Edited by: Johannes Petzold, University Hospital Carl Gustav Carus, Germany
Reviewed by: Guijun Dong, Quzhou University, China; Domenico De Berardis, Azienda Sanitaria Locale 4, Italy
This article was submitted to Addictive Disorders, a section of the journal Frontiers in Psychiatry
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
Achieving abstinence from drugs is a long journey and can be particularly challenging in the case of methamphetamine, which has a higher relapse rate than other drugs. Therefore, real-time monitoring of patients’ physiological conditions before and when cravings arise to reduce the chance of relapse might help to improve clinical outcomes. Conventional treatments, such as behavior therapy and peer support, often cannot provide timely intervention, reducing the efficiency of these therapies. To more effectively treat methamphetamine addiction in real-time, we propose an intelligent closed-loop transcranial magnetic stimulation (TMS) neuromodulation system based on multimodal electroencephalogram–functional near-infrared spectroscopy (EEG-fNIRS) measurements. This review summarizes the essential modules required for a wearable system to treat addiction efficiently. First, the advantages of neuroimaging over conventional techniques such as analysis of sweat, saliva, or urine for addiction detection are discussed. The knowledge to implement wearable, compact, and user-friendly closed-loop systems with EEG and fNIRS are reviewed. The features of EEG and fNIRS signals in patients with methamphetamine use disorder are summarized. EEG biomarkers are categorized into frequency and time domain and topography-related parameters, whereas for fNIRS, hemoglobin concentration variation and functional connectivity of cortices are described. Following this, the applications of two commonly used neuromodulation technologies, transcranial direct current stimulation and TMS, in patients with methamphetamine use disorder are introduced. The challenges of implementing intelligent closed-loop TMS modulation based on multimodal EEG-fNIRS are summarized, followed by a discussion of potential research directions and the promising future of this approach, including potential applications to other substance use disorders.
Addiction is defined as a strong need to use a particular substance or engage in a specific behavior, often in spite of harmful consequences. Addiction not only causes personal health problems but can have severe social impacts (
Methamphetamine (METH) has been among the most frequently misused drugs for the past two decades in Southeast and East Asia (
Conventional approaches to detect METH usage include analysis of sweat, saliva, or urine (
Various neuroimaging techniques have been used to study the influence of METH use on cognitive functions (
Electroencephalogram (EEG) and functional near-infrared spectroscopy (fNIRS) are promising tools for brain signal monitoring. Biomarkers in EEG signals have been explored in patients with METH addiction (
Various intervention approaches are used to help individuals with drug addictions. The conventional treatments are psychological counseling, family support, and legal restriction. These interventions are often planned according to a regular schedule with a set frequency. Family and social support often depend on the willingness and availability of others. Legal restrictions are often applied too late, when drug compulsive use is already established. Achieving abstinence from drug use is a long journey with a need for high self-motivation as well as external influences and legal constraints. Therefore, successful abstinence is a challenge. In contrast to the passive methods mentioned above, neurostimulation has the potential to alter brain activity to reduce cravings for drug use (
Neuromodulations exists as both invasive and non-invasive types. Deep brain stimulation (DBS) is the most frequently reported invasive neuromodulation solution for drug addiction (
The workflow examples of
To achieve effective treatment for METH addiction, three key elements are required for an intelligent closed-loop TMS neuromodulation system based on multimodal EEG-fNIRS measurements: (1) an appropriate measurement protocol for multimodal EEG-fNIRS monitoring, (2) intelligent signal-processing strategies, and (3) customized, user-friendly, wearable TMS devices. Section “2. Materials and methods” introduces the materials and methods of conducting the literatures collection for this review. Section “3. Detection and monitoring techniques for METH addiction” of this manuscript reviews the available techniques for physiological monitoring to detect drug use and addiction, with a particular focus is on wearable EEG and fNIRS neuroimaging approaches. In section “4. Biomarkers of neuroimaging techniques,” biomarkers in EEG and fNIRS recordings to identify METH addiction and the progression of recovery during abstinence are discussed. In section “5. Neuromodulation treatments for METH addiction,” the most common neuromodulation treatments for METH, tDCS, and TMS are introduced, and research supporting the efficacy of these treatments is summarized. The promising future of the application of intelligent closed-loop TMS modulation based on multimodal EEG-fNIRS for METH addiction is summarized in section “6. Challenges and future trends in treatment of METH addiction,” and the challenges to be overcome to achieve an optimized closed-loop system are discussed.
We did not conduct a systematic review since this type of review is less common in engineering than in the medical and public health fields. This review aims to summarize the challenges and propose a future trend of a wearable closed-loop neuromodulation system for METH addiction treatment from an engineering point of view based on the available evidence. Our review provides insights into combing the three key elements, biomarkers, real-time signal analysis approach and neuromodulations, of the proposed closed-loop system. We do not aim to find a fixed answer to a specific question or an optimal medical therapy as a standard systematic review does. Neither needs all available evidence to support the concept of the wearable closed-loop system. In addition, the need to reduce the total bias and quantify (statistical analysis) the available results is not the top priority. For building a biomedical system, interdisciplinary knowledge is needed. Therefore, a systematic review might not be the best approach to convey our perspectives.
Since this review contains a wide subtopic, the key words used for searching the published journal papers are introduced in this section. The databases “Web of Science” and “Google” were used. For section “4.2. Biomarkers of EEG and fNIRS in METH addiction,” key words used to search the EEG biomarkers in METH addiction were EEG or Electroencephalography or Electroencephalographic and Methamphetamine or Meth. The keywords used to search fNIRS biomarkers in METH addiction are functional near-infrared spectroscopy, fNIRS, NIRS, and Methamphetamine or Meth. The validated papers suggest potential EEG or fNIRS biomarkers to distinguish the subjects with METH use disorder from healthy ones. In addition, articles that provide biomarker information to classify the subjects of METH use disorder receiving different treatments or at a different phase of abstinence are included in this review article. For section “5.1. tDCS for methamphetamine addiction,” key words applied for searching are transcranial direct current stimulation or tDCS and Methamphetamine or Meth. For section “5.2. TMS for methamphetamine addiction,” keywords used to explore the related studies are transcranial magnetic stimulation or TMS or theta burst stimulation and Methamphetamine or Meth. Only the studies with solid conclusions that certain tDCS or TMS protocols are helpful to treat patients with METH use disorder are included in this review article. No matter the validation approaches for the outcomes of neuromodulation. The techniques include self-rating scales, questionnaires, cognitive tasks, physiological signals, or neuroimaging. For all the found literature, only those that conducted the experiments on humans are reported in this review article.
Several approaches can be used to determine if an individual meets the criteria for diagnosis of a dependence or addiction. Subjectively, questionnaires are highly accessible and easily administered to evaluate the condition of drug addiction. Also, various methods are available to quantitatively detect physiological parameters of patients who use illicit drugs or suffer from drug addiction. Some methods detect drugs in biological fluids, while others detect neurological signals. These approaches are discussed in detail in the following sections.
Various questionnaires have been developed for use before a person becomes addicted to drugs and during abstinence to predict the likelihood of relapse. The Inventory of Drug-Taking Situations is a questionnaire to judge the risk of drug addiction based on everyday situations. Abuse of drugs can be screened for using the assessment tools suggested by the National Institute on Drug Abuse (
As drug addiction affects has both physical and psychosocial effects, scales that characterize anxiety, depression, or impulsivity resulting from drug abuse are often used to obtain a broad view dependence and substance abuse. These scales include the 21-item Beck Anxiety Inventory, 21-item Beck Depression Inventory, and 30-item Barratt Impulsiveness Scale-11 (
Although questionnaires are the most convenient and broadly accessible approach for a variety of drug addiction-related applications, the results of the scales have limited reliability and accuracy. As the results of these scales are based on the answers of the respondents, the results exhibit individual variations and are influenced by the attitudes and conditions of the participants (
In addition to questionnaires, saliva, urine, and blood tests are conventional ways to quantitatively detect drug use (
In addition to ECG and HR, drug abuse has effects on pupil size (
The wearable techniques presented in this section are used to measure the physiological reactions of the autonomic nervous system caused by drug usage. However, drug addiction also affects the central nervous system (
Whereas the wearable technologies discussed in the previous section can be used to monitor physiological changes in the body as relates to drug use and withdrawal, clinicians and researchers are also keen to learn how drug use influences the control center of the body, the brain. The process of developing an addiction includes several phases, including drug intoxication, craving, binging, and withdrawal with loss of self-control (
Neuroimaging approaches can be separated into two categories: those that can be used to examine the structural changes at different stages of drug dependence and addiction to understand the physiological mechanisms of addiction (
Although these neuroimaging techniques provide high spatial resolution, they have limitations. For instance, the devices required are costly and bulky, with low accessibility, and trained operators are needed to conduct the examinations. Moreover, MRI may not be suitable for patients with metallic implants, and the use of radioactive agents in PET limits the frequency of examinations. Furthermore, those with claustrophobia may find it difficult to participate the examinations owing to the spatially confined test environment that is needed. Low temporal resolution also restricts the application of real-time monitoring.
Wearable devices, being compact and easy to access, are appropriate for real-time monitoring of neural activity. Owing to their user-friendly implementation, wearable neuroimaging devices have been widely used to study the outcomes of various treatments, including the effects of exercise on parameters of drug addiction (
Electroencephalogram can be used to record brain cortical electrical activity
When neurons change their activity patterns, such as during different phases of addiction, the local hemodynamic conditions in the brain change, resulting in neurovascular coupling (
Neurovascular coupling indicates that electrical neuron signals are closely related to hemodynamic conditions. In neurovascular coupling, when neuronal activity is elevated, more oxygen is delivered to the activated brain regions, resulting in a local increase in oxygenated hemoglobin. As EEG measures neural electrical activity whereas fNIRS monitors hemodynamic activity, multimodal EEG-fNIRS recording provides a more holistic measurement, enabling a more comprehensive understanding of the effects of drug use and withdrawal on the brain (
A further benefit of wearable systems is that they can be applied without the limitations of time and location, with a minimal influence on social activity. Consequently, multimodal EEG-fNIRS systems are suitable for evaluation of the efficacy of treatment and rehabilitation. An important goal during treatment and rehabilitation is the reduction of the incidence of relapse. Craving is a key symptom that promotes relapse to drug use; thus, identifying brain activity patterns during, immediately before, and immediately following cravings could help to optimize treatment and rehabilitation programs. For example, if we could identify the neuronal signatures of craving, we could trigger a closed-loop stimulation protocol to combat these activity patterns, or provide alternative interventions. In addition, brain signal monitoring can provide more information on changes in psychological and physiological conditions than the wearables measuring ECG, HR, and GSR mentioned in section “3.2. Conventional detection techniques and emerging wearable techniques.” Notably, including as many modalities as possible in the wearables would increase their diagnostic precision.
It has been suggested that wearables could be used to track the efficacy of treatment of those with a drug addiction (
In a closed-loop system, the treatment protocol can be optimized according to the real-time signals from the neuroimaging recording. Approaches to stimulate cues to further evaluate the level of desire for the drug are introduced in section “4.1. Stimulation cues.” The EEG and fNIRS biomarkers used to identify drug addiction are presented in section “4.2. Biomarkers of EEG and fNIRS in METH addiction.”
Approaches to investigating brain signal changes influenced by drug addiction include comparing recorded signals from those suffering from addictions with control populations, or comparing the signals of those suffering from an addiction across the addiction cycle. One approach to minimize the acute and long-lasting effects of drug intake on the participants during attempts to identify brain signal biomarkers, drug-related cues can be applied as alternatives to drugs. This approach is used because the variations in autonomic nervous system and brain signals occur not only after the presentation of drug use but also after cues. To identify useful biomarkers to distinguish addiction and control groups, protocols including various stimulations to provoke cue reactivity have been proposed (
As drug-paired cues are highly salient, emotional responses are influenced when an individual receives cues (
Although cues provide an opportunity to distinguish those suffering from an addiction from healthy controls, experiments in which drug-related cues are provided to the participants have potential ethical issues. For example, the risk of relapse may increase after stimulation by drug-related cues. One alternative approach has been recorded, brain signals when participants are involved in cognitive tasks, to differentiate those suffering from an addiction from healthy controls (
Furthermore, it has been reported that without cue stimulation and cognitive testing, the resting state brain of those suffering from an addiction and healthy individuals exhibit differences (
Biomarkers of neuroimaging techniques are not only used to distinguish those suffering from drug addiction from healthy controls, but they are also widely used to evaluate the efficacy of abstinence, exercise, and medical interventions. The biomarkers of EEG signals that have been shown to characterize the brain activity of those suffering from METH addiction are listed in
Biomarkers of EEG signals in frequency and time domains on patients with methamphetamine addiction.
References | Comparison conditions | Groups for comparison | Number of the electrodes and their locations | Biomarkers | Main limitations |
Newton et al. ( |
Eye-closed resting state during abstinence | METH (with 4 days of abstinence) versus HC | 35 electrodes distributed across the scalp | Increases: delta and theta bands across the scalp | Patients with another period of abstinence can be investigated. |
Newton et al. ( |
Eye-closed and cognitive tasks | METH (with 4 days of abstinence) versus HC | 35 electrodes distributed across the scalp | Increases: theta band increases with the increasing of the reaction time of cognitive tasks | Difficult to identify that the EEG biomarkers have resulted from METH use disorder or other health issues. |
Yun et al. ( |
METH users at abstinence stage. Eye-closed resting state. | METH versus HC | 16 electrodes distributed across the scalp | Decreases: approximate entropy | Patients are separated in to high- and low-dose of METH groups by their duration of METH use, not by the cumulated dose. |
Kalechstein et al. ( |
2.5 h of neurocognitive assessment tests | METH versus HC | 35 electrodes distributed across the scalp | Increases: theta band correlation with poor performances on cognitive tasks | The changes of biomarkers along with varies abstinence time can be further studied. |
Howells et al. ( |
Resting eyes closed, eyes-open and a cognitive task | METH versus HC | 6 channels (F3, F4, C3, C4, P4, and P4) | Increases: delta/alpha ratio | Future studies are needed, including a wider variety of mental disorders in METH patients. |
Ding et al. ( |
Drug-related and neutral VR | (1) METH versus HC |
5 channels (Fpz, AF7, AF8, TP9, and TP10) | (1) Increases: beta and gamma |
Be cautious when applying the machine learning modal built from male-only patients on female patients. |
Lu et al. ( |
METH users received anaerobic resistance treatment (RT) and aerobic cycling treatment (CT). |
(1) METH with exercise (RT or CT) versus METH without any exercise |
64 electrodes distributed across the scalp | (1) Increases: absolute power of theta, alpha, and beta bands on RT group during EC; the alpha block rate on RT group during EO and drug cues |
Lack of a healthy control group. |
Minnerly et al. ( |
Eye-closed resting state | METH versus HC | 19 electrodes distributed across the scalp | Increases: delta and theta bands across scalp |
Did not apply AI algorithm to reduce the analysis workload of extensive data. |
Zhao et al. ( |
Visual stimuli (video) then eyes-closed resting state | METH users in abstinence for 1–3 months versus other abstinence lengths | 128 electrodes distributed across the scalp | Increases: beta across scalp |
No comparison with healthy control and no longitudinal measurements on the same patient. |
Shahmohammadi et al. ( |
METH users at abstinence stage. Visual stimuli (drug-related, drugs and neutral images). | METH versus HC | 32 electrodes distributed across the scalp | Increases: P300 peaks of the event-related potentials (ERP) | All METH patients had history of cigarette smoking and no healthy subject had the history. This might influence to results. |
Khajehpour et al. ( |
Visual stimuli (drugs and neutral images) after tDCS | Biomarkers mean the difference of the biomarkers of watching drug related cues and neutral cues. METE users before versus after treated with tDCS. | 62 electrodes distributed across the scalp | Increases: P3-related late positive potential (LPP) component of the ERP |
Repetitive tDCS was not applied, only a single session tDCS is conducted. |
GSR, galvanic skin response; HC, healthy control; tDCS, transcranial direct current stimulation; METH, methamphetamine.
Biomarkers of EEG signals in functional connectivity (FC) network and network topological properties on patients with METH addiction.
References | Comparison conditions | Groups for comparison | Number of electrodes and their locations | Biomarkers |
Main limitations |
Ahmadlou et al. ( |
Resting state | METH (with 1–3 weeks of abstinence) versus HC | 31 channels distributed across the scalp | Increases: CC and the CC/L of gamma band in the small world network (SWN) |
The backgrounds of the METH and HC groups may not be similar. |
Khajehpour et al. ( |
Resting state | METH (during 1–6 months of abstinence) versus HC | 62 electrodes distributed across the scalp | Increases: CC and SWI in delta and gamma frequency bands |
HC can have a smoking, drinking, or caffeine history, which may affect the results. |
Khajehpour et al. ( |
Resting state | METH versus HC | 64 channels on the overall scalp | Decreases: WPLI of beta bends | Only male patients were included. |
Shafiee-Kandjani et al. ( |
Resting eyes closed and eyes open | METH versus HC | 19 channels on occipital, temporal, frontal, and parietal lobes | Decreases: coherences of the delta and theta band on the left frontoparietal cortices (F3Fz and C3Cz) | Coherences were used to study the linear relationship of the signals. However, brain signals seem to have more non-linearity properties. |
Zhao et al. ( |
Visual stimuli (video) then eyes-closed resting state | METH users abstinent for 1–3 months versus other abstinence lengths | 128 channels distributed across the scalp | Increases: WPLI between medial prefrontal cortex and bilateral orbital gyrus in the beta band | No comparison with healthy control and no longitudinal measurements on the same patient. |
Qi et al. ( |
Resting state with eyes open. METH users in control group, dancing group, and bicycling group | METH with exercises versus control group | 64 channels distributed across the scalp | Increase: brain flexibility and network connectivity entropy |
Lack of data from healthy subjects. |
Chen et al. ( |
Resting state | METH versus HC | 64 channels distributed across the scalp | Increases: GEV of 1 microstate (customized microstate C) |
Simultaneously MRI recording will be helpful to compare with the microstates data. |
Lin et al. ( |
Resting state with eyes open; then visual stimuli of METH cues with VR | (1) METH under cues versus resting; |
32 channels on the overall scalp | (1) Increases: coverage and occurrence of microstate B, transitions of microstates B → D and D→ B pairs |
The number of microstates was limited to 4 during the analysis. Results might change with other numbers of microstates. |
CC, clustering coefficient; GEV, global explained variance; HC, healthy control; L, characteristic path length; METH, methamphetamine; MMD, mean microstate duration; SWI, small-world index; WPLI, weighted phase lag index. *The microstates discussed in this table were derived based on individual EEG recordings (
The third type of EEG biomarker can be visualized by the topography of the data (
Other EEG biomarkers represented by topography include EEG microstates that show the spatial distribution of electrical signals recorded by the electrodes over the scalp (
In the EEG spectrum analyses summarized in
Some studies had explored EEG biomarkers when subjects had their eyes closed but were not asleep; this is done to reduce the disturbance due to non-task related visual stimuli (
In addition to the EEG signal, FC is often studied in the resting state as well. In the task state, the connectivity needs to be analyzed in every pair of channels at every point of interest, resulting in a heavy computational load. This is because the brain is engaged in various tasks at different stages along with the task. Only the data of a selected resting time period is calculated in the resting state. For this reason, in the task state, only the channels of interest are often analyzed to reduce computational load (
An fNIRS device is easy to wear without time-consuming preparation such as is required for EEG gel electrodes. Therefore, fNIRS devices have been widely used to study the effects of exercise on patients with METH addiction during abstinence. Optodes of fNIRS are often mounted on the prefrontal cortex and the motor cortex to study the influence of METH on decision-making as well as on cognition and motion abilities. The biomarkers of fNIRS signals are listed in
Biomarkers of fNIRS signals recorded on patients with methamphetamine addiction. The biomarkers include variation of hemoglobin concentration and functional connectivity (FC).
References | Comparison conditions | Groups for comparison | Locations |
Biomarkers | Main limitations |
Bu et al. ( |
METH users during resting and exercise: spinning training and strength training | (1) After exercises versus resting; |
Prefrontal cortex: 8, |
(1) Increase: △[OxyHb], wavelet phase coherence (WPCO) at frequency intervals II and IV |
Longitudinal recordings can be carried out to explore the changes in the biomarkers. |
Bu et al. ( |
METH users during resting and exercise: kick boxing | (1) METH group versus HC at resting and training states; |
Prefrontal cortex: 8, |
(1) Decrease: effective connectivity (EC) of some pair of channels |
Only signals when eyes closed were analyzed. |
Wang et al. ( |
METH users exercising; then visual stimuli of images with food | After exercises versus control group (no exercise) when receiving cues with high-calorie food | OFC: 4, |
Increase: △[OxyHb] of some channels at OFC | No measurements on healthy subjects for comparison. |
Zhou et al. ( |
METH users exercising: dancing or treadmill; then visual stimuli of images with food | After versus before treadmill training when receiving cues with high-calorie food | OFC: 4, |
Decrease: △[OxyHb] of one channel at left DLFPC | fNIRS cannot provide hemodynamic information in the deep brain. |
Tao et al. ( |
METH users exercising: dancing or cycling, then visual stimuli of images which caused negative emotions | After versus before dancing, when receiving cues with negative images | OFC: 4, |
Decrease: △[OxyHb] of one channel at DLFPC | No measurements on healthy subjects for comparison. |
Gao et al. ( |
METH users exercising: cycle ergometer with moderate or high intensity | METH with high intensity versus moderate intensity of exercises | OFC: 4, |
Increase: △[OxyHb] at PFC and DLFPC, FC of left DLPFC and OFC | No measurements on healthy subjects for comparison. |
Qi et al. ( |
METH users exercising: VR cycling, then visual stimuli of images with drug-related and neutral cues | (1) Drug-related versus neutral cues; |
DLPFC: 8, |
(1) Increase: △[OxyHb] at OFC and DLPFC; |
No measurements on healthy subjects for comparison. |
Qi et al. ( |
METH users exercising: VR cycling. Before and after the exercise session, a cognitive task (Stroop task) was carried out | (1) After versus before exercise during cognitive task; |
DLPFC: 8, |
(1) Increase: |
No measurements on healthy subjects for comparison. |
Gu et al. ( |
Patients with various drug addiction drug-related cues | METH group versus heroin group and mixed group | DLPFC: 16, |
Increase: activation of OFC | Limited number of subjects. |
FC, functional connectivity; Δ[OxyHb], concentration variation of oxygenated hemoglobin. *DLPFC, dorsolateral prefrontal cortex; FEF, frontal eye field; FPA, frontopolar cortex; M1, primary motor cortex; OFC, orbitofrontal cortex; PM, pre-motor cortex; S1, primary somatosensory cortex; SMA, supplementary motor cortex; VLPFC, ventrolateral prefrontal cortex. †Frequency intervals of fNIRS signals: interval I: 0.6–2 Hz, interval II: 0.145–0.6 Hz, interval III: 0.052–0.145 Hz, and interval IV: 0.021–0.052 Hz. ‡Network efficiency metrics of the small world properties: clustering coefficient (Cp), characteristic path length (Lp), nodal efficiency (Enodal), network global efficiency (Eglobal), and local efficiency (Elocal).
The fNIRS studies in individual with METH addiction were all conducted in the past 3–4 years. Interestingly, all these studies were carried out in China, specifically by the groups of Dong (
Most of the experimental groups in the studies summarized in
Advances in neurophysiology together with the neuroimaging technologies discussed here have led to the identification of some mechanisms underlying METH addiction disorders. Many studies have suggested that impaired self-control, irritability, compulsive consumption, etc., are caused by dysregulation and malfunction of specific brain circuits. Traditional pharmacotherapy, one of the most commonly applied interventions, can be viewed as a type of neural circuit modulation. However, traditional interventions lack spatial and temporal specificity of action. Neuromodulation, a novel approach that can modulate brain activity with spatiotemporal precision, has shown efficacy and is a promising treatment for addiction disorders (
Common neuromodulation techniques:
Transcranial direct current stimulation uses a constant low-intensity current that passes through two electrodes attached to the scalp of the participant to modulate neural activity. During tDCS modulation, a current flows between the electrodes and passes through the brain. A positive anodal current is generally considered to depolarize the neurons, thereby increasing cortical excitability and behaviors associated with the cortical region under the electrode. On the other hand, a negative cathodal current hyperpolarizes neurons, thereby inhibiting action potentials and behaviors in the corresponding cortical region.
A standard apparatus for tDCS stimulation, as shown in
The dorsolateral prefrontal cortex (DLPFC) is the area most selected for tDCS stimulation of subjects with METH addiction, as dysfunction in this area has been reported frequently among these individuals. Moreover, DLPFC can also be easily targeted in a non-invasive fashion.
Examples of tDCS treatments in patients with methamphetamine addiction.
References | Number of subjects METH patients | Treatment sessions | Stimulation parameters (current, duration, and location) | Effects | Main limitations |
Shahbabaie et al. ( |
30 males | 3 sessions. At least 72 h between two sessions. | 2 mA, 20 min. Anode: F4 (right); cathode: contralateral supraorbital area. | Reduced craving at resting state. Increased craving during meth-related cue exposure. | The effects might be transient. Long-term effects need to be explored. |
Shariatirad et al. ( |
1 male | 5 sessions a week, for 4 weeks. During 6-month follow-up, booster tDCS on days 67, 70, 72, and 88. | 2 mA, 20 min. Anode: right DLPFC; cathode: over right arm. | Reduced drug cravings as measured by DDQ and LDQ. | This is a case report. |
Shahbabaie et al. ( |
15 males | Two separate days, one-week washout period. | 2 mA, 20 min. Anode: F4 (right); cathode: F3 (left). | Significant decrease of craving after tDCS, modulation of DMN, ECN, and SN. | A limited number of subjects. |
Anaraki et al. ( |
30 males | 5 sessions. | 2 mA, 20 min. Anode: F4 (right); cathode: F3 (left). | Cue-induced cravings reduced significantly, no significant change in instant cravings. | Lack of longitudinal recordings to analyze the long-term effect of tDCS. |
Xu et al. ( |
75 females | CCAT + tDCS, 5 sessions per week, for 4 weeks. | 1.5 mA, 20 min. Anode: F4 (right); cathode: F3 (left). | Reduced cue-induced cravings. | Lack of longitudinal recordings to analyze the long-term effect of tDCS. |
Jiang et al. ( |
45 males | 5 days daily. | 2 mA, 20 min. Anodal: F4 (right); cathode: F3 (left). | Counterproductively increased impulsivity. | No simultaneous neuroimaging signals to provide the real-time effect of tDCS. |
Khajehpour et al. ( |
42 males | 1 | 2 mA, 20 min. Anode: F4 (right); cathode: F3 (left). | Mitigated initial attention bias but not sustained motivated attention to METH related stimuli. | Repetitive tDCS was not applied. Only a single-session tDCS is conducted. |
CCAT, computerized cognitive addiction therapy; DDQ, Desire for Drug Questionnaire; DLPFC, dorsolateral prefrontal cortex; DMN, default mode network; ECN, executive control network; LDQ, Leeds Dependence Questionnaire; METH patients, patients who were addicted to methamphetamine; SN, salience network.
The most common electrode locations in these studies are F4 (right DLPFC) for the anodal and F3 (left DLPFC) for the cathode electrodes. In these other studies, only two of the studies listed in
Transcranial magnetic stimulation techniques use a strong electrical current through an electromagnetic coil to generate magnetic pulses (
Examples of TMS treatments for methamphetamine addiction.
References | Groups for comparison | Treatment sessions | Brain area, coil type, and stimulation parameters (frequency, intensity, total number of pulses, and duration of treatment) | Effects | Main limitation |
Li et al. ( |
10 METH |
One session: 15 min of sham and real TMS separated by 1 h. | Left DLPFC, figure-of-eight, 1 Hz rTMS, 100% rMT, 900 pulses, 15 min. | Increase: cue-induced craving in METH. | The first studies to explore the TMS effect on METH addiction. Many stimulation parameters can be further optimized. |
Liang et al. ( |
50 males (1–15 days of abstinence). Real versus sham TMS. | 5 days treatments, then 2 days of rest, followed by another 5 days of treatments. | Left DLPFC, 10 Hz rTMS (5 s on and 10 s off), 100% rMT, 2,000 pulses, 10 min. | Decrease: craving and withdrawal symptoms. | The long-term effect of rTMS needs to be further explored. |
Chen et al. ( |
74 METH, separated into 3 real (A, B, and C) and 1 sham TMS. | One session/day and 5 days/week, in total 10 sessions over 2 weeks. | (A) Left DLPFC, figure-of-eight, 2 s on and 8 s off iTBS, 100% rMT, 900 pulses, 5 min; |
Decrease: cue-induced craving for all three groups. Group 3 was most effective. | The long-term effect of treatment needs to be further explored. |
Zhao et al. ( |
83 METH, separated into 3 TMS groups (A, B, and C) | Twice daily over 5 days for a total of 10 sessions. | (A) Left DLPFC, figure-of-eight, 2 s on and 8 s off iTBS, 70% rMT, 600 pulses, 3 min; |
Decrease: cue-induced cravings for groups (A) and (B). | The long-term effect of treatment needs to be further explored. |
Wang et al. ( |
66 METH (within 3 months of detoxification). Real and sham TMS. | 5 days/week, 20 sessions. | Left DLPFC, figure-of-eight, 2 s on and 8 s off iTBS, 100% rMT, 600 pulses, 3 min. | Decrease: cue-induced cravings. | The long-term effect of treatment needs to be further explored. |
Liu et al. ( |
20 male METH, separated into 2 TMS groups. | First 10 days daily, then on days 15 and 20. | (A) Left DLPFC, circular, 10 Hz rTMS (5 s on and 10 s off), 100% rMT, 2,000 pulses, 10 min; |
Decrease: cue-induced craving for both groups. | iTBS had a much shorter stimulation time compared to rTMS, which might affect the results. |
Wen et al. ( |
15 female METH. Real versus sham TMS. | Two separate sessions within 1 week | Left DLPFC, figure-of-eight, 2 s on and 8 s off iTBS, 80% rMT, 1,800 pulses, 10 min. | Decrease: frontal EEG theta/beta ratio during cue-related VR scenes. | The long-term effect of treatment needs to be further explored. |
cTBS, continuous theta burst stimulation; HC, healthy control; iTBS, intermittent theta burst stimulation; METH, individuals with methamphetamine addiction; rMT, resting motor threshold; vmPFC, ventromedial prefrontal cortex; VR, virtual reality. *Individuals with current methamphetamine dependence and non-treatment seeking.
All studies in
The time when the VAS and other physiological questionnaires (if available) were measured after TMS treatment for the studies is listed in
Another potential confound of the studies being discussed is that the patients population within a given study is either all male, or all female. This is because of a single-gender policy that applies at the rehabilitation centers. With the wearable closed-loop system we propose in the following section, assessing potential effects of TMS in both males and females would be easier to implement without these limitations.
In section “4. Biomarkers of neuroimaging techniques,” we summarized the evidence that neuroimaging biomarkers can be used to distinguish patients with METH addiction from healthy individuals and evaluated the efficacy of various types of treatment, such as exercise training and neuromodulations. Furthermore, in section “5. Neuromodulation treatments for METH addiction,” we discussed the results of studies that have used TMS on the DLPFC to reduce cravings resulting from drug-related cues in METH user groups. However, these results were based on offline signal processing, which does not the variations of brain activity during treatment in real time, which is essential for closed-loop therapeutics. In addition, bulky TMS systems hinder wider applications to increase the efficacy of the therapy. We propose a wearable closed-loop neuromodulation to efficiently treat METH addiction (
This system consists of multimodal EEG-fNIRS combined with TMS (
Proposed intelligent closed-loop TMS neuromodulation to treat methamphetamine addiction. It uses EEG-fNIRS measurements and is composed of three main parts: a real-time brain signal monitoring interface, an artificial intelligence signal processing block, and a customized neuromodulation system. Cp and CC, clustering coefficient; cTBS, continuous theta burst stimulation; Eglobal, network global efficiency; Elocal, local efficiency; Enodal, nodal efficiency; ERP, event-related potential; FC, functional connectivity; GEV, global explained variance; iTBS, intermittent theta burst stimulation; Lp and L, characteristic path length; rTMS, repetitive TMS; ML, machine learning; Δ[OxyHb], concentration variation of oxygenated hemoglobin; SWI, small-world index; WPLI, weighted phase lag index.
Limited studies have combined multimodal EEG and fNIRS for METH-related applications. Chen et al. used concurrent EEG-fNIRS monitoring to evaluate the influence of aerobic exercise on patients with METH addiction during cognitive tasks (
To demonstrate the efficacy of TMS treatment, neural signal analyses are carried out offline. Thus, the alternations in brain signals resulting from the stimulation cannot be identified in real time. Given the functionality of microchips for use in small systems, algorithms to detect biomarkers in real time can be embedded in a custom chip of a wearable system (
Another limitation is that when ML models are designed for METH addiction applications, the performance of classification for different types of drugs or different periods of METH use is the focus, rather than the features used in stratification and the links between neurological systems. This limits our understanding of how these algorithms work, and thus, makes it hard to improve upon them. A more detailed interpretation of ML models would increase the confidence of clinical professionals in the classification results. Yet other limitation of existing EEG or fNIRS biomarkers is that they cannot be used to determine the stage or severity of addiction; therefore, they cannot be used to evaluate treatment effects during the therapy or as features to predict the possibility of relapse. Until now, the most used approach for stratifying the severity of addiction has been questionnaires, although the scores are based on subjective answers. Using body fluidic tests, the amount of drug in one’s system can be quantified. It has been shown that the longer a person uses METH, the stronger their ΔSCR reactions are to drug-related cues (
The TMS protocol has not been extensively customized; previous studies have reported the effect of treatments in a fixed time frame owing to a pre-scheduled treatment protocol. To ensure a significant effect when comparing experimental and the controlled groups, which require analysis of performance over a fixed and limited time period, the protocols are often prescribed for a period from a few weeks to months (
With the proposed wearable closed-loop system, treatment could be conducted at any time when it is needed. First, the parameters of the stimulation protocol (amplitude, frequency, intertrain interval, and duration) and numbers of treatment sessions can be customized depending on the real-time-monitored neural signals. Ideally, the potential customized protocol would increase the efficiency of the TMS devices and the work of clinical professionals. Second, the efficacy of TMS treatment could be further improved by the use of a wearable TMS system. A wearable TMS device could provide the required stimulations without constraints of time and location. However, the wearability of the TMS system depends strongly on miniaturization of the magnetic coil and the control module. Achieving sufficient stimulation voltage for neuromodulation applications is an important issue to be addressed in the development of a compact wearable TMS system (
The treatment often has a predefined protocol in a rehabilitation center for drug abstinence. Patients receive a fixed amount of therapy in a fixed period. In addition, the effect of the treatment is difficult to quantify. This condition limits the flexibility of customizing the treatment protocol based on the latest conditions. One scenario of applying the closed-loop system in proposed.
The patients are recruited using EEG-fNIRS measurements to record the baseline brain signals. During the recording, the patients will be guided to watch the cues, such as drug-related pictures and videos, on a screen and rate the level of craving after seeing the cues. The differences of the various parameters of EEG and fNIRS signals from METH disorder users and healthy participants are analyzed. Moreover, various recorded signals when receiving neutral or drug-related cues are investigated. Brain signals’ features, found explicitly in patients receiving drug-related cues, are defined as biomarkers. During biomarker identification, no real addictive stimulus (drug) is required to arouse the thoughts of craving.
During daily life, no matter in the rehabilitation centers or not, patients wear the EEG-fNIRS cap as frequently as possible. Twenty-four hours per day would be optimal. Theoretically, when the users receive anything related to their previous experiences of METH use, the brain signals alter, and specific signals containing biomarkers can be detected. The potential cues can be a wide variety, such as the environment similar to where the users used METH, the people who look identical to who the users had METH with, and the items with similar shapes or functions as the tools used for METH. The advantages of the real-time signal processing chips are that the biomarkers can be extracted and compared with baseline biomarkers at every moment. As soon as the results of biomarkers comparison can reveal the increasing of craving, an alarm will be sent to the TMS module to active a treatment. Moreover, with the low-power and high-performance real-time processing chips, multimodal biomarkers can be analyzed without delay.
If this is the first time the TMS is activated, the protocol applied by previous publications can be used. During the TMS planned period, the biomarker changes are identified when being exposed to the cues which bring more craving, seeing real drug, for example. If a second round of TMS can be conducted after the planned stimulation period and the variance of the biomarkers before and after the treatment is not apparent, this can be a sign to reduce the TMS treatment intensity or even stop the stimulation (
With this wearable closed-loop system, many research questions which were listed in the limitations of previous published work (
The summary of the limitations of literatures listed in
The wearable closed-loop system we propose is highly accessible and user-friendly, and can be easily applied to a broader range of subjects. This means measurements on both genders and on both healthy and used disordered. Moreover, longitudinal recordings to track the long-term effect of the TMS treatment can be investigated. This system can provide a quantitative evaluation of the craving level in real-time. This is helpful for clinical doctors to adjust the therapy plan in the rehabilitation centers. Further quantitatively support the decision to leave the rehabilitation centers in advance or extend the stay. Furthermore, subjects not in the rehabilitation center, meaning those still using METH and those leaving the rehabilitation center, can be easily monitored. All scenarios, besides the rehabilitation centers, exit real METH. Therefore, the risk of relapse can be higher, which might result in more effective treatment of the system. With chips that can powerfully analyze the signals using ML algorithms in real-time, multimodal biomarkers can be analyzed efficiently.
We intend this review to provide a foundation for our current understanding of the potential for wearable closed-loop neuromodulation for treating methamphetamine addiction, which hopefully could be used for other addictive disorders. Here we detail wearable technologies and how those could be interfaced with neuroimaging techniques to understand how brain signals may relate to and influence biological signatures identified using the wearables. This information would be required to design closed-loop stimulation parameters that could be applied to patients. Our goal is that this review will provide a state of the field and a clear set of questions and next steps, including barriers that need to be overcome, to design such systems, which we believe hold great promise.
To boost the effectiveness of TMS treatment for METH addiction, we propose the concept of a wearable closed-loop neuromodulation with three main parts: a real-time brain signal monitoring system, an artificial intelligence signal processing system, and a customized neuromodulation system. In this review, we have summarized the research findings relevant to the essential modules required to achieve a wearable system that can be used to efficiently treat addiction, including biomarkers of EEG and fNIRS signals in patients with METH addiction, ML algorithms that can identify METH addiction, and applied TMS protocols to treat the addiction. Moreover, various cues that can be used to induce the desire for METH in validation experiments have been introduced. This novel approach currently focuses on METH but could be applied to other substance addictions in the future.
Y-HC, JY, HW, and MS: conceptualization and writing—review and editing. Y-HC and JY: data curation. Y-HC, JY, and KB: writing—original draft preparation. Y-HC: visualization. MS: supervision, project administration, and funding acquisition. All authors contributed to the article and approved the submitted version.
This study was supported by the Westlake University (No. 041030080118) and Zhejiang Key R&D Program (No. 2021C03002).
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.