Innovating transcranial magnetic stimulation treatment for post-traumatic stress disorder: Zapping away the bad memory

Abstract

Post-traumatic stress disorder (PTSD) is a debilitating psychiatric condition characterized by dysregulated fear memory, emotional disturbances, and impaired cognition, driven by dysfunctions across multiple neural circuits and neuroinflammatory processes. While current treatments demonstrate limited efficacy, repetitive transcranial magnetic stimulation (rTMS) has emerged as a promising non-invasive intervention. This review synthesizes preclinical and clinical evidence, highlighting rTMS mechanisms—including enhanced neuroplasticity, normalizing network connectivity, and reducing neuroinflammation. We then propose five innovative strategies, including accelerated protocols, precision targeting using neuroimaging, neuroinformatic approaches, closed-loop systems, and biomarker-guided treatment. By bridging circuit-level insights with clinical innovation, rTMS offers a transformative approach to PTSD treatment, though standardization and personalized paradigms require further development and validation. Future research should integrate multimodal biomarkers with hybrid therapies to optimize outcomes.

Graphic abstract

Schematic illustrating the therapeutic mechanisms of rTMS in PTSD and associated technological innovations (e.g., accelerated protocols, personalized targets, closed-loop systems) that enhance treatment precision and efficacy. This figure was created in BioRender. j, s. (2025) https://BioRender.com/cr72wj7.

1. Introduction

Post-traumatic stress disorder (PTSD) remains one major challenge in mental disorders, with symptoms such as impaired cognition, severe mood disturbances, flashbacks, nightmares, and intrusive thoughts related to traumatic events. PTSD develops following the experience or witnessing a terrifying event. Its pathophysiology involves aberrant neural connectivity and disrupted network dynamics, particularly within amygdala-centered circuits [1,2]. Current psychotherapeutic and pharmacological treatments demonstrate limited efficiency in clinical practice. Over the past decade, neural circuit-based brain stimulation approaches, especially non-invasive techniques like transcranial magnetic stimulation (TMS), have emerged as promising therapeutic interventions for depression and PTSD [[3], [4], [5]].

Originally developed in the 1980s to measure motor cortical excitability [6], TMS operates by passing transient electrical currents through a coil. This generates an alternating magnetic field that penetrates the skull, inducing cortical activation via electrical currents produced through electromagnetic induction (Faraday’s Law). TMS has since become a valuable technique for probing causal brain structure-function relationships [7]. Repetitive TMS (rTMS) represents the predominant clinical application, in which scalp-positioned electromagnetic coils generate targeted electric fields in cortical tissue. Clinically, figure-eight or H-coil configurations are typically employed for focal stimulation. These coils induce action potentials primarily through depolarization of myelinated cortical axons. The resulting neural activity propagates through both local (cortico-cortical) and distal (cortico-subcortical) pathways, ultimately influencing interconnected regions—including remote and contralateral areas [[8], [9], [10]]. rTMS produces lasting, plasticity-like alterations in cortical excitability, making it suitable for treating conditions such as post-stroke motor rehabilitation, depression, and drug addiction [[11], [12], [13]]. The recent development of wearable rTMS devices offers patients greater convenience and provides new tools for neuroscience research [14].

2. Brain circuit dysfunction in PTSD

2.1. The vulnerable brain regions involved in PTSD

Neurological studies indicate that PTSD symptoms—including abnormal fear conditioning, dysregulated discrimination patterns, and emotional dysregulation—are closely associated with neuroanatomical and neurocircuitry alterations [15]. Structural and functional magnetic resonance imaging (MRI) reveals that changes in the prefrontal cortex, anterior cingulate, amygdala, hippocampus, and insula contribute significantly to PTSD pathophysiology [16,17] (Fig. 1).

Fig. 1.

Brain regions and neurocircuits implicated in PTSD [49,51,58]. (a) Key brain regions identified in human neuroimaging studies and rodent models of PTSD. Schematics illustrate dysregulated neural circuits, including: (b) Fear learning and threat detection. Sensory inputs activate the amygdala (LA→BLA→CeA), driving fear responses. In PTSD, amygdala hyperactivity enhances fear conditioning, while PFC dysregulation impairs inhibition of BLA activity. (c) Contextual processing: The hippocampus encodes contextual memories and interacts with the PFC for retrieval. Weakened hippocampal-PFC connectivity in PTSD compromises safe-context discrimination. (d) Emotional regulation: The thalamus relays emotional stimuli to the ACC and insula, which modulate physiological responses. PTSD patients exhibit emotional overmodulation (e.g., numbness) due to insula/ACC hyperactivity.

2.1.1. Prefrontal cortex (PFC)

The prefrontal cortex (PFC), situated at the anterior portion of the frontal lobe, integrates and transmits information from diverse brain regions. It mediates critical functions including cognitive processing, motor control, and emotion regulation [18]. In PTSD, symptoms such as intrusive thoughts, hyperarousal, and altered cognition/mood are closely associated with PFC dysfunction. Structural MRI studies reveal volume reductions in key PFC subregions of PTSD patients, including the ventromedial PFC (vmPFC) [19] and dorsal anterior cingulate cortex (dACC) [20]. Notably, PTSD symptom severity is negatively correlated with anterior cingulate cortex (ACC) volume.

Functional neuroimaging further demonstrates decreased activation in the medial PFC (mPFC) and ACC during fear response processing tasks in PTSD [19,20]. Meta-analyses of emotion processing tasks consistently report hypoactivation in the vmPFC, dorsomedial PFC (dmPFC), and ACC. Crucially, reduced ACC activation is linked to impaired fear memory extinction [21], and the lower mPFC activation correlates with greater symptom severity [22]. Importantly, the therapeutic benefit of serotonin reuptake inhibitors in PTSD patients has been associated with increased activation within PFC regions [23].

2.1.2. Hippocampus

The hippocampus, a limbic structure located in the medial temporal lobe, is critically involved in memory processing, emotional regulation, and spatial navigation. It integrates information from multiple brain regions and facilitates fear memory encoding in PTSD. Structural MRI (sMRI) studies consistently reveal hippocampal volume reductions in PTSD patients. Meta-analyses confirm bilateral volume loss [24,25], with the most pronounced reductions occurring in the Cornu Ammonis 3 (CA3) and dentate gyrus (DG) subfields. Notably, diminished volumes in CA2/3 and DG correlate negatively with intrusion symptom severity [26,27]. High-resolution sMRI further validates these subfield-specific deficits. Complementary magnetic resonance spectroscopic imaging (MRSI) indicates that volume loss reflects neuronal reduction, evidenced by decreased N-acetyl aspartate (NAA)—a neuronal density marker. These deficits appear gender-independent [28].

Functionally, fMRI studies associate impaired fear extinction learning in PTSD with reduced hippocampal activation [29]. Paradoxically, increased hippocampal blood flow correlates with symptom severity [30], consistent with findings linking flashback intensity to elevated regional cerebral blood flow (rCBF) in left parahippocampal regions in PTSD patients [31].

2.1.3. Amygdala

The amygdala, a limbic structure in the medial temporal lobe, is primarily responsible for threat detection, fear responses, and emotional memory enhancement. In PTSD, fear reactivity and avoidance behaviors are closely linked to amygdala dysregulation. While structural alterations in the amygdala remain inconclusive in PTSD, functional neuroimaging (e.g., fMRI) consistently demonstrates amygdala hyperactivity that positively correlates with symptom severity [32]. This hyperactivity drives hypervigilance toward potential threats and potentiates the acquisition of conditioned fear.

2.1.4. Animal models

Preclinical studies using established PTSD animal models—including inescapable foot shock, early life stress, single prolonged stress (SPS), social defeat stress (SDS), acute restraint stress (ARS), and contextual fear conditioning—consistently implicate the prefrontal-amygdala-hippocampal circuitry in PTSD pathophysiology. These models demonstrate the critical involvement of these neural circuits in stress response and fear processing [[33], [34], [35], [36], [37], [38]]. Reduced dendritic spine density on pyramidal neurons with concomitant microglial proliferation and increased process complexity were identified in the mPFC and dorsal hippocampus of PTSD-modeled mice [39]. Furthermore, excitatory synaptic transmission from basolateral amygdala (BLA) projection neurons to both the lateral central amygdala (CeL) and anterodorsal bed nucleus of the stria terminalis (adBNST) was significantly attenuated. Notably, deep brain stimulation (DBS) restoring BLA-CeL and BLA-adBNST circuit function reduced anxiety and fear behaviors [40]. Studies in ARS-exposed mice reveal that upregulated dmPFC-BLA glutamate release drives acute anxiety-like behaviors [35].

The SPS rat model revealed elevated tumor necrosis factor-α (TNF-α) and interleukin-1β (IL-1β) expression in hippocampal microglia 72h post-trauma, indicating neuroinflammatory processes preceding observable fear extinction deficits [41]. Studies in an acute SDS mouse model demonstrated that traumatic social avoidance engrams are physically encoded within BLA neuronal ensembles [38]. In a rat model of PTSD induced by electric shock and situational reminders, altered oxytocin receptor (Oxtr) expression and neuroinflammation were observed in the mPFC and BLA regions. Notably, intra-mPFC microinjection of 3,4-methylenedioxymethamphetamine (MDMA) pre-extinction enhanced fear extinction, improved social interaction, and reduced freezing [42], indicating targeted neuromodulation of these circuits can ameliorate PTSD-like phenotypes.

Collectively, synaptic dysfunction and neuroinflammation within prefrontal-amygdala-hippocampal circuits drive behavioral pathology, with circuit-targeted interventions demonstrating therapeutic potential.

2.2. Prefrontal cortex circuits dysfunction in PTSD

2.2.1. Fear learning/threat detection circuitry

The neural circuitry mediating fear learning and threat detection involves the amygdala, PFC, ACC, and insula. The amygdala integrates sensory inputs from the thalamus and orchestrates threat responses through projections to the hypothalamus, basal ganglia, and brainstem. The mPFC provides top-down regulation of subcortical structures to modulate behavioral responses [43], while the dorsal ACC (dACC) critically facilitates fear expression [44]. The insula functions as a key node within the salience/threat detection network, enabling identification of environmental safety signals. Dysregulation within this network disrupts salience-default mode network balance, contributing to PTSD pathophysiology [45].

The fMRI studies reveal that trauma induces heightened bottom-up amygdala→mPFC excitation coupled with diminished top-down dorsolateral PFC (dlPFC)→amygdala regulation [46]. Resultant amygdala hyperactivity potentiates fear memory consolidation. The prelimbic cortex integrates ventral hippocampal and basolateral amygdala (BLA) inputs to orchestrate fear responses [47]. During threat exposure, suppressed PFC activity compromises both emotion regulation and amygdala inhibition. Critically, ventromedial PFC (vmPFC) hypoactivation underlies impaired fear extinction in PTSD [29].

2.2.2. Context processing circuitry

The hippocampus-mPFC circuit is primarily responsible for contextual learning and memory. As previously noted, the hippocampus plays an important role in episodic memory and spatial representation. It forms unique contextual representations by integrating spatial information with non-spatial elements (e.g., time, prior experience, emotion) into a unified gestalt [48]. Additionally, hippocampus-mPFC interactions facilitate contextual retrieval [49]. The core PTSD symptom of pattern discrimination impairment is closely linked to hippocampus memory dysfunction. Normally, the hippocampus establishes new memories and recalls existing ones through environmental cues, primarily via two complementary processes: pattern separation and pattern completion [50]. Pattern separation enables differentiation between similar contexts (e.g., distinguishing artificial stage flowers from real garden flowers), which supports safety-threat discrimination. Pattern completion allows full memory retrieval from partial cues, enabling rapid protective responses during perceived threats.

In PTSD patients, pattern separation dysfunction contributes to recurrent trauma-related memories triggered by non-threatening partial cues, resulting in generalized fear. Concurrent pattern completion hyperactivity impedes threat-safety discrimination, leading to hypervigilance and hyperarousal [51]. The mPFC connects directly to the hippocampus and contributes critically to contextual memory processing. It integrates multimodal event-related features—including associations between events, contexts, locations, and emotions—organizing experiential components into coherent memories [52]. Dysfunction in this system causes inappropriate contextual responses in PTSD, manifesting as exaggerated reactions to potential threat indicators and heightened vigilance. Functional neuroimaging reveals decreased mPFC activity in PTSD, which impairs both amygdala inhibition and functional hippocampal interactions. This interconnected circuitry dysfunction underlies impaired contextual memory and compromised safety recognition [53].

2.2.3. Emotion regulation circuitry

Emotional regulation abnormalities constitute a significant contributor to PTSD symptomatology [51]. Emotional dysregulation manifests primarily through two distinct syndromes: emotional undermodulation and emotional overmodulation [54]. Emotional undermodulation reflects diminished emotional control, wherein patients exhibit heightened responses to fear, anger, and depression. Conversely, emotional overmodulation involves excessive emotional restraint, manifesting as numbness, environmental disengagement, and analgesia [55].

In PTSD development or maintenance, dysfunction within the mPFC-limbic circuitry and disrupted connectivity contribute to these regulatory impairments. Decreased vmPFC activity correlates with increased amygdala activation and reduced inhibition of limbic regions, driving heightened emotional responses and traumatic re-experiencing [56]. Concurrent hyperactivity in the insular cortex and ACC reflects amplified interoceptive awareness and depersonalization. Critically, emerging evidence indicates that fluctuating activity patterns across these regions better explain the abrupt symptom shifts observed when patients attempt emotional regulation [57,58].

2.3. The neural connectivity dysfunction in PTSD

The three-brain-network model has been proposed as critical to cognition regulation in psychiatric and neurological disorders [59]. These comprise the default mode network (DMN), salience network (SN), and central executive network (CEN) [54,59,60]. The DMN—consisting of mPFC, posterior cingulate cortex, and lateral parietal lobes—primarily mediates self-referential processing, social cognition, autobiographical memory, and future-oriented thinking. Studies in PTSD patients reveal altered DMN connectivity with other brain structures during both rest and task conditions [61,62], with DMN hyperconnectivity correlating with dissociative symptoms [63].

The SN, composed of dorsal anterior cingulate cortex (dACC) and anterior insular cortex, detects personally salient internal and external stimuli to guide homeostasis-maintaining behaviors. Its anterior insula dynamically switches connectivity between CEN and DMN during attentional and cognitive processing [64]. Altered SN connectivity may disrupt threat detection circuits, contributing to hyperarousal symptoms [65].

The CEN—a frontoparietal network essential for working memory and cognitive control—shows reduced resting-state connectivity in PTSD patients. Conversely, during memory tasks, CEN connectivity strengthens but fails to recruit auxiliary memory regions [66]. CEN impairment underlies cognitive deficits. Particularly, treatment strategies targeting these neural networks to restore function are proposed later in this review [59].

3. Current advances in rTMS treatment of PTSD

Considerable evidence suggests that targeting bilateral or unilateral dlPFC reduces PTSD symptoms and associated depression [4,[67], [68], [69], [70]]. The treatment protocols were modeled after depression paradigms, employing high-frequency (5 or 10 Hz, presumed excitatory) rTMS over left dlPFC or low-frequency (1 Hz, presumed inhibitory) rTMS over right dlPFC. The novel form of TMS—theta burst stimulation, which efficiently induces long-lasting cortical plasticity [71]—has recently proven effective for PTSD treatment [[72], [73], [74]]. A meta-analysis indicated that high-frequency protocols may yield superior treatment effects compared to low-frequency approaches [4]; different protocols might also benefit distinct clinical symptom domains and comorbidities (e.g. depression) [75]. Nevertheless, multiple factors contribute to cross-site discrepancies, including standardization of TMS administration, limited sample sizes, heterogeneous patient severity and trauma etiology, targeting precision, and technical variations across TMS systems.

Apart from dlPFC as the most common cortical target, dmPFC stimulation has demonstrated efficacy in reducing PTSD comorbid with eating disorders in an open-label case series [76]. Another study reported that a single session of rTMS at one pulse per minute over the bilateral primary motor cortex (M1) yielded symptom improvement [77]. In addition, low-frequency (1 Hz) TMS over the early visual cortex reduces the emotional intensity of intrusive memories [78]. Evidence for non-dlPFC targets remains limited, requiring further investigation to establish consistency.

With advancing TMS technology, deep TMS (dTMS) targeting deeper cortical and subcortical structures has become clinically approved for depression treatment. A recent study exploring mPFC-targeted dTMS using Brainsway’s H7 coil reported negative therapeutic outcomes [79]. One potential confounder was brief script exposure preceding TMS, previously effective in their trial [80]. Memory retrieval may theoretically “re-open” plasticity windows for extinction training, as demonstrated in fear extinction and addictive memories [81,82]; however, interactions between memory retrieval paradigms and brain stimulation for traumatic memories still require further investigation (Fig. 2).

Fig. 2.

Innovative TMS treatment approaches for PTSD. (a) Traditional mode: Conventional daily sessions (e.g., one session per day for 20 continuous days). (b) Accelerated mode: Multiple daily sessions (e.g., 5–10 sessions per day over 5 days). (c) Network-based mode: Stimulation focused on specific neural circuits of deep brain regions using neuronavigation. (d) Information-based mode: Combined approaches (e.g., VR exposure therapy synchronized with TMS). (e) Closed-loop mode: Wear-on TMS with EEG-guided real-time modulation (emerging approach).

As a promising therapeutic intervention for PTSD, the clinical efficacy of rTMS is typically assessed using standardized measures such as the Clinician-Administered PTSD Scale (CAPS) or PTSD Checklist (PCL-5). Multiple clinical studies demonstrate that rTMS treatment yields significant improvements in both CAPS and PCL-5 scores, correlating with reduced PTSD symptom severity [68,79,83]. Integrating TMS with neuroimaging (TMS-fMRI or TMS-EEG) provides a powerful tool to investigate network activation states and predict clinical improvement through brain connectivity assessment. TMS modulates connectivity within PTSD-relevant circuits—including the mPFC, amygdala, and anterolateral temporal lobe—with these changes correlating with therapeutic outcomes. Excitatory stimulation of this circuit further potentiates fear extinction efficacy [84]. Therefore, HF-rTMS targeting the left dlPFC may offer therapeutic benefits for fear-related disorders by simultaneously enhancing extinction memory formation while suppressing original fear traces [79].

4. Innovating the TMS treatment for PTSD

4.1. Acceleration of treatment

Acceleration refers to delivering a higher number of TMS pulses within a single session or given treatment period. It has been demonstrated that daily treatment sessions can be replaced by two or three sessions per week for depression, with comparable clinical efficacy and tolerable side effects [85,86]. This approach was successfully replicated in high-dose theta-burst stimulation (TBS) protocols invented by Stanford scientists, which employ 10 daily sessions of intermittent TBS (iTBS) delivering 1800 pulses per day over five continuous days (totaling 18,000 pulses daily compared to conventional 3000-pulse protocols) [87,88]. These protocols have proven highly efficient for treatment-resistant depression and are therefore likely to be readily adopted and validated for PTSD treatment.

4.2. Uncovering population heterogeneity in treatment responses

PTSD is characterized by high clinical heterogeneity alongside the underlying neurobiological abnormalities across patients [89]. Significant efforts have been made to identify more homogeneous subgroups within PTSD to improve diagnosis and treatment. For instance, neuroimaging studies have identified subgroups based on functional connectivity abnormality in clinical PTSD populations [90], revealing associations with both clinical symptoms and neurocognitive differences. Notably, distinct resting-state connectivity patterns are well established in predicting rTMS treatment responses for depression [91]. Clarifying neural network connectivity and physiological responses to TMS protocols is therefore critical for predicting treatment outcomes.

4.3. Identification of peripheral biomarkers for protocol guidance

Peripheral biomarkers remain elusive during TMS treatment for PTSD, potentially reflecting symptom severity or predicting outcomes. Neuroendocrine factors, neuroinflammation signaling molecules, neurotrophic factors, and epigenetic modifications in peripheral mononuclear cells have been proposed as biomarker candidates in PTSD [[92], [93], [94], [95], [96]]. Previous studies report predictive value for inflammatory signaling and stress hormones in PTSD treatments (e.g. psychotherapy) [97,98]. Of note, rTMS studies in depression, chronic pain, and stroke have identified peripheral γ-aminobutyric acid (GABA), beta-endorphin, and brain-derived neurotrophic factor (BDNF) as potential biomarkers, respectively [99]. These biological markers require validation in clinical trials of TMS for treating PTSD.

4.4. Designing TMS therapy with individual precision

Beyond the neural network variables, precise TMS delivery requires spatial navigation based on sMRI imaging, refined TMS coil designs enabling better electromagnetic field penetration into the cortex, and improved cortical atlas “GPS” defining PTSD treatment’s most responsive “hot spot”. For instance, alleviation of distinct clinical symptom clusters in depression (e.g., anhedonia vs. anxiety) correlates with different TMS targets associated with independent brain circuits in fMRI connectome databases [100]. These possibilities are being elucidated through advances in human brain atlas with fine-grained parcellations and functional connectivity information—specifically the human brainnetome [101]. Recent neuroimaging and neurophysiological studies demonstrate default mode network, salience network, and cingulate cortex changes following rTMS treatment in PTSD patients [102]. Establishing a “PTSD brain connectome” with symptom-specific cortical targets for TMS stimulation sites remains crucial.

4.5. Information-based approaches

The stimulated cortical area comprises various neurons from distinct neural circuitries, and may or may not encode PTSD-associated symptoms. To improve the information specificity of rTMS protocol, one strategy involves applying rhythmic TMS that recruits neurons firing at specific frequencies, which are consequently implicated in encoding particular information (e.g., aversive memory) [103]. Another approach is to combine TMS treatment with behavioral interventions, including novelty exposure, memory retrieval, psychotherapy, or relevant cognitive tasks. These may either increase the specificity of TMS targeting or enhance treatment effects. For instance, cognitive training that elevates frontal theta oscillation prior to rTMS treatment has been shown to improve rTMS's antidepressant efficacy [104].

As mentioned above, the memory retrieval approach has been reported to be effective in PTSD treatment [80], which originates from animal study findings that concurrent memory retrieval and deep brain stimulation of the insula removes aversive memory [105]. However, potential discrepancies should be noted and elucidated in future studies: (1) the neuronal ensembles activated by PTSD memory retrieval may be distributed across different cortical areas, with varied importance of contribution to PTSD symptoms; (2) it will be important to understand the time interval precision between cue and TMS pulse. For instance, one study showed that precise timing (100 ms interval) of TMS pulse over the frontal cortex facilitated fear extinction, with only 28 pulses in total [106]; (3) it will be necessary to personalize the memory retrieval cue to maximize the reactivation effect, since there are huge individual differences in response to the same video content.

4.6. Close-loop stimulation

Accurate decoding of neural signatures associated with intrusive memories and depressive mood may enable close-loop brain stimulation. Recent studies utilizing chronic intracranial electrophysiology have established personalized symptom biomarkers and demonstrated their therapeutic utility for depression treatment [107]. This closed-loop paradigm could be clinically validated in PTSD patients receiving DBS implants. Furthermore, its implementation may extend to wearable TMS systems integrated with real-time EEG monitoring, enhancing both therapeutic precision and patient accessibility.

5. The neuromodulation effect of TMS on PTSD

5.1. TMS enhances neural plasticity of PTSD

Fear-and anxiety-related circuits in PTSD and other anxiety disorders show impairments in memory and dysfunctional synaptic plasticity [108]. Numerous studies have demonstrated the efficacy of TMS as an adjunctive or stand-alone treatment for PTSD, and other commonly used TMS methods, such as rTMS, paired associative stimulation (PAS), and TBS, can induce changes in cortical excitability and neuroplasticity. High-and low-frequency rTMS exert frequency-dependent and opposite effects on brain activity in local and distal regions [109]. High-frequency rTMS (HF-rTMS) stimulation (> 5 Hz) typically increases cortical excitability, whereas low-frequency rTMS (LF-rTMS) stimulation (≤ 1 Hz) decreases cortical excitability, and right-sided hemisphere TMS targeting the right dlPFC may be a better choice for PTSD treatment [69,110].

In addition, rTMS also affects morphological plasticity in the prefrontal cortex, such as increased spine densities of apical and basal dendrites in layers II/III and V of the mPFC in mice that received HF-rTMS (15 Hz) for 5 days [111]. Long-term changes in synaptic strength can occur after high-frequency stimulation. 5 Hz rTMS treatment of the dlPFC increased the synaptic plasticity-related proteins Synaptophysin and PSD-95 in aged rats, improved synaptic structural plasticity in the hippocampus, and activated the BDNF/ cAMP-response element binding protein (CREB) pathway [112]. In a single prolonged stress (SPS) rat model of PTSD, HF-rTMS (15 Hz) significantly increased N-methyl-d-aspartate receptor (NMDAR) and α-amino-3-hydroxy-5-methyl-4-isoxazole-propionic acid receptor (AMPAR) expression, significantly alleviated impaired glutamatergic receptors in the ACC, and remodeled neuronal synaptic plasticity through the phosphatase and tensin homolog deleted on chromosome 10 (PTEN)/Akt signaling pathway [113]. HF-rTMS enhances cognitive performance by modulating NMDAR-dependent brain plasticity. Healthy rats receiving 10Hz rTMS showed a significant increase in the expression of NMDA receptors NR1, NR2A, and NR2B in the PFC, hippocampus, and the M1, and activated BDNF/TrkB/Akt signaling pathway, as compared with pre-stimulation or control groups [114]. Cortical axon NMDARs are essential for triggering BDNF secretion and mediating TBS-induced long-term potentiation (LTP) at corticostriatal synapses [115]. Additionally, low-frequency pulsed magnetic field (LFPMF, 1 Hz) treatment can improve the expression of postsynaptic proteins, hippocampal synaptic plasticity, and cognition in depressive rats [116] (Fig. 3).

Fig. 3.

Mechanisms of rTMS treatment for PTSD. (a) Key rTMS protocols: Established techniques used in PTSD treatment include high-frequency rTMS (HF-rTMS), low-frequency rTMS (LF-rTMS), continuous theta burst stimulation (cTBS), and intermittent theta burst stimulation (iTBS) [110,119,150,151]. (b) Schematic illustrating how rTMS modulates PTSD pathophysiology. rTMS exerts therapeutic effects through multiple mechanisms: regulating cortical excitability, inducing synaptic plasticity (including LTP/LTD via NMDAR modulation), and enhancing neurotransmitter release. It additionally reduces astrocyte and microglia reactivity, exerts anti-inflammatory effects, and strengthens functional connectivity across neural networks.

Abbreviations: rTMS, repeated transcranial magnetic stimulation; HF-rTMS, high-frequency rTMS; LF-rTMS, low-frequency rTMS; cTBS, continuous theta burst stimulation; iTBS, Intermittent theta burst stimulation. LTP, long-term potentiation; LTD, long-term depression; NMDAR, N-methyl-d-aspartate receptor.

Patients with PTSD exhibited a significantly lower motor evoked potential (MEP) inhibition than healthy controls at 2 ms, 3 ms and 4 ms inter-stimulus intervals (ISIs). Paired-pulse transcranial magnetic stimulation (pTMS) of the motor cortex in PTSD patients showed a significant effect on the MEP amplitude at different ISIs, reflecting the selective vulnerability of cortical GABAergic interneurons to glutamate-mediated excitotoxic events [117]. Therefore, rTMS may improve the clinical symptoms of PTSD by inducing alterations in cortical inhibitory circuits.

TBS, such as iTBS and continuous TBS (cTBS), can rapidly induce synaptic plasticity, and TBS targeting the hippocampal network has beneficial effects on memory-related hippocampal neural activity and memory formation [118]. When TBS is applied to the motor cortex, it leads to corticospinal and corticocortical aftereffects, which may reflect LTP/long-term depression (LTD)-like synaptic effects [119]. iTBS affects the excitability of excitatory synaptic inputs to pyramidal neurons in both the stimulated hemisphere and the contralateral hemisphere [120], whereas cTBS may reduce the responsiveness of pyramidal neurons to excitatory stimuli [121].

5.2. TMS modulates the neural circuits of PTSD

PTSD is a stress-induced fear circuit disorder that involves dysregulation across a range of neural circuits mediating emotional processing, particularly those related to fear extinction, emotion regulation and memory processing [122,123]. PTSD primarily involves amygdala hyperreactivity to threatening stimuli, impaired prefrontal cortical regulation of fear extinction, and hippocampal memory dysfunction [124,125]. Impaired top-down control of the amygdala by the PFC and hippocampus in patients with PTSD results in abnormal fear extinction [124]. The amygdala, PFC, and hippocampus form the fronto-limbic neural circuitry, which plays a critical role in the acquisition, regulation, and extinction of fear memories in animals. These three brain regions are central to processing fear responses, with the amygdala primarily involved in fear learning and expression, the PFC in modulating and controlling fear, and the hippocampus in contextualizing and encoding fear-related memories.

The link between the amygdala and the mPFC is critical for fear conditioning and extinction [69,123]. The amygdala mediates the expression of conditioned fear and the enhancement of emotional memory, the vmPFC mediates the extinction of conditioned fear and the regulation of negative emotions, and the dlPFC plays an important role in executive function [126,127]. TMS can induce indirect changes in amygdala activity by directly stimulating functionally or structurally connected cortical sites. TMS applied to the vlPFC in 45 healthy participants revealed the most substantial reductions in TMS-induced fMRI signals within the basolateral amygdala, indicating that stimulation enhanced cortical-subcortical communication via the vlPFC-amygdala pathway [128]. In a randomized, sham-controlled clinical trial, 10 Hz rTMS applied to the left dlPFC in major depressive disorder (MDD) patients attenuated the inhibitory effects of the dlPFC on the amygdala [129].

Hippocampal structures mediate memory and learning deficits in PTSD. Repeated HF-rTMS treatment applied to a parietal location over 5 days in healthy adults demonstrated that long-term memory changes were associated with enhanced functional connectivity between cortical-hippocampal network regions, significantly improving associative memory [130]. TBS targeting the hippocampal network in the parietal cortex network also had a direct, beneficial effect on memory-related hippocampal neural activity [118]. PTSD is also associated with altered resting state functional connectivity (rsFC) involving the amygdala and hippocampus [131]. Therefore, increased connectivity between the amygdala-PFC and the hippocampus-PFC networks may serve as an important predictor of the therapeutic effect of TMS in PTSD patients (Fig. 3).

5.3. TMS modulates neural connectivity of PTSD

PTSD is associated with altered brain network connectivity, involving multiple brain regions in different functional and structural network patterns [70]. In addition to the activation of specific brain regions, studies of dysfunctional connectivity within and between circuits involved in mood disorders such as anxiety and depression suggest that dysfunction in large-scale circuits such as the DMN, SN, negative affect, positive affect (reward), attention and executive control network (ECN) is associated with mood disorders [132,133]. The ventral and dorsal attention networks (VAN/DAN) and areas related to the social network are also indirectly affected [134]. Compared to healthy controls, connectivity of the SN and ECN in the PTSD was enhanced [127], whereas connectivity in the DMN was reduced [135,136]. DMN intra-connectivity has been negatively correlated with PTSD severity and anxiety, and reduced DMN intra-connectivity with imbalanced DMN-SN and DMN—CEN connectivity may underlie impaired intrusive traumatic recall and situational autobiographical recall in PTSD [125].

5.4. TMS targets anti-inflammation in PTSD

Some individuals with PTSD exhibit increased levels of pro-inflammatory markers such as TNF-α, IL-1β, interleukin-6 (IL-6), and C-reactive protein (CRP), and decreased levels of anti-inflammatory markers such as interleukin-10 (IL-10) [137,138]. Studies have observed dysregulation of immune cells, human leukocyte antigens, and immune-related genes in PTSD [139]. Moreover, growing evidence indicates that inflammatory processes contribute significantly to PTSD pathogenesis [140,141]. Notably, rTMS demonstrates beneficial anti-inflammatory effects [142]. rTMS may indirectly modulate inflammatory and immune responses [143]. rTMS treatment resulted in significant downregulation of pro-inflammatory mRNA expression in both the infralimbic cortex and vCA1 [144]. Low-intensity rTMS decreased inflammatory pathway gene expression in cultured astrocytes [145]. Furthermore, high-frequency rTMS treatment reduced TNF-α levels while elevating IL-10, suppressed astrocyte A1 marker expression and classical activation, and attenuated microglial reactivity [146,147] (Fig. 3). These findings suggest that TMS may serve as a non-pharmacological anti-inflammatory intervention for PTSD.

Application of TMS near the cortical surface induces changes distal to the stimulation site. fMRI analyses suggest that altered patterns of intra-and inter-network connectivity play an important role in the clinical improvement of PTSD treated with TMS [73]. TMS enhances neuroplasticity in PTSD patients by increasing the function of prefrontal and subcortical/limbic structures and related networks [148]. Assessing TMS effects and changes in brain network connectivity patterns may aid in PTSD diagnosis and treatment monitoring. Several studies have shown that the DMN may serve as an important predictor of TMS treatment efficacy [149]. Clinical improvement in PTSD has been shown to be associated with enhanced connectivity within the DMN and increased negative cross-network connectivity (greater negative connectivity between DMN and ECN), along with significant improvement in social and occupational functioning [73]. In a prospective open study, stimulation of the left dlPFC with up to 40 daily sessions of 5 Hz TMS revealed that reductions in PTSD and MDD comorbidity symptoms were associated with reduced connectivity between subgenual ACC (sgACC) and DMN, between dlPFC and insula, and between hippocampus and SN [102]. After 4 weeks of HF-rTMS (10 Hz) treatment, spatial memory learning and cognitive abilities were significantly improved in rats receiving rTMS, and several brain regions within the interoceptive/DMN and cortico-striatal-thalamic network showed an increased regional homogeneity (ReHo) compared to the pre-TMS or control groups [114].

Available evidence suggests that TMS may enhance neuroplasticity, inhibit overactivation of the amygdala, improve functional connectivity within the DMN, facilitate extinction of fear memories, and improve clinical symptoms of PTSD. TMS represents a promising non-invasive treatment option to improve the quality of life of PTSD patients who do not respond well to psychotherapy and pharmacological treatments.

6. Conclusion

TMS, particularly rTMS, has emerged as a promising non-invasive neuromodulation technique capable of inducing neuroplasticity and restoring disrupted connectivity in PTSD-related brain networks. Innovations such as accelerated protocols, personalized stimulation targets, closed-loop systems, and biomarker-guided approaches may further enhance treatment precision and efficacy. Additionally, TMS demonstrates anti-inflammatory effects, which may contribute to symptom improvement among PTSD patients with immune dysregulation.

Future research should focus on refining stimulation protocols, identifying predictive biomarkers, and integrating TMS with behavioral interventions to maximize therapeutic outcomes. As a safe and non-invasive tool, TMS holds significant potential as a standalone or adjunctive therapy for PTSD, providing a hopeful alternative for patients with treatment-resistant symptoms.

CRediT authorship contribution statement

Ti-Fei Yuan: Writing – review & editing, Funding acquisition, Conceptualization. Ting Wang: Writing – review & editing, Writing – original draft. Shuoshuo Li: Writing – review & editing, Writing – original draft. Pengfei Wei: Writing – review & editing. Liping Wang: Writing – review & editing. Bingxing Pan: Writing – review & editing. Ji Hu: Writing – review & editing. Shengxi Wu: Writing – review & editing. Yuan Shen: Writing – review & editing. Zengqiang Yuan: Writing – review & editing. Haitao Wu: Writing – review & editing, Validation, Supervision, Funding acquisition, Conceptualization.

Biography

Dr. Wu Haitao, (BRID: 09631.00.96228) is a professor and Director of Beijing Institute of Basic Medical Sciences. He graduated from Fudan University and received his PhD degree from Academy of Military Medical Sciences. He finished his postdoc training at the Medical College of Georgia. He is a recipient of the National Science Fund for Distinguished Young Scholars and the National Science Fund for Excellent Young Scholars, a distinguished guest researcher at the Chinese Institute for Brain Research (CIBR) and a ``Beijing Brain Scholar''. His lab has long been dedicated to studying the mechanisms underlying cortical development, social behavior, and stress-induced cognitive dysfunctions. As the principal investigator, he has led more than 20 projects funded and published over 50 papers and reviews as a corresponding author in journals such as Neuron, PNAS, Science Advances, Mol. Psychiatry, Cell Death Differ., Advanced Science, Cell Reports (cover article) etc., and has obtained 20 national invention patents and software copyrights. He currently serves as the executive committee member of the Chinese Neuroscience Society and the chair of its Neurodevelopment and Regeneration Branch.