. The 40.8% incidence of HR alteration and the anatomical distribution of the relevant insula stimulation sites in our study of 487 insula E-stims, after excluding stimulation sites that had induced clinical symptoms and unifying the intensity of E-stim, remained broadly consistent with previous findings.

Current evidence remains controversial regarding the functional lateralization and subdivision-specific modulation of HR by the insular cortex. Leona et al. demonstrated a significant positive correlation between resting HR and the left insular cortex volume, contrasting with a significant negative correlation for the right insular cortex volume (Danilin et al., 2022). Nevertheless, animal experiments, revealed symmetrical cardiovascular responses (including blood pressure changes, HR changes, or renal sympathetic nerve activity) to chemical stimulation of bilateral insular cortices (Marins et al., 2016). In our study, the direction of HR response induced by insular E-stim is not significantly associated with insular lateralization, although the left insula appeared more prone to elicit cardiac reactions, which aligns with previous animal experimental findings. Notably, low-current-intensity stimulation tended to increase HR, while high-intensity stimulation produced the opposite effect. We cannot exclude the possibility that the HR elevation during low-current E-stim reflects an anxiety response to the initial phase in the experimental procedure rather than a direct current effect (Martha et al., 2021); conversely, high-current stimulation elicited more robust cardiovascular responses that were likely unaffected by external confounders and thus more accurately reflected the true cortical stimulation effects. These intensity-dependent observations corroborate Sanchez-Larsen et al.’s (2021) findings demonstrating more substantial HR alterations associated with higher-intensity insular stimulation. Additionally, the higher activation thresholds of autonomic or nociceptive fibers may lead to unintended stimulation of non-target nerve fibers (Umair et al., 2022), potentially introducing confounding effects on HR alteration through current spread to adjacent brain regions. Concurrently, antiepileptic drugs (AEDs) are known to modulate autonomic function in epilepsy patients. Gradual AEDs withdrawal has been associated with enhanced parasympathetic and sympathetic tone, accompanied by increased HRV (Lossius et al., 2007). Carbamazepine has been shown to induce significant systolic blood pressure changes during Valsalva maneuvers, indicating reduced sympathetic reactivity (Francesca et al., 2020). Lamotrigine has been documented to prolong QT intervals (Kevan and Djans, 2007). However, under experimental conditions comparable to the present study—whether comparing different monotherapy regimens (Francesca et al., 2020) or contrasting monotherapy versus polytherapy approaches (Omnia Fathy et al., 2015)—the observed alterations in HRV parameters remain relatively limited.

In our study, the posterior insula evoked a greater increase in HR, while the anterior part evoked a greater decrease. The insular cortex can be anatomically and functionally subdivided into three distinct subregions—the posterior granular, middle dysgranular, and anterior agranular areas—based on cytoarchitectonic organization, connectivity patterns, and functional specialization (Benarroch, 2019). In the present study, we demonstrate a clear functional dichotomy within this structure: E-Stim of the anterior agranular insular cortex elicits pronounced bradycardia, likely mediated through parasympathetic pathways (Chouchou et al., 2019), while activation of the posterior granular insular region produces significant tachycardia, presumably via sympathetic modulation. These findings align with previous neuroanatomical evidence (van der Kooy et al., 1984) that identified potential direct projections from the granular insular zone to the dorsal motor nucleus of the vagus nerve using fluorescent retrograde tracing techniques, further supporting the concept of topographically organized autonomic control within distinct insular subregions. Further analysis revealed that HR showed an immediate and significant decrease following stimulation onset in the middle short insular gyri, with a recovery trend observed within 5 s post-stimulation. In contrast, no significant HR variations were detected following stimulation of other insular subregions. These findings provide the first evidence that the middle short insular gyri may serve as the key functional locus within the anterior insula responsible for mediating HR deceleration. The study by Shinpei et al. (2016) demonstrates that the middle short gyrus serves as an upstream hub within a functional neural circuit involving the anterior short gyrus (ASG), which collectively orchestrates viscero-autonomic reflexes. Neurofunctional mapping studies have localized the middle short gyrus as the predominant cortical region responsible for insula-mediated functions, including pain perception (Afif et al., 2008) and speech production (Afif et al., 2009).

Neurocardiac damage is strongly associated with sudden unexpected death and increased perioperative risks in epilepsy surgery (Oveisgharan et al., 2022; Laoprasert et al., 2017). Notably, paroxysmal tachycardia or bradycardia frequently occurs in cases of insular epilepsy (Sami et al., 2017). Investigating the functional mapping between insular subregions and HR alteration may enhance the precision of epileptogenic zone localization while preserving critical autonomic neural circuits, thereby supporting cardiovascular homeostasis during the perioperative period. Importantly, structural or functional insular impairment—whether due to epileptic foci (Lacuey et al., 2016), stroke (Romano et al., 2019), or surgical resection (de Morree et al., 2016)—elevates mortality risk. The incidence of SUDEP is estimated to range from 0.7 to 1.3 per 1,000 patient-years, escalating to 3.5–4.1 per 1,000 patient-years in those with refractory epilepsy (Buchanan et al., 2023; Walczak et al., 2001). Several clinical studies have demonstrated a correlation between SUDEP, HR variability, and the insula (Kassinopoulos et al., 2023; Scorza et al., 2022). In a 3-year longitudinal study of 320 patients, two cases of SUDEP were observed, both involving individuals with prior insular injury who exhibited progressive HRV deterioration preceding death (Lacuey et al., 2016). To date, the pathophysiology of SUDEP remains incompletely understood, though it often involves bradyarrhythmia progressing to asystole and fatal cardiac arrest (Devinsky et al., 2016; Li et al., 2017). In our study, we found that the middle short insular gyri may be a key node for triggering bradycardia in the HR regulation pathway. Future research should conduct prospective, causally designed clinical trials based on closed-loop stimulation or telemetry-integrated monitoring to specifically assess the therapeutic potential of the middle short insular gyri in preventing SUDEP and improving the efficacy of epilepsy surgery.

The present study has several limitations. First, it is a retrospective, observational study that could not standardize all individual-specific factors, including AEDs, stimulation parameters, and electrode placement. Consequently, we could not fully control for confounding factors arising from patient heterogeneity, individual neuroanatomical variations, and differences in electrode coverage. Second, potential confounding factors from procedural stress on HRV could not be completely eliminated. Third, the spatial resolution of E-Stim is limited by current spread to adjacent gyri or subcortical regions, potentially causing nonspecific activation. Fourth, applying mirror processing could potentially obscure these biologically meaningful hemispheric differences. Furthermore, we did not monitor cardiac function, blood pressure, and respiratory function, which precludes analysis of these additional autonomic parameters and may reduce ECG signal quality. Compared with traditional leads, deltoid leads are inferior in terms of electrical signal quality and artifacts. In the future, the use of complete electrocardiogram detection equipment synchronized with EEG signals will improve data quality and reliability, and the short E-stim durations and intervals cannot reveal long-term autonomic effects in this study. Lastly, there were multiple E-Stims originating from the same site or the same patient, which may have introduced clustering-related bias into the analysis results. Future studies could design prospective experiments with stratified analysis to eliminate any existing bias as much as possible.

Approximately 40.8% of insular E-stim induced a cardiac response (19.5% increased the HR and 21.3% decreased the HR). Our findings suggest that insula subdivisions and lateralization may play an important role in the regulation of HR. Stimulation sites inducing HR elevation were anatomically localized to posterior insular regions relative to bradycardic response. Among these, we found a correlation between the middle short gyrus and HR reduction. This study is the first to observe the role of the middle short gyrus in cardiac rhythm regulation and may provide a new direction for pre-surgical assessment of epilepsy and the prevention and treatment of SUDEP.