Trigeminal neuralgia (TN) is a chronic neuropathic disorder characterized by sudden, electric shock-like facial pain - either spontaneous or triggered - that typically affects a specific facial region and may persist between paroxysms. This disease can significantly impair quality of life, and in refractory cases, may be associated with suicidal ideation (Zakrzewska et al., 2017). The diagnosis of TN is based on three clinical criteria: (1) pain localized to one or more trigeminal nerve branches, (2) paroxysmal, electric-shock-like pain episodes, and (3) pain triggered by innocuous stimuli in the affected facial or intraoral regions. Triggered paroxysmal pain is pathognomonic of TN, occurring in 91%–99% of reported cases (Di Stefano et al., 2018; Cruccu et al., 2016; Maarbjerg et al., 2014). TN is currently classified into three distinct subtypes: classical (the most common), secondary, and idiopathic. The classical form arises from neurovascular compression (NVC) at the trigeminal nerve root entry zone (REZ), typically causing focal demyelination. NVC can be visualized using functional magnetic resonance imaging (fMRI) and 3D reconstruction. Anticonvulsants such as carbamazepine (typically 200–1,200 mg daily) and oxcarbazepine (300–1800 mg daily) are considered first-line treatments for managing paroxysmal pain in TN. Although these drugs are effective, they are associated with significant side effects and often have limited efficacy in treating persistent pain (Bendtsen et al., 2019). Microvascular decompression (MVD) has emerged as the treatment of choice for medication-refractory TN (Cruccu et al., 2016). MVD represents the most effective surgical treatment for classical TN. Postoperative outcomes demonstrate 68%–88% initial success rates (1–2 years), with 61%–80% of patients sustaining pain relief at 4–5 years follow-up (Tuleasca et al., 2014; Zakrzewska and Akram, 2011). Resting fMRI (rs-fMRI) has become an invaluable tool for examining central mechanisms, particularly in structure and mapping functional connectivity within the human brain (Biswal et al., 2010). Patients with TN show significant cortical thinning in the anterior cingulate cortex (ACC), posterior cingulate cortex (PCC), and midcingulate cortex (MCC), accompanied by extensive subcortical volume reductions (Mo et al., 2021). In addition, TN patients exhibited significantly greater cortical thickness in the contralateral primary somatosensory cortex and frontal pole compared to controls (DeSouza et al., 2013). Compared with healthy controls (HC), TN patients showed significantly altered regional homogeneity (ReHo) and fractional amplitude of low-frequency fluctuations (fALFF) in multiple brain regions, including the cerebellum, temporal lobe, cingulate cortex, putamen, limbic lobe, occipital lobe, precuneus, insula, and medial/superior frontal gyri (Yuan et al., 2018). TN patients displayed increased functional connectivity between the mPFC and left planum temporale, alongside decreased connectivity between the mPFC and left superior frontal gyrus (Wu et al., 2025). Another study revealed that patients with TN exhibited increased functional connectivity between the right insula/secondary somatosensory cortex (S2) and several key regions, including the ACC, PCC, medial prefrontal cortex (mPFC), and bilateral dorsolateral prefrontal cortex (dlPFC) (Wang et al., 2017). The sensorimotor network plays an important role in the pain modulation, as it integrates sensory input and motor output, which is crucial for bodily movements and responses to sensory stimuli. Xu et al. found that distinct patterns of functional connectivity are found in default mode network, somatosensory network, and salience network in TN (Xu et al., 2022). Disruptions in this network can be indicative of neurological disorders or conditions like TN. Costs et al. found that The TN group performed significantly worse than controls on both sensorimotor (tracking and aiming) and cognitive tasks (Coats et al., 2020). Recent fMRI studies indicate that sensorimotor impairments in low back pain (LBP) may be linked to alterations in brain function and structure (Goossens et al., 2018). Voxel-based morphometry (VBM) analyses demonstrated widespread gray matter volume (GMV) reductions throughout the brain in classical TN patients, particularly affecting the insula/secondary somatosensory cortex (S2), cingulate cortex, and multiple temporal regions (Wang et al., 2017). Zhou et al. reported that TN patients exhibited enhanced functional connectivity between the somatosensory-motor network and the dorsal attention network (Zhou et al., 2024). Tsai et al. found the activity in the sensorimotor network and default mode network was reduced in TN patients but increased following surgery (Tsai et al., 2019). Applying transcranial direct current stimulation (tDCS) over motor cortex reduces somatosensory cortex activity, which may lead to pain relief (Khodashenas et al., 2019). Currently, there remains a paucity of research investigating TN-related alterations in sensorimotor networks, as well as the impact of MVD surgery on these networks. Previous studies have analyzed blood oxygen level-dependent (BOLD) signal time series using general linear modeling (GLM), which are well-established correlates of task-evoked brain activity. However, GLM has two key limitations: it relies on a predefined experimental design matrix, and it cannot detect functional connectivity between brain regions (Kim et al., 2009; Specht and Laeng, 2011). To mitigate these limitations, we used independent component analysis (ICA), a blind-source separation method, to identify distinct clusters of brain regions with coherent hemodynamic temporal patterns without relying on prior temporal response information (e.g., an experimental design matrix). Unlike previous research that examined isolated regions, we used ICA to assess functional connectivity changes within sensorimotor networks in patients undergoing microvascular decompression surgery for classic TN, yielding potential insights for future interventions.

This study employed ICA to examine functional connectivity changes after MVD surgery for TN. Unlike hypothesis-driven approaches, ICA is data-driven and can reveal patterns of brain dysfunction beyond the constraints of predefined experimental models. Our findings successfully revealed changes in individual functional connectivity associated with MVD surgery.

Previous studies have used ICA to identify consistent neural activity patterns across groups of individuals (Eyler et al., 2019). Compared to HC group, we found increased activity in the right frontal pole, right frontal operculum cortex, right insular cortex and right inferior frontal gyrus, pars opercularis pre-surgery of the TN group and decreased activity in the right cerebellum (lobule IX) and the left cerebellum (lobules VIII and lobules IX). The frontal pole is associated with various functions, including motor control and sensory processing (Koechlin, 2011) Changes in connectivity in this region may indicate altered neural processing related to pain. Recent neuroimaging studies have reported that pain induces both structural and functional plasticity in pain-processing brain regions, including the PFC, insula, ACC, and somatosensory cortex (Coppieters et al., 2016; Niddam et al., 2017), these changes affect sensory, emotional, and cognitive dimensions of pains. Our results are consistent with the previous study.

The insula plays a critical role in both physiological and possibly pathological pain (Garcia-Larrea, 2012; Brooks and Tracey, 2007). The insular cortex is one of the earliest brain regions activated by nociceptive stimuli. Its activation level strongly correlates with both the intensity of noxious stimuli and the subjective perception of pain (Coghill et al., 1999; Oshiro et al., 2009). The relationship between altered insular activity (Hsieh et al., 1995), structural changes (Gustin et al., 2011), and pain chronification has been increasingly recognized. In chronic LBP patients, studies have demonstrated disrupted functional connectivity of the insular with the somatosensory (Kim et al., 2019) and the medial prefrontal cortices (Baliki et al., 2011). On the other hand, pharmacological interventions for chronic pain have been shown to partially reverse these aberrant insular connectivity (Čeko et al., 2015). The observed increase in activity in the right insular cortex may indicate a neural adaptation or maladaptive changes due to chronic pain in our study. This change could reflect heightened processing of pain. In our study, we found the MVD surgery can effectively reduce the active areas in the insular cortex. However, the persistent insular activation postoperatively may contribute to the recurrence of TN in the future. Recent clinical studies have proposed the insular cortex as a potential therapeutic target for deep brain stimulation in chronic pain management. In healthy subjects, transcranial magnetic stimulation (TMS) of the posterior opercular insular cortex was found to elevate heat pain thresholds mediated by Aδ fibers (Lenoir et al., 2018). Similarly, repetitive TMS targeting the posterior insula produced significant increases in thermal pain thresholds among central neuropathic pain patients (Galhardoni et al., 2019). Notably, a study demonstrated TMS in posterior superior insula has a significant analgesic effects when applied to patients with peripheral neuropathic pain (Dongyang et al., 2021). These consistent findings across multiple pain models suggest the insular cortex may serve as a promising target for non-pharmacological pain interventions. Although our results confirm that MVD surgery is effective in reducing insular activation area, further interventions target insular cortex may be needed in the future to enhance efficacy.

The right inferior frontal gyrus (rIFG) comprises part of Broca’s area homolog, a region traditionally implicated in speech production and language processing. However, it is also involved in emotional processing, executive function, and integration of sensory information, making it relevant to pain procession and responses (Symonds et al., 2006). Recent tDCS and repetitive TMS studies suggest that the rIFG may be involved in pain-related empathy, yielding novel insights into the neurobiological substrates underlying empathy for others’ suffering (Wu et al., 2018; Li et al., 2021). Our results demonstrate that increased rIFG activity persists both pre- and post-operatively, potentially reflecting enhanced pain empathy resulting from the long-lasting pain experience in TN.

The operculum, defined as the cortical region adjacent to the insula, comprises three major subdivisions and plays well-established roles in various neurological and psychiatric disorders. This structurally complex area mediates multiple functions including sensory processing, motor control, autonomic regulation, and cognitive operations (Garcia-Larrea, 2012). The dorsal posterior insula and adjacent medial operculum are considered nociception-specific regions. Lesions in this area may predispose to neuropathic pain development, while the adjoining medial operculum mediates non-noxious thermal processing (Garcia-Larrea, 2012). Patients with chronic traumatic neck pain exhibited enhanced left pallidum-left frontal operculum connectivity compared to HC (Coppieters et al., 2021). The operculo-insular cortex plays a crucial role in pain processing. Building on this understanding, Ulf Baumgärtner and colleagues investigated somatotopic organization within this region across 11 healthy subjects using noxious heat and pinprick stimulation paradigms. Their study revealed multiple discrete pain representations within this region, likely reflecting its critical role as a multisensory integration hub that generates body site-specific emotional and behavioral responses to injury (Baumgärtner et al., 2010). Our results indicate increased activity in this brain area related neuropathic pain development in TN. MVD surgery is effective in reducing operculum activation area.

The cerebellum plays a modulatory role in pain through its communication with sensorimotor, executive, reward, and limbic functional regions. Although substantial evidence suggests that the cerebellum may be critically involved in trigeminal nociceptive pain, few studies have investigated this function. Even less is known about which specific cerebellar regions participate in trigeminal surgery. In a recent meta-analysis, Moulton and colleagues demonstrated that multiple cerebellar regions are typically activated during nociceptive processing in human, beyond this, the cerebellum contributes to pathological nociceptive sensitivity and processing in chronic pain (Moulton et al., 2010). In addition to evidence of altered activation patterns during stimulation in chronic pain, PET studies have also investigated baseline activity levels in individuals experiencing chronic pain, with some reporting sustained cerebellar blood flow changes (Hsieh et al., 1995). Beyond altered blood flow, the cerebellum in patients with chronic orofacial neuropathic pain also exhibits modified patterns of ongoing activity (Alshelh et al., 2016). In addition to observed cerebellar activation in chronic pain patients, recent neuroimaging studies have also identified abnormal functional connectivity between the cerebellum and other cortical/subcortical areas in these individuals (Simons et al., 2014). We are the first found abnormal connectivity in the right cerebellum (lobules IX) and the left cerebellum (VIII and IX) in TN. Our results indicated MVD can effectively treat this abnormality, this could be considered an important criterion for postoperative evaluation. Investigating connectivity changes pre- and post-peratively in TN is very important. They could lead to the identification of biomarkers for TN severity or potential treatment evaluation.

The precentral gyrus, also known as the primary motor cortex or Brodmann area 4, is located in the frontal lobe of the brain and is responsible for initiating and controlling voluntary movements (Banker and Tadi, 2019). The precentral gyrus also play a modulatory role in pain processing. Previous studies have identified significant group differences in functional connectivity between the right precentral gyrus and regions such as the cingulate gyrus, precuneus, and left paracentral lobule in patients with chronic shoulder pain (Wei et al., 2022). Our findings suggest that the motor cortex may serve as a key brain region in chronic pain, and that MVD can effectively enhance its activity, thereby mitigating chronic pain progression. This aligns with prior moderate-quality evidence demonstrating that high-frequency rTMS of the primary motor cortex induces significant analgesic effects in neuropathic pain (NP) and fibromyalgia (Xiong et al., 2022).

While our findings demonstrate potential clinical relevance, this study has limitations. Most notably, the modest sample size may reduce statistical power and limit the generalizability of the results. Future research with larger samples is needed to validate these results. Second, our fMRI preprocessing used head motion thresholds of 2.5 mm translation and 2.5° rotation, which may have been insufficiently conservative. This leniency could have introduced motion-related artifacts, potentially compromising the validity of the functional connectivity results. In addition, longer-term data tracking would be beneficial to further investigate the correlation between TN recurrence post-surgery and activation in these brain regions. Notwithstanding these limitations, our findings demonstrate the potential utility of ICA for evaluating post-operative pain outcomes following MVD surgery, while providing a neurobiological foundation for developing targeted interventions.

Our study demonstrates that ICA can effectively identifies distinct patterns of functional connectivity associated with TN and MVD surgery. In the sensorimotor network, TN was found to increase neural activity in broad brain regions, including the right frontal operculum cortex, right insular cortex, right inferior frontal gyrus (pars opercularis), and right frontal pole. Conversely, decreased activity was found in the right cerebellum (lobule IX) and left cerebellum (lobules VIII and IX). These regions are critically involved in altered pain processing, including nociceptive stimuli integration, subjective pain perception, and pain chronification. Additionally, TN enhanced activity in areas linked to pain-related empathy while disrupting cerebellar function. Compared to the pre-surgey, increased activity was found in the right precentral gyrus in the post-surgery group. Although MVD surgery is an effective treatment, long-term hyperactivity persisted in the right frontal operculum cortex, right insular cortex, right inferior frontal gyrus (pars opercularis), and right frontal pole. However, MVD significantly reduced aberrant activation in the operculo-insular cortex and normalized cerebellar abnormalities. Our findings identify promising biomarkers for TN and provide insights for developing targeted treatments.