This is an open-access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in JMIR Research Protocols, is properly cited. The complete bibliographic information, a link to the original publication on http://www.researchprotocols.org, as well as this copyright and license information must be included.
Impulsivity and substance use disorder (SUD) are strongly interconnected, with persons scoring high on impulsivity being more vulnerable to develop substance abuse, facing more challenges for successful treatment, and being more prone to engage in criminal behavior. Studies have shown that impulsivity and craving for substances are strongly correlated. Neurofeedback is an effective treatment to reduce impulsive behavior. This study intends to determine to what extent a neurofeedback-intervention that is aimed at reducing impulsivity can also reduce levels of craving in forensic patients with SUD and comorbid Axis I and/or II diagnoses.
The main objective of this study is to investigate to what extent a reduction in impulsivity by a sensorimotor rhythm (SMR)-neurofeedback intervention will lead to a reduction in craving in a population of forensic psychiatric patients with a diagnosis of SUD.
Participants will be male SUD patients with various comorbidities residing in an inpatient forensic treatment facility approached through treatment supervisors for participation. Participants have tested positive for drug use in the past 24 months. The study consists of 2 parts: a randomized controlled trial (RCT) and a n-of-1 clinical series. In the RCT, 50 patients will be randomly assigned to an intervention (n=25) or a control (n=25) condition. Patients in the intervention group will receive 20 SMR neurofeedback sessions aimed at reducing impulsivity; participants in the control group receive treatment-as-usual (TAU). Additionally, 4 in depth n-of-1 clinical trials will be conducted where effects of an SMR neurofeedback intervention will be compared to effects of sham neurofeedback.
Results of this study are expected by the end of 2017.
This protocol describes the design of a study testing the effects of an impulsivity-based neurofeedback protocol among forensic patients with SUD and various comorbidities. We expect a significant reduction in impulsive behavior, level of craving, and actual drug-use for participants receiving the SMR neurofeedback protocol. The n-of-1 approach might help to explain effects possibly found in the RCT study since it allows for a more direct focus on treatment effects by following participants more closely and thereby being able to directly attribute behavioral and neurophysiological change to the SMR neurofeedback protocol employed.
Dutch National Trial Register NTR5386; http://www.trialregister.nl/trialreg/admin/rctview.asp?TC=5386 (Archived by WebCite at http://www.webcitation.org/6nXLQuoLl)
Impulsivity has been defined as a dysfunctional trait, leading to a tendency for an individual to display behavior that is performed with little or inadequate forethought [
Several studies have demonstrated that substance use disorder (SUD) is strongly associated with elevated impulsivity scores on various measures [
The concept of impulsivity has been of particular interest in studies involving criminal offenders, as these individuals often suffer from major mental disorders and are therefore more likely to be involved in criminal acts than persons without major mental disorders [
Electroencephalographic (EEG) spectral analysis is a frequently used method to compare healthy controls with prolonged drug users by focusing on differences in the (relative) strength of naturally occurring rhythms in the EEG (in Alper et al [
Neurofeedback is an intervention that uses real-time EEG measurements and displays information about these EEG measurements back to the participant, allowing them to not only see but also change their brain electrical activity over time [
Several studies have shown neurofeedback to be a promising intervention for various disorders, ranging from SUD to attention deficit hyperactivity disorder (ADHD) [
Although the relationship between impulsivity and symptoms of SUD such as craving and actual drug use has been established, to date there is no evidence about the effects of an impulsivity-based neurofeedback protocol and its effectiveness on impulsivity and on symptoms of SUD. This study aims to examine the treatability of impulsivity with an SMR-neurofeedback intervention in a population of forensic psychiatric patients with SUD and comorbid Axis I and/or II disorders. It also aims to investigate whether a reduction of impulsivity through an SMR-based neurofeedback protocol will also result in a reduction of SUD symptoms such as craving and actual drug use.
The study will combine a randomized controlled trial (RCT) design with an n-of-1 clinical trial. The RCT allows for investigating to what extent an SMR-neurofeedback protocol can reduce craving and actual drug use by augmenting levels of impulsivity for forensic psychiatric patients at a group level. However, RCTs have several disadvantages. First, they focus on between-group differences, making it difficult to determine the exact working mechanisms of neurofeedback at the single patient level. Despite the fact that the number of studies employing neurofeedback has increased over the past 2 decades, to date the underlying working mechanisms of neurofeedback remain unclear. Success of treatment is usually determined by a reduction in subjective complaints or based on other behavioral measures, independent of patients’ responses to neurofeedback on a neurophysiological level (eg, change in mean amplitude of brain frequencies). Second, most RCTs focus on participants with single, well-defined disorders or diagnoses, making it difficult to apply previous findings to patients who have a more complex psychopathology as is usually the case in forensic patients. Third, finding a reduction in subjective complaints could partially be explained by the interaction with the person giving the treatment, as this occurs with almost all frequently given types of therapy in the psychological field [
Primary outcome variables are the degree of impulsivity as measured with the Dutch version of the BIS-11 [
Primary objective: To what extent does a reduction in impulsivity by using SMR-neurofeedback result in a reduction of core symptoms of SUD such as craving and actual drug use in a population of forensic psychiatric patients with a diagnosis of SUD?
Secondary objectives:
1. To what extent can an SMR-based neurofeedback intervention reduce levels of impulsivity as measured by BIS-11 and a cued go/no-go task in a population of forensic psychiatric patients with a diagnosis of SUD?
2. To what extent can an SMR-based neurofeedback intervention reduce levels of craving as measured by self-report questionnaire DAQ-SF (short form) in a population of forensic psychiatric patients with a diagnosis of SUD?
3. To what extent can an SMR-based neurofeedback intervention reduce actual drug use as measured with urine, saliva, or breathalyzer analysis in a population of forensic psychiatric patients diagnosed with SUD?
This study will be conducted according to the principles of the Declaration of Helsinki (version 59, Seoul, October 2008) and in accordance with the Medical Research Involving Human Subjects Act. It has been approved by the medical ethical council of Brabant, the Netherlands (study number NL46390.008.13).
This study takes place in Forensic Psychiatric Centre (FPC) Dr S van Mesdag, a maximum security inpatient forensic treatment facility in Groningen, the Netherlands. Patients in this treatment facility are male criminal offenders with at least one Axis I or II diagnosis and considered to be at risk for criminal recidivism if not treated properly. About 70% of all patients treated in this facility have a comorbid diagnosis of SUD [
A randomized controlled trial with N=50, where 25 participants are randomly assigned to treatment as usual (TAU) combined with 20 SMR-based neurofeedback sessions and 25 participants are randomly assigned to TAU only, without neurofeedback intervention. The 2 groups are compared pretreatment (T0) and posttreatment (T1) on variables linked to the research questions. Both groups will receive pre- and posttreatment measurements with an interval between T0 and T1 of approximately 10 weeks in which participants in the intervention group will receive 20 neurofeedback treatment sessions and participants in the control condition will follow TAU.
The design of this part of the study is a 2×2 design with the condition (neurofeedback vs TAU) as a between-subjects factor and time as a within-subjects factor (pre- and postintervention).
To zoom in on specific treatment effects, 4 single case studies with an A1B1A2B2design (single time series) will be conducted, of which 2 single case studies will apply an actual SMR-neurofeedback protocol and 2 single case studies will apply a sham neurofeedback training. The clinical trial will be single-blinded, indicating that participants do not know which part of the training they will receive. Participants are selected from the control group of the previously described RCT protocol who have already completed pre- (T0) and posttreatment (T1) measurements. Inclusion in the n-of-1 trial will be selective: participants with the highest scores on outcome measures on T1 of the RCT will be approached first as it is believed that these patients have the highest need for treatment. However, allocation to treatment (sham or real) will be random.
For a detailed description of this design of n-of-1 studies, see Rizvi and Nock [
To test for transient effects of the neurofeedback intervention, a follow-up measurement of resting state EEG, BIS-11, DAQ-SF, and cued go/no-go task will be performed 12 months after completing the posttreatment measures for both participants in the intervention group of the RCT and for participants in the n-of-1 clinical trial.
A power analysis calculation for the RCT using G*Power 3 (Department of Psychology) based on a 1-tailed alpha value of .05, a power value of 0.80, and an effect size (
Participants are male patients diagnosed with SUD (substance dependency or substance abuse) according to the
Participants are allowed to continue using prescription medication (as prescribed by a psychiatrist or general physician of the treatment facility) but are required to inform researchers of any medication they are currently using or any change in medication during treatment with neurofeedback.
Recruitment will start with the selection of patients for the RCT part of the study. Participants are approached through treatment supervisors for participation. Treatment supervisors are informed about the general inclusion criteria for this study. Out of all participants that meet the requirements, a random sample of 50 will be drawn and randomly assigned to 1 of the 2 conditions (intervention and control). Prior to participation in the trial all participants are asked to provide written consent. If at this point a participant chooses not to participate in the trial, this will be coded as a nonresponse. Missing numbers of participants will be complemented by randomized allocation of other suitable participants who are willing to participate in order to guarantee the sample size. Once all patients for the RCT have been recruited, recruitment for the n-of-1 clinical trial will begin. All participants will receive a financial reward after completing pre- and posttreatment measurements.
Participants will undergo a 21-channel EEG measurement with Nexus-32 hardware and Biotrace+ software (Mind Media BV). The EEG will be collected from 19 standard 10/20 positions [
For participants in the neurofeedback group, a 1-minute baseline recording over 3 conditions will be conducted before start of the first neurofeedback session and after the last session. EEG signal will be recorded from electrode position Cz against a right ear mastoid reference across the conditions (1) eyes open, (2) eyes closed, and (3) cognitive task (where participants are instructed to solve simple mathematical calculations). These measurements will be used to determine neurofeedback threshold values and to assess change in mean magnitude of frequency bands before and after neurofeedback training.
The Dutch version of the BIS-11 (eleventh edition) [
The cued go/no-go task is a continuous performance test measuring impulse control by the ability to inhibit prepotent responses [
The DAQ-SF [
All questions of the original questionnaire are designed to measure craving purely for alcohol; however, within the treatment facility alcohol use is less common than other drug use (such as marijuana and/or cocaine). Therefore, questions from the questionnaire have been altered so they can fit any type of drug dependency. An example of this is “My desire to drink seems overpowering” which has been altered to “My desire to use drugs seems overpowering.”
The Instrument for Forensic Treatment Evaluation (IFTE) is an observational treatment evaluation instrument consisting of 22 items measuring 3 factors: Problematic behavior, protective behavior, and resocialization skills. It is scored on a 17-point Likert scale with 5 anchor points: none, rarely, sometimes, often, and always [
Drug testing is performed on a regular basis, usually once every 2 weeks. Whenever staff suspects illegal use of substances within 2 moments of drug testing, spontaneous and unexpected drug testing can be performed. Number of drug tests will be counted, as will be positive (meaning drug use in the period of time since last drug test) and negative (meaning no drug use since last testing) outcome scores. Drug testing is done in the form of urine, saliva, or breathalyzer (for alcohol use only) analysis.
Covariates are sociodemographic characteristics; specific psychopathology; duration of forensic treatment; actual drug use during the past 24 months (or as long as patients reside in the treatment facility); medication use; clinical risk assessment score (Historical/Clinical/Future-Revised, HKT-R) [
All participants already receive TAU at the moment of inclusion. They will continue to do so during the course of this trial. Type of TAU is dependent on disorder and behavior but can range from cognitive behavioral therapy, psychotherapy, and psychomotor therapy to relapse prevention treatment and can be either individual treatment or in-group treatment. Treatment can also be supplemented by medication for psychotic symptoms or depressive symptoms, for example. In some rare cases, aversion or craving reducing medication is prescribed.
Participants in the intervention condition of the RCT will receive 20 neurofeedback sessions, each lasting approximately 40 minutes. EEG magnitude is measured across delta (0.5-3.5 Hz), theta (3.5-7.5 Hz), alpha (7.5-12 Hz), beta (12-20 Hz), SMR (12-15 Hz), high beta (20-32 Hz), and gamma (32-49 Hz) frequency bands. To reduce inattention and impulsivity, a conventional neurofeedback protocol will be used that consists of suppressing theta magnitude and enhancing SMR magnitude [
For the n-of-1 design of the trial, 2 participants will receive the SMR-neurofeedback intervention and 2 participants will receive sham neurofeedback.
Real and sham neurofeedback procedures will be similar (eg, electrode position, preparation, instructions given to participants) except that for the sham neurofeedback training group, participants are instructed to enhance an irrelevant frequency band that is randomly chosen from higher beta bands (20-23 Hz, 23-26 Hz, 26-29 Hz, and 29-32 Hz). Therefore, no specific frequency band is systematically modulated and thus should not result in desired treatment outcomes. Participants will still be given positive feedback and be able to influence the video games in order to minimize possible irritation of participants.
Neurofeedback will be applied as implemented within the BrainMarker software engine (BrainMarker Device, Brainmarker BV Gulpen).
Participants will be shown simple video games implemented in the software that will provide feedback about their brain activity. During the video games, they are instructed to be attentive to the feedback (no movement/movement of objects) in the video game and to find the most successful strategy to reach the goal of the game. Example of such video games are a car moving on a road, where participants are instructed to keep the car in the right lane of the road, or a basketball court where participants are instructed to try to throw the ball in the basket. The video game–based neurofeedback rounds will last 1 minute at a time, with a short break between rounds. Also, movie-based neurofeedback will be applied. During movie-based neurofeedback participants will watch a digital video disk of their own choice and be instructed to keep the monitor as free as possible from black curtains appearing on both sides of the monitor and keep the volume of the movie at an audible level. Movie-based training will last 90 seconds at a time with a short break when necessary. Participants will receive both game- and movie-based neurofeedback in each session.
Thresholds will be set manually in a way that if a participant maintains the reinforced frequency band above a threshold for 80% of time, positive feedback will be received. To determine threshold values, mean magnitude of the baseline measurement across the 3 conditions described above will be used to roughly assess threshold values for the neurofeedback training. For each training session, mean magnitude values will be calculated for all frequencies.
All statistical analysis will be conducted using SPSS version 19 (IBM Corp). Summarizing descriptive statistics and frequency tables will be provided.
Resting state EEG data will be analyzed using custom-made Matlab R2012b scripts [
Repeated measurement with time (pre- [T0] and postintervention [T1]) as the within-subject factor and group (control vs intervention) as the between-subject factor will be conducted for the DAQ-SF, BIS-11, IFTE, and IFTE-SR. If main or interaction effects are observed, post hoc test will be used to determine which levels of the factors are explaining the observed effects. A repeated measures analysis of covariance will be conducted to examine differences in actual drug use as dependent variables to test for a moderating effect of impulsivity on craving and actual drug use.
First, a time-plot will be inspected using the autocorrelation coefficient (ie, correlogram) [
Results of all measurements will be expected by the end of 2017 and will be published in corresponding articles.
This study aims to evaluate the efficacy of an SMR-based neurofeedback treatment on reducing impulsivity in a population of inpatient forensic patients. Possible effects of a reduction in impulsivity on substance abuse will be assessed as well. We expect a significant reduction in impulsive behavior, level of craving, and actual drug use for participants receiving the SMR-neurofeedback protocol. The n-of-1 approach might help to explain effects possibly found in the RCT study since it allows for a more direct focus on treatment effects by following participants closely and thereby being able to directly attribute behavioral and neurophysiological change to the SMR-neurofeedback protocol employed. The study aims to extend previous findings on the efficacy of neurofeedback treatment in reducing impulsivity, not only by linking possible findings regarding a reduction of impulsivity to substance abuse symptoms but also by examining effects in a forensic psychiatric population with various comorbid disorders.
Studies about the efficacy of neurofeedback in a psychiatric forensic setting, in which the population is characterized by various comorbidities and various kinds of medication, are lacking. In our study, exclusion criteria are kept to a minimum to include as many participants with SUD as possible and to be able to generalize effects of an SMR-neurofeedback treatment over different types of comorbidities.
Although RCTs with a treatment and a sham treatment arm are considered the gold standard in research, conducting large trials is often times difficult in forensic settings; treatment motivation might be low for the type of patients in the treatment facility because they are placed under compulsory inpatient custody and are not seeking treatment due to inner motivation for change. In RCTs, number of participants usually has to be quite high to reach the desired effect size [
By employing an n-of-1 approach combined with an RCT, this study might help shed light on the underlying mechanisms of neurofeedback because an n-of-1 approach allows closer monitoring of treatment effects and provides valuable insight into an individual’s treatment progress that might otherwise be lost in a between-group design [
If effective, neurofeedback could be a noninvasive treatment option for the reduction of impulsivity, which may lead to a reduction in feelings of drug craving and in actual drug use. Both impulsivity and drug-seeking behavior are known to hamper treatment progress and are strongly linked to criminal behavior [
There are several important issues to consider that might influence the results. First of all, participants are not selected based on their level of impulsivity. Even though the most commonly observed disorders in the treatment facility are schizophrenia and personality disorder and both types of disorders are associated with increased impulsive behavior, not all suitable participants might show elevated levels of impulsivity. Studies have shown that although there is evidence that heightened impulsivity can be found across different types of substance use disorders, there is still substantial heterogeneity on impulsivity levels within these groups [
Also, participants will be included who have tested positive for drug use in the past 24 months. This implies that there will also be participants whose substance use disorder is in early remission. Although substance abuse–related symptoms such as craving are known to persist even after drug use is terminated, this period of time might be too long for these participants to report any craving at the moment of the administered questionnaire.
CONSORT-EHEALTH checklist V1.6.1 [
attention deficit hyperactivty disorder
Barratt Impulsiveness Scale
Desire for Alcohol Questionnaire
Desire for Alcohol Questionnaire–Short Form
Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, Text Review
electroencephalograph
Forensic Psychiatric Centre
Historical/Clinical/Future–Revised
Instrument for Forensic Treatment Evaluation
Instrument for Forensic Treatment Evaluation–Self-Report
neurofeedback
randomized controlled trial
sensorimotor rhythm
substance use disorder
treatment as usual
We would like to thank Harry Beintema and Sander van der Kroft for facilitating this study at the FPC Dr. S van Mesdag, as well as FPC De Kijvelanden for their financial support of this study. This research received no grant from any funding agency in the commercial, public, or not-for-profit sectors.
None declared.