Cocaine use disorder (CUD) can lead to impaired social and vocational functioning, lost opportunities, and the increased risk of morbidity and mortality (1, 2). However, effective pharmacologic treatment remains elusive (3, 4). Clinical trials have searched for medications that could help patients reach remission, but no clear options have emerged. While studies suggest that some pharmacotherapy (such as long-acting stimulants or combined medications) may help some patients, there remains a need for therapeutic options (1, 2, 5).

Research shows that transcranial magnetic stimulation (TMS) may serve as a treatment for CUD by reducing craving and cocaine use (6, 7). Although a pivotal trial has yet to be published, early studies using heterogeneous approaches – including imaging, open-label trials, retrospective record reviews, and randomized controlled trials – have shown promise for this refractory disorder.

Seizure is a rare but concerning potential adverse event associated with TMS (8, 9). While the incidence is low, there is a concern of increased seizure risk among individuals with CUD based on data showing that cocaine exposure itself can cause a seizure (10–26).

Given the promise of TMS for CUD, our goal was to characterize the risk of seizure and to identify methods to reduce it. We began with a review of seizures and cocaine use, followed by an analysis of studies where cocaine was safely administered in laboratories. We then reviewed clinical trials of TMS in CUD, with a focus on reports of seizures in this population. From this analysis, we provide suggestions that can be used to mitigate the seizure risk on TMS in CUD.

In the 1980s, cocaine production and consumption increased across the globe. Clinical researchers soon identified the association between cocaine exposure and seizure, which is now a recognized potential adverse event. We reviewed the literature on this topic, starting with a systematic review by Sordo et al. (27), which identified 11 cross-sectional studies (10–20) and one case-control study (21) reporting on seizure prevalence among individuals using cocaine. We then replicated the search of Sordo et al., to identify additional studies published since their 2013 review by searching Medline, EMBASE, and PsycINFO using the terms “cocaine, cocaine-related disorders, cocaethylene, and seizures” and excluded unverified reports of seizure.

This returned a total of 17 studies (Table 1), which show that seizure was seen in 1% to 27% of individuals who had used cocaine recently (intravenous, smoked, insufflated, or oral) (10–26). As shown in Table 1, all studies included patients receiving acute medical care in emergency departments or hospitals due to cocaine use alone or due to illness where cocaine use was a contributing factor. The highest prevalence (27%) was reported in a retrospective analysis of cocaine-related deaths, suggesting that the greatest risk of seizure may be during a fatal overdose (23).

We then reviewed studies that included self-reports of seizure by cocaine users, which was not observed by medical personnel (28–32). The prevalence rates varied from 0.9% to 18.1%. This large range can be partly attributed to seizures being described as “fits, convulsions, fainting, and loss of consciousness” – such that these reports included non-seizure events. Nonetheless, these studies also indicated that seizures occurred in the setting of recent, high dose cocaine use that caused symptoms of physical and/or mental distress. Thius, this finding is similar to the studies with confirmed seizures witnessed by medical personnel.

Taken together, this research indicates that individuals who use cocaine and consequently require acute medical attention are at risk of seizure. We found no reports of increased seizure in persons with CUD outside of this setting. While the lack of data cannot provide definitive evidence, it is remarkable that there are no reports of increased rates of seizure associated with CUD itself.

With respect to TMS, this data suggests that seizure risk would be elevated following recent cocaine use, especially at toxic doses that induce illness – and could be low outside of this setting. In the next sections, we review methods that can be used to assess recent use and toxic doses.

A urine drug screen (UDS) is a reliable method for determining recent cocaine use. However, it detects benzoylecgonine (an inactive metabolite), which can persist in the urine for 2–4 days, well after cocaine has cleared the plasma (33). Additionally, although uncommon, a UDS can return a false negative result where cocaine is in the blood but benzoylecgonine has yet to appear in urine (34, 35). This can occur when cocaine use is very recent (within about 2 hours) before the metabolite is detectable (35, 36).

We sought to identify clinical parameters that could assess recent use – instead of a UDS or in addition to this test. For this goal, we reviewed studies where cocaine was taken by participants in a laboratory setting under close monitoring. This provided information on the signs and symptoms of recent cocaine use, which correlate with its pharmacokinetic profile.

The half-life of cocaine in the plasma is 45 to 90 minutes following smoked, intranasal, or intravenous administration (37–40). Once five half-lives have passed (or about 7.5 hours), plasma levels become undetectable (37–40). The subjective effects of cocaine intoxication include feeling high, alert, self-confident, social, talkative, anxious, irritable, and ‘on edge’ (41–46). These symptoms peak within about 5 minutes and resolve within about 30–45 minutes following a single dose (40, 45, 47–50). Repeated dosing extends the subjective effects of cocaine longer, beyond 100 minutes or more (41, 51, 52).

Increases in vital signs are another sign of cocaine intoxication. Studies show that cocaine increases heart rate by about 20–40 beats per minute and increases blood pressure (systolic and diastolic) by about 10–30 mmHg over baseline (41–46). These remain elevated for 2 to 3 hours following smoked or intravenous cocaine use (41, 44, 53).

Thus, the signs and symptoms of cocaine intoxication can be used to evaluate individuals for recent use, including subjective effects and vital signs. However, an additional approach is to establish a collaborative, non-judgmental dialogue with individuals about their patterns of cocaine use This could include the estimated dosage, timing, and route of administration – as well as a discussion of the potential risk of TMS in the setting of recent cocaine exposure.

The difference between a euphorigenic dose of cocaine (which produces a desired effect) and a toxic dose (which causes distress) can be miscalculated, even among individuals with regular cocaine use. A toxic dose, also called an overdose, leads to symptoms of physical and mental distress and requires a medical assessment. Thus, in addition to recent use, evaluating this parameter would improve safety.

The signs and symptoms of cocaine toxicity have been divided into three stages of increasing severity (54). Stage 1 includes central nervous system effects (headache, nausea, mydriasis, vertigo, twitching, pseudo-hallucinations, and pre-convulsive movements); changes in vital signs (hypertension, tachycardia, tachypnea, hyperthermia, ectopic beats); and psychiatric symptoms (paranoia, euphoria, confusion, aggression, agitation, emotional lability, restlessness). Stage 2 consists of additional symptoms of greater severity: encephalopathy, seizures, increased deep tendon reflexes, incontinence, arrhythmias, peripheral cyanosis, gasping, apnea, and irregular breathing (54). Stage 3 includes symptoms that can be life-threatening: areflexia, coma, fixed and dilated pupils, loss of vital function, hypotension, ventricular fibrillation, cardiac arrest, respiratory failure, cyanosis, and agonal breathing (54).

In reviewing the symptoms of cocaine toxicity, it is key to recognize that some individuals will experience stage 1 symptoms without progressing to stage 2 or 3 – and that not all symptoms listed above are required to meet criteria for cocaine overdose. Furthermore, some individuals might not have sought medical care despite having severe symptoms.

With respect to TMS, an evaluation of cocaine overdose should include both recent and past episodes. A recent episode could indicate that a clinical evaluation is needed to rule out ongoing illness, such as neurological or cardiac events. A past event may not be exclusionary for TMS, but should be assessed, since an overdose could have included neurological events, like ischemic stroke and intracerebral or subarachnoid hemorrhage (55, 56).

We reviewed studies of TMS in CUD, in order to evaluate two parameters: 1) whether an assessment of recent use was included prior to the delivery of TMS; and 2) whether seizure was reported. In May 2025, we searched PubMed using the term “cocaine and transcranial magnetic stimulation,” which returned 91 studies (Figure 1). Inclusion criteria consisted of investigational studies examining the use of TMS for CUD that were published in English. Publications were excluded if they were literature reviews, conference abstracts, commentaries, secondary analysis of already included data, or non-human investigations, leaving 32 studies for further review.

Of these, 11 studies required abstinence or a urine drug screen negative for cocaine prior to the delivery of TMS sessions (57–67). Notably, these studies used outcomes that were not affected by the requirement of abstinence, such as imaging, measures of craving, or cocaine self-administration. There were no reports of seizures.

We then reviewed publications that did not specifically require abstinence, in order to determine whether recent cocaine use was assessed prior to the delivery of TMS. This resulted in 21 studies (Table 2). As shown in Table 2, these publications assessed cocaine use with heterogeneous methods, including UDS, hair analysis, self-report, and reports from significant others (68–88). However, only five (of the 21) studies clearly reported that recent use was evaluated prior to the delivery of TMS (70, 74, 75, 81, 85). The remaining studies did not specify whether cocaine use was assessed prior to each TMS session. Future studies that clarify this parameter would benefit the field of TMS research.

Among the studies of TMS in CUD, we found two reports of a seizure (75, 85). Madeo et al. (75) performed an observational study (n = 248) and reported that one participant experienced a seizure following cocaine use. The authors reported that 66 days had passed since the first TMS session, but the precise time between the most recent TMS session and the seizure was not published (although the authors stated that TMS had not recently been delivered). Martinotti et al. (85) performed a randomized, sham-controlled study (n = 80) and reported a seizure-like episode in one participant in the active TMS group. The event occurred following intravenous cocaine use, but the proximity to the TMS session was not reported. Moreover, both studies were limited in their description of seizure-like activity and appeared to rely on self-report (75, 85).

Steele et al. (74) reported a transient neurological event of uncertain etiology in an open-label pilot study (n = 19) where a participant experienced rhythmic right-hand supination and pronation at the wrist. The event began 10–15 minutes following TMS and persisted for about 3 minutes (74, 89). The episode resolved within 1 hour, and no further TMS sessions were administered. Steele et al. (90) also reported a new-onset tremor in a participant with CUD who received sham intermittent theta burst stimulation.

Taken together, this data suggests that seizure is uncommon in studies using TMS in CUD. This finding is consistent with a recent meta-analysis, which reported a seizure prevalence of 2.32 per thousand participants receiving active TMS for individuals with CUD (91). However, clinical studies that describe the temporal association between cocaine use and the delivery of TMS, and the time between TMS and serious adverse events, such as seizure, would further demonstrate the safety of this intervention.

Previous TMS studies have shown that participants with CUD have a higher motor threshold compared to controls (63, 66, 92, 93). Hanlon et al. (93) investigated the mechanism behind this using intracortical facilitation and functional MRI. The findings suggest that cocaine use is associated with reduced cortical facilitation, which may increase motor threshold and alter motor cortex BOLD signal, potentially mediated by glutamate and GABA signaling (93).

It remains unknown whether the elevated motor threshold seen in CUD varies with the amount of use or the duration of abstinence. Some TMS in CUD studies measured motor threshold prior to each session (59, 61, 74), which would address this potential variability. Alternatively, a recent study obtained a UDS prior to each TMS session, and the motor threshold was reassessed in the event of a newly positive result (NCT04907357). Either of these approaches would be expected to accommodate possible changes in cortical excitability.

Pharmacological treatment for CUD has been elusive, despite numerous clinical trials (3, 4). Additional treatments are needed, and research investigating TMS for CUD shows that this intervention reduces craving and cocaine use (6, 7, 94). Although pivotal clinical trials are still needed, there is a strong rationale for pursuing TMS as a therapy.

Given the potential benefit of TMS, there is a need to characterize the risk of seizure in CUD (95). To this end, we reviewed the research on cocaine exposure and seizure risk, which showed that patients requiring urgent medical care due to recent, high dose cocaine use were at risk for seizure. Outside of this scenario, there are no published reports of increased seizure rates among individuals with CUD. This suggests that CUD itself does not increase seizure risk – although this absence of documentation is not definitive.

With respect to TMS, the following steps, outlined in Figure 2, would be expected to mitigate seizure risk:

Our review of the literature has several limitations. As noted, there is substantial heterogeneity across studies in TMS protocols, approaches to assessing recent cocaine use, and methods for reporting seizure-like activity. These discrepancies hinder our ability to understand how TMS, cocaine use, and seizure events are related in terms of timing and sequence. In addition, a recent meta-analysis (91) that included several of the cocaine studies discussed here found a high risk of bias in these studies.

Furthermore, our publication did not review the efficacy of TMS for CUD, nor the occurrence of adverse events other than seizure. For these topics, we refer the reader to systematic reviews of TMS in SUD (6, 96) and substance use and to Blyth et al. (91), who reviewed the overall risk of adverse events in this patient population.