While the reward system primarily motivates the initial drug use by reinforcing pleasure, the craving/relapse circuitry sustains SUD by driving compulsive drug-seeking behavior. Changes in memory and emotional processing systems increase vulnerability to relapse. Although the Nucleus Accumbens (NAc) is the central hub of the two systems, the Amygdala (AMY) regulates emotional processing, and the Hippocampus (HP) functions in memory processing (30, 31). The Prefrontal cortex (PFC), on the other hand, regulates the function of the three regions (32, 33).

The NAc, divided into core and shell regions, receives inputs from the ventral tegmental area (VTA), AMY, and HP and sends outputs to the PFC (34–36). The core differs from the shell’s inputs, with the former receiving from the central AMY nucleus and PFC and the latter from the lateral hypothalamus (LH) (37–42). Both regions are involved in reward behaviors, but the core shows drug-specific adaptations, suggesting potential for SUD treatments, though further research is needed (43–45). Notably, the dorsomedial PFC (dmPFC), a crucial regulator of cravings and drug-seeking behavior, can become compromised in SUD (46, 47). Within the dmPFC, the prelimbic (PLC) cortex is more implicated in relapse susceptibility, while the infralimbic (ILC) cortex potentially aids in suppressing drug-seeking behavior (46, 47). This region’s activity evolves during withdrawal, initially showing hypoactivity in early withdrawal but hyperactivity in late withdrawal (48–51). Similarly, the amygdala undergoes structural changes and adaptations from chronic drug use, amplifying circuits linked to craving and relapse (26, 52). Imbalances in neurotransmitters within the amygdala can significantly alter how reward signals are processed, thus contributing to addictive behaviors (53–57). Different subdivisions of the AMY respond differently to drug versus natural rewards. While the basolateral AMY (BLA) is involved in both, the central AMY (CeA) primarily mediates drug-reward-driven behaviors (55–57). While the HP is not typically a primary target in treating SUD, disruption in HP theta oscillations could occur in SUD, leading to imbalances in neurotransmission in connected brain regions, including the AMY, HP, and PFC.

Addictive drugs, such as psychostimulants, opioids, and alcohol, change neurotransmission in both craving and reward systems, albeit with varying mechanisms and effects (58–61). For example, exposure to psychostimulants can lead to dysfunction in the glutamatergic projections between the dmPFC and NAc while affecting NAc dopamine by dopamine transporter blockage, eventually leading to drug-seeking reinforcement (62–65). Conversely, the amygdala’s response to psychostimulants is expressed by surges in dopamine activity (66–68). Opioids engage opioid receptors, inducing euphoria, pain relief, and dopamine release while affecting glutamate and dopamine transmission (69). Alcohol increases BLA glutamate, CeA GABA, and NAc core dopamine, fostering intoxication, reward perception, emotional reactions, and impulsivity (70–72). Chronic psychostimulants and opioid use can heighten craving, desensitize natural reward responses, and induce enduring neurological alterations, leading to adaptations, tolerance, and a shift from positive to negative reinforcement effects (73–75). In the same way, prolonged alcohol use can disrupt PFC glutamate signaling, impairing cognitive faculties and decision-making (76). Therefore, neuromodulation of the reward and craving/relapse systems is crucial for developing targeted interventions for disrupting SUD-related cycles and promoting recovery.

The Insular cortex (IC), a vital component of the interoceptive system, works alongside the ACC in processing internal bodily signals and drug cravings (77). Functionally divided into ventral, dorsal anterior, and posterior regions, the IC retrieves memories of drug effects stored in the ventral anterior IC (AIC) during drug craving, triggering the posterior IC (PIC) to process physical sensations associated with the drug (78–80). Collaborating with the HP, PFC, and AMY, the IC amplifies emotions during drug withdrawal, with PFC activity varying based on the withdrawal stage (78–81). This collaboration intensifies the conscious experience of withdrawal symptoms.

The IC’s adaptation to drug versus natural rewards is complex, involving local communication within the ventral anterior IC to regulate activity and balance glutamate influence (79). Further research is needed to compare IC adaptation to drugs versus natural rewards. In the PFC, ACC processes information, integrating physiological signals about the body’s internal state and assigns salience to these signals, influencing perception and behavior (82). The ACC, interacting with the IC, modulates attention, decision-making, and emotional regulation, driving motivation to seek relief from drug use (83, 84). However, drug memories can override its guidance, leading to relapse (80, 85).

Psychostimulants, opioids, and alcohol disrupt the interoceptive system’s delicate equilibrium via neurotransmitter modulation, obscure overdose symptoms, and foster hazardous behaviors (86). Chronic substance use can precipitate enduring modifications in interoceptive processing, reinforcing drug use as a coping mechanism (87). This intricate network interaction, termed allostasis dysregulation, intensifies craving and compulsive drug use in individuals with SUD, often manifesting as withdrawal symptoms and recurrent substance-seeking behavior (88, 89). Thus, the interoceptive system is an important target to disrupt the craving-compulsion cycle and address SUD effectively.

Numerous techniques have been developed to treat psychological disorders by facilitating neuromodulation (90). Traditional methods like TMS and TDCS have provided valuable insights into noninvasive strategies for altering brain function and treating disease. Treatments with these methods are relatively painless, do not require surgery, and can be combined with other therapies (20). However, their effects can vary, and they have limited spatial resolution to target specific brain regions (20, 21). Emerging technologies, such as auricular nerve stimulation and near-infrared optogenetic stimulation (NIR), address some of these challenges but face their limitations (see Table 1 for a summary of the advantages and limitations of various neuromodulatory technologies).

LIFU emerges as a leading neuromodulation technique due to its precision, safety, and versatility. It boasts millimeter precision, facilitating targeted modulation of specific neural populations within the brain and minimizing unintended impacts on surrounding tissues (91, 92). LIFU’s mechanism of action involves direct interaction with neuronal tissue through mechanical energy, which allows deeper tissue penetration than techniques using electrical and magnetic field interactions (93, 94). Thus, LIFU can access deeper brain structures, including subcortical nuclei, making it valuable for addressing conditions like SUD (95, 96).

LIFU has a good safety profile when operated within minimal and nonsignificant risk recommendations of the International Transcranial Ultrasonic Stimulation Safety and Standards Consortium (ITRUSST) (97). LIFU operates at frequencies that prevent heat generation, mitigating the risk of magnetothermal lesions associated with high-intensity magnetic fields used in specific TMS and NEMS protocols (20, 98). LIFU also reduces potential complications like immune responses to nanoparticles, which may arise in NEMS technology (98). This favorable safety profile makes it a more biocompatible and patient-friendly neuromodulation technique. While NIR optogenetic stimulation offers high spatial resolution, its application requires invasive genetic modification of targeted cells, limiting its versatility (99). In contrast, LIFU presents a noninvasive approach that does not require genetic manipulation and nanoparticle injection, simplifying its application and broadening its suitability across a diverse spectrum of neurological disorders.

One of the unique and critical advantages of LIFU is its capability for bidirectional modulation of neural activity. LIFU can reversibly inhibit or excite neuronal circuits depending on the parameters applied (92). The flexibility offered by LIFU is particularly relevant in treating SUD. In addition, specific circuits may be hyperactive (during late withdrawal) or underactive (during early withdrawal). Adjusting LIFU’s parameters makes it possible to tailor the treatment to the individual’s neurophysiological SUD profile.

Neurons are viscoelastic materials capable of propagating mechanical energy and storing it elastically (100). Mechanical interactions have led to the proposal of several mechanisms of action for LIFU, including mechanosensation, electrophysiological-mechanical coupling, microtubule resonance, thermal mechanism, and cavitation.

Mechanosensation converts mechanical energy into neural signals through mechanosensitive ion channels. LIFU is hypothesized to physically displace and activate these channels, leading to changes in ion transport, neuronal depolarization, and altered neural signaling (101–103). LIFU has been shown to interact with calcium-permeable mechanosensitive channels like TRPP1/2, TRPC1, TRPA1, and Piezo1 (101–104). Additionally, mechanosensitive ion channels from the two-pore-domain potassium channel family (e.g., TREK-1 and TRAAK) have been shown to respond to LIFU (100, 105). LIFU can also activate touch sensation MEC-4 channels (including DEG/ENaC/ASIC ion channel), large conductance mechanosensitive channel MscL, and sodium ion channels (106). Cytoskeleton also plays a part in the LIFU mechanosensation mechanism. Duque et al. recorded calcium influx and membrane currents in hsTRPA1-expressing cells of rats and mice, likely due to the interaction of the sonication-sensitive N-terminal tip domain of hsTRPA1 with the actin cytoskeleton (107). In another study, acoustic pressure waves were generated when LIFU traveled through the extracellular matrix, ultimately activating ASIC1a in a cytoskeleton-dependent manner (102). This activation likely occurs in concert with the simultaneous activation of other mechanoreceptors, suggesting a complex interplay within the mechanosensitive system underlying LIFU’s actions. However, only a few mechanosensitive ion channels have been studied for LIFU’s effects.

Beyond individual channels, LIFU modulates broader mechanosensitive machinery, leading to changes in mechanically coupled electrophysiological signals. These signals are associated with changes in membrane conformational state changes and mechanosensitive ion channels. Changes in membrane conformational states involve mechanical signals influenced by surface tension, elasticity, and intracellular fluid viscosity (108). These conformational changes can be externally induced via LIFU’s mechanical energy, altering membrane fluidity and permeability (109). This high-energy state causes embedded proteins and lipids to adapt, changing the membrane’s capacitance and modulating neural activity. Mechanical deformations redistribute dipoles in lipid bilayers in neuronal membranes, causing surface polarization- a process termed direct flexoelectricity (DF) (110). LIFU’s mechanical energy provides a possible membrane deformation, which could lead to DF. Exogenous LIFU pulsation could also interfere with these native thermodynamic waves generated by lipid phase transition by transferring acoustic energy and generating pressure waves, which could alter action potentials depending on the neuron’s initial state and orientation (111, 112).

LIFU stimulation can also cause cavitation, where gas bubbles within tissues resonate, expand, and collapse depending on the frequency, creating mechanical effects. Though bubbles are generally negligible in the nervous system, micro cavitation can increase membrane permeability via sonoporation, creating pores in the lipid bilayer. At the same time, mechanosensitive channels could be activated during micro cavitation. LIFU combined with ultrasound microbubble contrast agents could have controllable cavitation effects. However, clear models for this application are still being developed. The neuronal intramembrane cavitation excitation (NICE) models predicts cell-type-specific responses that correlate indirectly with experimental data, and the SONIC model addressed the computational speed limitation of the NICE model (113–115). Both models describe how US-induced cavitation can modulate neuronal activity. However, many studies obtain results that do not follow the predictions of these models partly because LIFU stimulation does not always target singular cell types at a time.

So far, the effects of LIFU have mostly been routed through mechanical interactions. In addition to the earlier described mechanisms, Hameroff et al. propose that the LIFU in specific megahertz frequency bands can resonate with microtubules, causing them to vibrate when aligned with their long axis (116). Even with these microtubule vibrations, electrophysiological-mechanical coupling, and cavitation are likely. Some studies have proposed thermal mechanisms. For example, Darrow et al. suggest a 2°C increase might contribute to neuro inhibition under specific conditions (117). However, unlike HIFU, which utilizes high intensities for tissue ablation, direct LIFU stimulations typically induce negligible temperature increases (<1°C), which is generally considered insufficient for direct neuronal modulation (117, 118). Therefore, thermal effects are unlikely to be the primary driver, but thermal modeling and reported thermal indices remain valuable in accounting for variations in sonication parameters, tissue properties, and beam configurations during treatment optimization. While LIFU has been used for various successful applications of neuromodulation, much work remains to be done to understand the complex interactions that account for LIFU’s action mechanisms with varying parameters.

Ultrasound is applied using transducers with specific fundamental frequencies or acoustic frequencies to target brain regions, and neuromodulation with LIFU relies on specific frequencies for precise targeting of brain regions. While LIFU studies have used f ₀ ranging from 200 kHz to 10 MHz ( Figure 1 ), compared to higher f ₀ for medical imaging (around 15 MHz), skull properties present a significant challenge. The acoustic pressure weakens through the skull due to the conversion of sound waves, absorption by bone, reflection, and scattering. Standing waves can also form on the other side of the skull, potentially causing unintended sonication effects. Acoustic pressure attenuation is particularly problematic when using higher frequencies, which offer excellent spatial resolution but penetrate less deeply. In contrast, while a low fundamental frequency is favorable for deeper targets, excessively low values make ultrasound waves travel deeper than intended, potentially reaching unintended brain regions. Therefore, the fundamental frequency is in the middle range of frequencies to allow sufficient penetration to reach the target region, and avoiding these complications is vital to successful LIFU applications.

The pulse repetition frequency (PRF) is the rate at which ultrasound pulses are delivered. According to ITRUSST, the PRF could be reported as either the individual pulses in a pulse train (PRF_{pulse train}) or the frequency at which the pulse train is repeated (PRF_{pulse train repeat}) (119). Both PRF_{pulse train} and PRF_{pulse train repeat} are crucial parameters in focused ultrasound, as they define the pattern of every ultrasound waveform. While PRF_{pulse train} influences the observed effect of LIFU, PRF_{pulse train repeat} is usually employed to avoid excessive heat buildup. However, repeated stimuli separated by fixed times can initiate Long-Term Potentiation (LTP) and long-term memory (LTM) encoding (120, 121). For instance, we have previously used a 0.2 Hz PRF_{pulse train repeat} in the downstream inhibition of dopamine to avoid heating with a 10 MHz f _c transducer (122). A theta burst protocol was also developed by Zeng et al. to induce consistent corticospinal excitability (123).

While the exact influences of PRFs are still under investigation, a growing body of research suggests that PRF is more than an on/off switch for inhibition or excitation. Stimulation with 250 kHz and 1 kHz PRF_{pulse train} can suppress NAc activity, suggesting a wide range of possible inhibitory PRF_{pulse train} (124, 125). In recent studies, the He group used 3 kHz to depress field excitatory postsynaptic potentials in rats HP, Kim et al. utilized a 3 kHz PRF_{pulse train} to reduce heat pain sensitivity by targeting human IC, and the Legon group achieved similar effects using a 1 kHz PRF_{pulse train} (126–129). However, Niu et al. showed excitation of GluA1 expression alongside inhibition of GluA2 and GluA3 in mice NAc with 1 kHz (130). Therefore, predicting inhibitory vs excitatory mechanisms is challenging.

In the PFC, the complexity increases. Our group found that applying a 1 kHz PRF_{pulse train} to the PFC led to downstream dopamine inhibition in the NAc, and 1.5 kHz targeting the PFC has been used to inhibit inflammation in rodents (122, 131). PRF_{pulse train} lower than 1 kHz have also been used in neuromodulation. Yi et al. showed downregulation of inflammation markers with 100 PRF_{pulse train} (132). Huang et al. showed excitation of postsynaptic current and GluN2A expression with 500 Hz, Ren et al. used 200 Hz to reverse the expression of depression-related genes in the PFC, and the Lee group showed excitation at 140 Hz and inhibition at 10 Hz PRF_{pulse train} (133–135). Similarly, Pan et al. used a PRF_{pulse train} of 1 Hz to increase neuronal activity in mPFC cells, and Xie et al. showed an increased firing rate upon LIFU sonication at the same PRF_{pulse train} (136, 137). Chou et al. and Kuhn et al. used a 10 Hz PRF to inhibit functional connectivity in the amygdala (138, 139). However, a PRF_{pulse train} of 1 kHz produced a similar result in a study by Folloni et al. (140) From the distributions depicted in Figure 1 , the PRF_{pulse train} from both excitation and inhibition span similar ranges. The same trend is observed for the PRF_{pulse train repeat}. These findings suggest that PRF is not the sole determinant of inhibition or excitation in LIFU applications.

It is crucial to consider that each brain region in every organism has a natural local oscillation frequency (141). However, it is noteworthy that apart from the studies by Chou et al. and Kuhn et al., which used 10 Hz, and Xie et al., and Pan et al., which used 1 Hz, the PRFs tested to date predominantly fall within the high gamma domain or high-frequency oscillation domains (128, 133, 137). This observation raises the question of how LIFU might interact with other frequency domains, such as delta, theta, alpha, and beta, each known to correlate with specific tasks (142, 143). EEG spectral measurements indicate that diseases and behavioral changes may shift native oscillation frequencies (144–146). Consequently, the susceptibility of different brain areas to specific PRF ranges could be dictated by the constructive or destructive interference of LIFU waves with the specific brain waves of a particular region (141). This complex interplay of factors underscores the need for further research to understand the full implications of PRF in LIFU applications. Future research should delve deeper into the interplay between pulse repetition frequency (PRF), other parameters, and brain region, to understand its neuromodulatory effects to unlock the full potential of LIFU therapy.

There are three different layers of duration and duty cycle in the LIFU parameter selection. The pulse/pulse train/total sonication duration (PD/PTD/TSD) refers to the time-on duration of the pulse/pulse train/sonication waveform. At the same time, the duty cycle is the percentage ratio between pulse/burst/sonication duration and the total pulse/burst/sonication period. Notably, while some studies employ both pulse train and pulse duration in parameter selection, others use either pulse or burst duration. There is a wide range of PD and PTD, from microseconds (5 µs) to milliseconds (360 ms). For instance, Mahoney et al. employed 100/900 ms (on/off) PD in humans, and Niu et al. used 5 μs pulse duration in mice (130, 147). There is a broader range of pulse train duration values (1 ms - 5 min). TSD also varies widely (40 s- 1 hour). Mahoney et al. applied LIFU for 10 minutes per hemisphere in humans, whereas Deveci et al. used a longer 30-minute duration in rats (124, 147). In the studies highlighted in Table 3 , excitatory or inhibitory effects are not necessarily favored by shorter or longer TSD. Kim et al. observed both excitation and inhibition with 20 min TSD, while Lin et al. found that 20 min of LIFU could lead to excitation in neuronal activity and reduction in drug-seeking (134, 148). However, the choice of TSD is crucial as it appears to determine the level of effects observed. For example, Deveci et al. observed reduced alcohol SUD-related gene expression from 454 gene expression changes after short-term stimulation to 382 gene expression changes after long-term stimulation (124). There may be a correlation between the effectiveness of long-term stimulation and the process of encoding long-term memories. This process usually occurs in a time scale of minutes (120, 121).

The duty cycle is a function of both duration and repetition frequency ( Table 2 ). As with other parameters, the duty cycle may have been chosen based on the targeted brain region, desired effects (excitation vs. inhibition), and safety considerations. Studies on humans focus on safety and tolerability, while studies on animals have more flexibility in exploring different stimulation parameters, including the duty cycle. The parameters might involve continuous or near-continuous LIFU for initial assessment, and pulsed LIFU with varying duty cycles. Lower duty cycles have been thought to cause inhibitory effects, while higher duty cycles lead to excitatory effects (92). A similar trend is observed in the studies by Kim et al. and Wang et al., where DC_{pulse train} of 70 and 60% cause excitation (134, 149). However, Kim et al. also observed excitatory effects with a DC_{pulse train} of 0.8% and a DC_{pulse train repeat} of 10%, an exception to this trend. Moreover, Figure 2 shows that more studies found inhibition at DC_overall greater than 10% compared to excitation. A similar trend is observed with the DC_{pulse train repeat}. Therefore, there is no conclusive rule for duty cycle selection. As with the PRF, each brain region could possess a natural duty cycle for wave oscillations, and selected LIFU parameters could lead to constructive or destructive interference, leading to the neuromodulatory effect. However, further research is needed to establish these relationships and optimize LIFU protocols for targeted modulation of neural circuits in SUD.

Intensity is a critical parameter influencing LIFU neuromodulatory effects. It is typically quantified in units of either Isppa (spatial peak intensity averaged) or Ispta (spatial peak temporal averaged). While some studies report acoustic pressure instead of intensity, and I_sppa and I_spta are not always differentiated, they convey the same concept. Studies investigating direct NAc stimulation with 4.06 W/cm² I_sppa [0.305 W/cm² I_spta] and acoustic pressures of 590 kPa have demonstrated suppression of NAc gene expression and activity (124, 125, 130). Conversely, 304 KPa targeted at the NAc showed simultaneous inhibition and excitation of different subunits of AMPA receptors, ultimately allowing ion flow.

The effects on the PFC are more nuanced. Different PFC regions stimulated with LIFU can exhibit direct and downstream excitatory or inhibitory effects. Our previous work demonstrated effective downstream inhibition of NAc dopamine release via PFC sonication at 13 W/cm² (122). Wang et al. reported inhibitory effects of LIFU with at 500 mW/cm², Pan et al. reported upregulation of the NR1 NMDA receptor and cFOS downregulation in the mPFC at 500 mW/cm², while Kim et al. observed both inhibitory and excitatory effects on the PFC at 3 W/cm² (131, 134, 136). Notably, several studies have achieved excitatory effects on the PFC using various intensities (3 W/cm² I_sppa, 20 kPa, and 328 kPa) (131, 134, 148). For the insular cortex, stimulation with intensities of 3.5 W/cm², 208.46 mW/cm², and 300 kPa resulted in inhibition, while 8.66 and 0.35 W/cm² yielded excitation (126–128). Additionally, Chou et al. employed 14.4 W/cm² I_sppa (0.72 W/cm² I_spta) to decrease Amygdala-IC functional connectivity, while the Legon group used 3.5 W/cm² I_sppa to attenuate pain ratings in the insula (127, 138).

Many specific studies find effects of intensity, but trends across studies are difficult to elucidate. Figure 3 shows a wide range of distribution of I_sppa variation between 0- 90 W/cm² across studies. I_spta applied is generally lower than 1 W/cm² partly due to FDA regulations. Consequently, safety becomes the primary concern regarding LIFU intensity for SUD treatment. Higher intensities are generally associated with higher tissue heating, but the exact relationship between intensity and therapeutic effect remains under investigation. Future research should prioritize establishing optimal intensity thresholds within brain regions relevant to SUD.

Exploring the frontiers of SUD treatment, LIFU emerges as a promising tool, demonstrating significant potential in mitigating addictive behaviors as evidenced by recent scientific studies. LIFU’s reduction of craving has been shown in both animals and humans. Lin et al. highlighted LIFU’s potential in reducing methamphetamine (MA)-seeking behaviors in rats in a conditioned place preference (CPP) test model (148). Notably, LIFU applied to the infralimbic cortex was most effective in reducing MA-seeking behaviors induced by MA priming, with no observed impact on MA-induced locomotor activity. This observation suggests a specific effect of LIFU on craving behavior rather than general locomotion. Mahoney et al. explored the potential of LIFU neuromodulation for treating SUDs by targeting the bilateral NAc ( Figure 4 ) (147, 151). A 43-year-old participant underwent both sham and active LIFU sonication, reporting an immediate reduction in cravings for primary substances. Cue-induced craving for alcohol and illicit substances was utterly suppressed during post-LIFU follow-up assessments up to 90 days, alongside the elimination of substance-related dreams and improvements in anxiety and overall functioning. Urine toxicology analyses confirmed abstinence throughout the follow-up. In a follow-up study, four participants with opioid use disorder underwent active LIFU targeting the NAc, showing reduced craving during and after treatment, with lasting reductions observed during a 90-day follow-up. These findings suggest that LIFU targeting the NAc is a safe and effective intervention for reducing substance craving and improving outcomes in SUDs. However, more extensive and randomized trials are needed to validate these findings and fully assess the impact of LIFU on substance use and relapse, potentially offering a noninvasive adjunctive therapy for SUDs with the potential to reduce cravings and improve treatment outcomes.

Addiction disrupts communication pathways between reward circuits and areas governing decision-making and impulse control, exacerbating the challenge of resisting cravings as reward circuits become hyperactive while decision-making areas become hypoactive. However, the brain’s inherent neuroplasticity presents both a hurdle and an opportunity in this context. While SUD rewires neural circuits, treatment interventions have the potential to restore them to a healthy state. Notably, studies indicate that LIFU modulates neuronal activity and influences responsiveness to other neural inputs, thereby facilitating observed alterations in activity coupling. For instance, Folloni et al. examined the impact of LIFU on neural activity in the human amygdala, revealing reduced activity coupling between the stimulated area and its interconnected regions (140). Wang et al. also showed improved theta oscillation synchronization after LIFU stimulation of the mPFC (149). Peng et al. used functional magnetic resonance imaging (fMRI) to investigate the effects of LIFU on the reward network in ten healthy adults (152). Bilateral NAc inhibition was observed during LIFU on the left NAc compared to sham, with increased functional connectivity between the NAc and medial PFC (mPFC). These findings suggest that LIFU is a promising tool for the direct and noninvasive modulation of the NAc, shedding new light on treating SUD and other brain diseases involving reward processing.

Many individuals with SUD initially turn to substances to cope with introceptive disorders, such as hypersensitivity to pain, social anxiety, fear, and depression. Addictive drugs can provide temporary relief from these interoceptive responses, leading to a cycle of self-medication and potential SUD. However, introceptive dysfunctions of the IC and ACC following chronic drug use amplify these responses. Therefore, even when such individuals withdraw from drug use, underlying emotional problems often lead to relapse (27). LIFU has exhibited promise in regulating anxiety, fear, and depression. In a study by Chou et al., Blood Oxygen Level-Dependent (BOLD) activation in several regions of the fear network, including the AMY, HP, and dorsal ACC, was altered by LIFU, suggesting the reorganization of functional connectivity of fear-related brain activity (138). Yi et al. found that LIFU ameliorates depressive-like behaviors and anxiety-like behaviors while inhibiting proinflammatory cytokine upregulation in the PFC (132). Similarly, Wang et al. observed amelioration of depression-like behaviors through improvements in theta oscillation synchronization and synaptic functional plasticity within the vCA1 - mPFC pathway following LIFU treatment (149). Wynn et al. demonstrated LIFU’s effectiveness in treating pain by applying LIFU to both anterior and posterior IC, attenuating contact heat evoked potentials (CHEP) peak-to-peak amplitudes, indicating a reduction in neural responses to the painful stimulus (127). Additionally, In et al. showed that LIFU to the PIC significantly attenuated pain ratings in both temporal summation of pain (TSP) and the conditioned pain modulation (CPM) protocols, while LIFU to AIC did not affect either TSP or CPM pain ratings (126). Wang et al. also demonstrated the alleviation of chronic allodynia pain by ACC inhibition, potentially offering a novel approach to managing pain hypersensitivity induced by opioid misuse (153).

LIFU holds the potential to recalibrate disrupted neural pathways and instigate molecular changes controlling synaptic plasticity. In a recent study by Kuhn et al., LIFU was used in conjunction with ASL MRI and simultaneous BOLD fMRI to investigate its impact on deep brain regions in humans (139). The results demonstrated that LIFU could selectively enhance regional blood flow and modulate network connectivity of subcortical regions in a targeted manner. Interestingly, tFUS parameters aimed at disrupting activity led to decreased functional connectivity in the amygdala network while increasing BOLD activity and functional connectivity in the targeted ErC and its network. Both LIFU protocols led to increased perfusion exclusively in the targeted brain region, highlighting the focal and region-specific modulatory effects of tFUS. Thus, LIFU selectively enhances regional perfusion while modulating regional activity and connectivity, opening up potential clinical applications for emotion regulation and memory function.

Substance use disorder, particularly involving fentanyl, can worsen neurodegenerative processes. Recent research by Lim et al. suggests that LIFU could promote neurogenesis in the brain, potentially aiding in treating neurodegeneration linked to SUD (102). However, additional research is needed to fully understand the extent to which LIFU can induce neurogenesis in the brain regions most impacted by SUD. Nonetheless, preclinical studies show that LIFU, a noninvasive neuromodulation technique, holds considerable promise for addressing SUD-related behavioral issues. While further research is necessary to determine its efficacy and safety in human populations, LIFU presents a promising new approach for developing novel therapeutic interventions.

The precise mechanisms underlying the modulation of neural circuits by LIFU warrant exploration, alongside elucidation of how such changes manifest in downstream behavioral effects. Prior investigations have focused on LIFU’s influence on calcium signaling and mechanosensitive ion channel expressions at the cellular and molecular levels (101, 102). However, LIFU modulation entails a complex interplay of neurophysiological processes beyond these aspects. Comprehensive mechanistic and behavioral studies are imperative to better understand the criteria for LIFU parameter selection. For example, studies should explore the effect of LIFU on upstream and downstream neurotransmission in SUD-related circuits. Methods to measure cell firing and real-time changes in neurotransmission are essential to comprehend the neurotransmission basis for observed behavioral effects of LIFU.

Moreover, mechanistic studies aimed at elucidating the underlying neural mechanisms of LIFU-induced neuromodulation in SUD-related behaviors should include the pre- and post-synaptic terminals, surrounding astrocytes, and presynaptic extracellular matrix between these elements. These studies may involve examining changes in neural activity, protein expression analysis, and real-time neurotransmitter changes following LIFU treatment. Understanding the precise molecular mechanisms and neural circuits involved in LIFU-mediated effects can inform targeted interventions and optimize treatment outcomes.

Once LIFU’s effect on neurotransmission and downstream molecular signaling cascades is established, further behavioral studies can assess the impact of LIFU treatment on SUD-related behaviors using preclinical animal models and clinical trials. Preclinical studies in rodents, mainly focusing on the NAc, have reported promising results, with LIFU application reducing drug-seeking behavior in models of cocaine, alcohol, and opioid dependence. However, further investigations on other abused substances, including fentanyl, are warranted. By targeting additional brain regions and investigating SUD in various addictive drugs, LIFU may offer comprehensive interventions for addressing multiple aspects of SUD pathology.

Utilizing low-energy sound waves, LIFU provides a noninvasive and targeted treatment modality when combined with microbubbles, tiny gas-filled spheres containing drugs that are introduced into the bloodstream. LIFU can be multiplexed with pharmacological treatments and could be used for targeted drug delivery for SUD and relapse treatments. This innovative approach may improve the delivery of drugs across the blood-brain barrier, a significant challenge in treating neurological conditions. By amplifying the effects of medications designed to alleviate SUD symptoms, such as cravings and withdrawal, LIFU holds promise as a potent therapeutic tool. Clinical trials, such as those conducted by InSightec, are currently exploring LIFU’s safety and efficacy as an adjunctive treatment for Opioid Use Disorder (OUD), potentially paving the way for broader applications in SUD treatment (154, 155).

Integrating LIFU with cognitive-behavioral therapy (CBT) and other behavioral interventions could modulate neural circuits that underlie addictive behaviors. Although this approach has only been explored in methods other than LIFU, combining its advantages over other neurostimulation methods with CBT could be revolutionary (156). This synergy could enhance the brain’s receptivity to therapy, potentially leading to more durable treatment outcomes and reduced relapse rates. This method of drug delivery facilitates delivery to target locations and overcomes blood-brain barrier permeability. This combination holds immense potential across various medical domains, including neurological disorders, cancer treatment, and gene therapy. By enabling targeted drug delivery while potentially reducing side effects, this approach offers a promising avenue for improved treatment paradigms. While still in the early stages of development, ongoing research aims to validate the efficacy of this approach through clinical trials, paving the way for its widespread adoption and transformative impact on patient care. The precise mechanisms by which LIFU affects neural activity and behavior are still under investigation, necessitating further research to optimize its use in conjunction with behavioral therapies.

Future research directions include pairing LIFU with real-time brain monitoring techniques, such as fMRI, to tailor treatments to individual neural activity patterns. This personalized approach could allow for the adjustment of LIFU parameters during treatment, potentially improving efficacy and patient outcomes. Studies investigating the integration of LIFU with monitoring technologies are essential to develop protocols that can adapt to the dynamic nature of brain activity during SUD treatment. Integrating LIFU into the multi-modal treatment landscape for SUD holds excellent promise. By combining this innovative technique with established pharmacological and behavioral interventions and harnessing the power of real-time brain imaging, researchers aim to create more effective and personalized treatment strategies. As evidence grows, LIFU could become a key component in the comprehensive treatment of SUDs, offering new hope to those affected. Rigorous clinical trials and continued scientific inquiry are vital to unlocking LIFU’s full potential and ensuring its safe and effective application in clinical settings (157).