Behavioral addictions (BA), characterized by excessive and compulsive engagement in non-substance-related activities (Grant et al., 2010; Robbins & Clark, 2015), have garnered increasing attention in the field of psychiatry due to the increasing prevalence and changing epidemiology of BA (Abbott, 2020) and the consequent interference with the quality of life of the patients (Potenza et al., 2019). Gambling disorder (GD), a prototypical BA, shares striking parallels with substance use disorders (SUDs) (Griffiths, 2017). Key shared features include impaired control over the behavior, craving, tolerance, withdrawal, and continued engagement despite adverse consequences (Alavi et al., 2012; Leeman & Potenza, 2012; Rash et al., 2016), all of which can lead to profound personal, social, and financial disruptions (Rodriguez-Monguio et al., 2018; Browne et al., 2017). Neuroscientific research has advanced our understanding of the neuronal substrates underpinning both behavioral and substance addictions (Murch & Clark, 2016; Paulus, 2022). In a Research Domain Criteria (RDoC) perspective, a framework for research on mental disorders that includes multiple dimensions with a strong focus on neurocircuitries (Insel et al., 2010), GD patients exhibit deficits in the domain of “positive valence systems,” particularly in the “approach motivation” and “reward learning” constructs, as well as in the “cognitive systems,” primarily in the “cognitive control” construct (Pallanti et al., 2021).

Among the neuronal regions implicated, the pre-supplementary motor area (pre-SMA) has emerged as a critical node in the neurocircuitry of addiction (Lohse et al., 2023; Kaufman et al., 2003). The pre-SMA is considered a non-primary motor area (Woolsey et al., 1952; Matelli et al., 1985, 1991; Barbas & Pandya, 1987; Matsuzaka & Tanji;, 1996, Zilles et al., 1995; Zilles et al., 1996, Geyer et al., 2000), and is involved in higher-level aspects of motor behavior such as motor selection and inhibition, which is of critical relevance in current behavioral investigations and psychiatric interventions. This array of functions is reflected by pre-SMA’s distinct structural connectivity (Eickhoff et al., 2011; Picard & Strick, 1996; Rizzolatti & Luppino, 2001; Nachev et al., 2005, 2007, 2008; Shima et al., 1996). This cortical area is located in the mesial aspect of the caudal part of each frontal lobe and rostral to the SMA and the primary motor cortex as illustrated in Fig. 1. Its structural connectome is complex, including short- and medium-range fibers as well as long association fiber tracts such as the cingulum bundle (CB). A critically important structural connection of pre-SMA is with the subthalamic nucleus via the hyperdirect pathway. This structural connection makes the pre-SMA a cortical target (or “portal”) with direct access to the striatal circuitry via the subthalamic nucleus (STN). This is particularly relevant in the treatment of SUD and behavioral addictions such as gambling disorder, considering the association of the corticostriatal circuitry in compulsive behaviors (see, e.g., Robbins et al., 2019; Greenberg et al., 2010; Fineberg et al., 2010; Fettes et al., 2017; Alexander et al., 1986; Milad & Rauch, 2012; Graybiel & Raucht, 2022). Overall, given its important connections and bio-behavioral functions, the pre-SMA has become recently a very important neuroanatomical target for neuromodulation treatments in psychiatry.

Fig. 1
figure 1

Pre-supplementary motor area (pre-SMA) and its structural connectome. Pre-SMA is located in the mesial aspect of each hemisphere and occupies the anterior portion of Brodmann’s architectonic area 6 (BA6) (Rushmore et al., 2022 – HOA ComPaRe paper). In the Talairach coordinate system (Talairach & Tournoux, 1988), the pre-SMA lies approximately between an anterior coronal plane at the callosal genu and a posterior coronal plane at the anterior commissure (AC). Its inferior border is the cingulate sulcus and extends to the hemispheric margin superiorly. Pre-SMA’s morphology and topographical location are illustrated in A on a medial magnetic resonance imaging (MRI) view of a human subject’s hemisphere, whereas similar information is shown in B using an inflated cortical surface representation of T1-weighted MRI data. A schematic diagram of pre-SMA’s structural circuitry is illustrated in C; this circuitry includes short- and medium-range fibers (s.r.f; m.r.f.) and long association fiber tracts such as the cingulum bundle (CB). An important structural connection of pre-SMA is with the subthalamic nucleus via the hyperdirect pathway (h-d.p.)

Full size image

Recent studies have shown that alterations in pre-SMA activity are associated with the severity of addiction-related behaviors across addiction types (Huang & Ma, 2023; Bell et al., 2014; Lohse et al., 2023), in light of the causal link emerged between the level of pre-SMA activity and the propensity for impulsive risk-taking behavior in the context of gambling behaviors (Lohse et al., 2023).

Theta-burst stimulation (TBS) is a brief form of repetitive transcranial magnetic stimulation (rTMS) that holds promise in targeting specific brain regions and altering neuronal activity patterns (Huang et al., 2005), by inducing long-term potentiation (LTP) and long-term depression (LTD)–like phenomena (Suppa et al., 2016). TBS can be intermittent (iTBS) or continuous (cTBS). iTBS involves administering brief TBS trains at intervals, maintaining a prevailing excitatory effect and eliciting LTP. In contrast, cTBS is administered continuously for a duration sufficient to enable the inhibitory effect to surpass the facilitatory effect, leading to LTD (Suppa et al., 2016). As such, TBS offers a potential avenue for modulating pre-SMA function and, by extension, influencing addictive behaviors. The application of TBS to addiction treatment represents a novel and innovative approach, with the potential to mitigate the devastating impact of both behavioral addictions like GD and SUDs, as it has been shown in major depressive disorder (MDD) and obsessive compulsive disorder (OCD) (Blumberger et al., 2018; Williams et al., 2021; Cole et al., 2022). In a randomized-controlled trial, Pallanti et al. (2023) have investigated the efficacy of continuous theta-burst stimulation (cTBS) in gambling disorder when targeting the pre-SMA, showing a significant reduction of craving and gambling severity only in the active stimulation group. These findings provide a model for exploring its utility across different addiction types. Furthermore, Pallanti et al. (2022) showed the beneficial effects of cTBS over the pre-SMA in SUDs in comorbidity with schizophrenia.

However, at our knowledge, no study directly investigated the effects of cTBS in SUDs and compare the efficacy of such a protocol in BA and SUDs, with the aim not only to understand the treatment potential of this intervention but also to investigate whether targeting the pre-SMA could yield to different outcomes in these two groups.

In this naturalistic study, we seek to explore the therapeutic effects of pre-SMA cTBS on individuals diagnosed with GD and SUDs, comparing the outcomes between these two distinct forms of addiction. By utilizing a comprehensive battery of outcome measures, we aim to shed light on the potential divergent and shared neuronal mechanisms underlying these addiction types. We expect to observe an improvement in symptoms’ severity in both groups, with specific affected domains. Considering the lack of FDA-approved treatment, through this research, we hope to contribute to the development of targeted interventions that can improve the lives of individuals grappling with the complex challenges of addiction.

Methods

Participants

The data of patients aged 18–55 with a diagnosis of SUD (polysubstance use, cocaine, cannabis) or GD, who underwent rTMS, were extracted from databases containing information on patients of the psychiatric clinic at the Istituto di Neuroscienze, Florence (Italy). In addition, a previous rTMS treatment, concomitant alcohol or drug use (all patients of both conditions were abstinent from 1 month before the start of the intervention), a diagnosis of MDD, bipolar disorder, psychotic or neurological disorder, illiteracy, or cognitive impairment were all exclusion criteria, with the aim to remove potential confounding variables. All diagnoses were based on criteria of the DSM-5. All patients underwent a safety screening for TMS, to exclude history of epilepsy, seizures or increased risk of seizures for any reason, history of any metal in the head (outside the mouth) or metallic implant, and known or suspected pregnancy or lactation. rTMS was added to ongoing pharmacological treatments, which were unchanged during 1 month prior to the start of TBS protocol and remained stable throughout the duration of the neurostimulation sessions. The data of 58 patients were included for the analysis (Mage = 37, SDage = 11.64, nmales = 35 (60%), nfemales = 23 (40%)). After the complete description of the study to participants, written informed consent was obtained from each one for the inclusion of their data in this study.

Procedure and TMS Protocols

cTBS was administered with the MagVenture MagPro R30 stimulator with add-on theta-burst option (MagVenture INC.) using a Cool D-B80 figure-of-eight coil. cTBS consists of bursts of 3 pulses separated by 20 ms (i.e., 50 Hz) delivered repeatedly at theta frequency on the pre-SMA bilaterally (Fig. 1).

Stimulus intensities were set at 80% of resting motor threshold (RMT). Two trains of 600 pulses (a total of 1200 pulses) were administered with an 8-s intertrain interval. The protocol consisted of 20 daily sessions. The bilateral pre-SMA was targeted using individual MRI and a neuronavigation system (SofTaxic Optic 2.0). The TMS coil was held by a mechanical arm. RMT was defined as the minimum pulse strength needed to elicit a response in a resting target muscle (abductor pollicis brevis) in 5/10 trials using single-pulse TMS administered to the contralateral primary motor cortex.

Assessment

Participants’ addiction severity was assessed using standardized scales: the Addiction Severity Index (ASI; McLellan et al., 1980), a structured interview (Cronbach’s α = 0.82), which assesses the severity of a person’s substance abuse (e.g., “How much money would you say you spent during the past 30 days on drugs?”), assigning a severity index on a scale from 0 to 9; and the Visual Analog Scale (VAS) for craving intensity, which is a single-item scale (Cronbach’s α = 0.91) in which the patient is asked to indicate the point on a line that represents your current state with respect to a substance “With zero (0) meaning ‘No Craving At All’ and 10 meaning ‘Strongest Craving Ever’” (Mottola, 1993). Additionally, UPPS-P Impulsive Behavior Scale, a 59-item scale (e.g. “When I am upset I often act without thinking”) with a maximum score of 236 (Cronbach’s α = 0.86), which evaluates impulsivity across multiple domains (Whiteside & Lynam, 2001) was administered, considering that impulsivity is one of the main features of the addicted clinical picture (Dawe & Loxton, 2004; Lee et al., 2019) and that both impulsive and compulsive dimensions involve motor/response disinhibition (Robbins et al., 2012), which is associated with pre-SMA (Obeso et al., 2017). Sheehan Disability Scale (SDS; Sheehan, 1983), a 3-item scale (e.g. “The symptoms have disrupted your ability to function at work”) assessing work, social, and family disability with a maximum score of 30 (Cronbach’s α = 0.83), was used to assess functional impairment and disability across multiple life domains, in order to identify whether addiction severity is related to a higher degree of impairment in everyday activities.

These assessments were conducted at baseline and post-treatment.

Statistical Analysis

The baseline demographic and clinical characteristics of the sample were tabulated with descriptive statistics. Parametric (t test) and non-parametric (Wilcoxon) tests were used according to variables’ distribution (tested with the Shapiro–Wilk test) to analyze changes in scores over time and the interaction effect “group × time.” For all statistical analyses, the alpha level of significance was set at 0.05. All the statistical analyses were performed with the statistical programming language R (version 4.0.5).

Results

Of the 58 subjects included in the study, 25 had GD and 33 SUD. Our study demonstrated significant improvements in both BA and SUD groups following cTBS to the bilateral pre-SMA. Both groups (see Figs. 2 and 3) showed a significant reduction in ASI scores (GD: V = 170, p = 0.0023, r = 0.43; SUD: V = 345, p = 0.005, r = 0.23), indicating a decrease in overall addiction severity, in craving intensity, as measured by VAS, (GD: V = 226.5, p = 0.001, r = 0.45; SUD: V = 245, p = 0.023, r = 0.16) and in CPFQ (p < 0.01).

Fig. 2
figure 2

Addiction Severity Index (ASI), a measure of addiction’s symptoms severity, median scores before (pre-test) and after treatment (post-test) in the two groups: gambling disorder (GD; red dotted line) vs. substance use disorder (SUD; blue solid line)

Full size image
Fig. 3
figure 3

Visual Analogue Scale (VAS), a measure of craving, median scores before (pre-test) and after treatment (post-test) in the two groups: gambling disorder (GD; red dotted line) vs. substance use disorder (SUD; blue solid line)

Full size image

Impulsivity (see Fig. 4), measured with UPPR-S decreased significantly in SUD (t = 5.86, p < 0.001, d = 1) but not in GD (t = 1.25, p > 0.05). When a mixed-effect ANOVA was run to assess the interaction effect “group × time,” it was not significant for ASI and CPFQ scores, while a significant interaction effect was observed for craving (F = 4.47, p = 0.037) and UPPR-S (F = 4.77, p = 0.031).

Fig. 4
figure 4

UPPS-P Impulsive Behavior Scale, a measure of impulsivity, mean scores before (pre-test) and after treatment (post-test) in the two groups: gambling disorder (GD; red dotted line) vs. substance use disorder (SUD; blue solid line)

Full size image

Discussion

In this comparative study examining the impact of cTBS on the pre-SMA in gambling disorder versus SUDs, the treatment demonstrated effectiveness in both disorders, albeit with distinct outcomes. In both conditions, there was a reduction in addiction severity, indicating that pre-SMA cTBS might have a broad influence on addictive behaviors irrespective of the specific addiction type. Craving also diminished, with a more pronounced improvement observed in BA, suggesting that pre-SMA cTBS may modulate craving-related neural circuits relevant to both BA and SUD. However, alternative mechanisms may be implicated in SUD-related craving. Notably, impulsivity showed improvement only in SUD, not in GD. The efficacy of pre-SMA cTBS in GD corroborates prior findings (Pallanti et al., 2023). While this study presents novel insights into the effects of this protocol on SUD, earlier research indicated preliminary findings in schizophrenia with comorbid SUD (Pallanti et al., 2022).

Implications for a Shared Framework for Addiction Treatment

The observed reduction in addiction severity for both disorders underscores a promising common therapeutic dimension for cTBS in addressing addictive behaviors across the spectrum. This suggests that TBS may possess the potential to alleviate the adverse consequences of both BA and SUDs, aligning with a growing body of research indicating overlapping neuronal circuits in these conditions (Balodis & Potenza, 2020). Approaching this from an RDoC perspective aids in identifying the neurobiological factors underlying the two disorders. The deficits evident in both disorders are linked to the RDoC domains of positive valence systems (specifically approach motivation and reward learning) and cognitive control (particularly the construct of response inhibition), respectively (Pallanti et al., 2021), with the latter associated with the functioning of the pre-SMA. Our results, in this sense, suggested a shared framework for shared domains common to different forms of addiction, with one of this emerged through the targeting of response inhibition and the cortical portal represented by the pre-SMA.

Impulsivity in BA and SUD

The absence of improvement in impulsivity in GD aligns with the findings of Pallanti et al. (2023). However, noteworthy is the observed reduction in impulsivity reported in the SUD group in this study. It is crucial to highlight distinctions that have surfaced when comparing impulsivity in SUD and GD (Leeman & Potenza, 2012). Additionally, prior reports have indicated that individuals with BA tend to display lower levels of impulsivity compared to those with SUD, potentially linked to the neurotoxic effects of drug exposure (Albein-Urios et al., 2012).

Cognitive Rigidity and Addiction

These results warrant interpretation within a broader context that considers cognitive rigidity in both SUD and GD patients. Shared neural indicators, such as diminished ventrolateral prefrontal cortex (vlPFC) activation associated with cognitive inflexibility, have been observed in both gambling and cocaine addictions (Verdejo-Garcia et al., 2015). However, the effects linked to cocaine use extend to a wider range of task-related dysregulation patterns, manifesting as irregular signaling in dorsolateral and dorsomedial PFC regions. A study utilizing a reversal learning task as a measure of cognitive flexibility deficits identified cognitive rigidity as a key dimension of addiction disorders, encompassing aspects of both impulsivity and compulsivity (Izquierdo & Jentsch, 2012; Lee et al., 2019; Winstanley et al., 2010). While addiction models suggest a shift from initial impulsivity to dependence-related compulsivity, it is noted that impulsivity retains a role even in the later phases of the disorder (Winstanley et al., 2010; Lee et al., 2019). In this sense, new models of addiction should not focus on the distinctions of impulsivity and compulsivity, but on the cognitive rigidity dimension.

The Evolutionary Roots of Craving: Toward a New Model of Addictive Behaviors

The most significant clinical insight from this study lies in the notable improvement in craving, a pivotal aspect of addiction. A comprehensive understanding of addiction and craving necessitates delving into the evolutionary origins of craving (Liedtke et al., 2011) and acknowledging its dual polarities. One polarity characterizes craving as a response to the need to take risks in a negative contextual setting, associated with impulsive behaviors. The other polarity involves craving and compulsivity aimed at enhancing resources in a favorable context. Phenomenologically, these two behavioral modules can be differentiated. The human inclination to use drugs has capitalized on these evolutionarily relevant mechanisms of gratification (Liedtke et al., 2011).

Beyond the transition from ventral striatum activation to dorsal striatum activation during dependence (Everitt & Robbins, 2013), the role of context and ancient survival mechanisms becomes pivotal. Consequently, the activation of a gratification component can potentially be distinguished at three moments: activation preceding the behavior, activation following the behavior, and activation associated with context uncertainty. These forms may be linked to distinct reward circuits (Koob & Volkow, 2010). This discussion aims to enhance existing models of addictive behaviors and treatment strategies by proposing the consideration that recovery from addiction does not entail the absence of craving—a evolutionarily necessary phenomenon—but rather a reduction in craving below a threshold where it becomes controllable.

Brain and Potential Therapeutic Mechanisms

The reported improvements may be attributed to the utilization of two distinct brain mechanisms. Firstly, our results could stem from the direct modulation of pre-SMA activity, leading to an enhancement in inhibitory control. This improvement, in turn, contributes to a reduction in behaviors associated with gambling or substance intake, as well as a decrease in related thoughts. Secondly, these outcomes may be elucidated by the modulation of other brain regions interconnected with the pre-SMA. It is crucial to underscore the significance of the pre-SMA structural connectome’s anatomy, particularly its direct anatomical link with the subthalamic nucleus (STN) through the hyperdirect pathway projection. Additionally, its association connectivity via short-range, middle-range, and long fiber tracks plays a critical role (refer to, e.g., Makris et al., 2023). The hyperdirect pathway acts as a direct mechanism for modulating corticostriatal circuitry through transcranial magnetic stimulation (TMS) stimulation of the pre-SMA. Given the involvement of the striatum in pathological mechanisms of gambling (Pallanti et al., 2010; Hollander et al., 2005) and substance use (Kravitz et al., 2015), this pathway may be implicated in the observed results. To validate these hypotheses and provide insight into the neurofunctional changes underlying the observed therapeutic differences between the two groups, future neuroimaging studies should be conducted.

Limitations and Future Research

The TMS device here used did not allow to target specifically the left or right pre-SMA; indeed, bilateral pre-SMA was stimulated. It is not clear whether our results could have been affected by this. In the future, network-informed targeting (Lioumis & Rosanova, 2022) will be more helpful to decide whether lateral or bilateral targeting is optimal and multi-locus TMS (mTMS) to simultaneously modulate more than one cortical structure (Nieminen et al., 2022). Moreover, the 80% RMT here adopted is a low intensity. Future studies should dose intensity with TMS–EEG, since neuronal perturbations may depend on the regional and personal cytoarchitecture (Rosanova et al., 2009), with the aim also to reduce the number of sessions needed to have the desired result (Casarotto et al., 2022; Lioumis & Rosanova, 2022). Another limitation is associated with the retrospective design of the study. Retrospective studies depend on preexisting data, which may be susceptible to recall bias and constrained by the availability of historical records. Additionally, given that the study was not sham-controlled to mitigate bias, it remains possible that unblinding or placebo effects could have influenced the outcomes. Moreover, the potential impact of comorbidity with ADHD should be considered. Many individuals with substance use or gambling show attentional deficits (Luo & Levin, 2017), which relate to executive functioning alterations, impulsivity, and emotional dysregulation.

Conclusions

The promising results of this study highlight the potential of pre-SMA TBS as an adjunctive treatment for both GD and SUDs. By targeting shared neuronal mechanisms related to craving, cognitive control, and impulsivity, cTBS offers a novel approach to addressing behavioral and substance addictions.