Double-Blind 2-Site Randomized Clinical Trial of Neurofeedback for ADHD
Principal Investigator(s): View help for Principal Investigator(s) L Gene Arnold, The Ohio State University; Roger Debeus, University of North Carolina-Asheville
Version: View help for Version V1
Version Title: View help for Version Title OSU/UNCA ICAN Research Project
| Name | File Type | Size | Last Modified |
|---|---|---|---|
| Data_Files | 03/26/2024 04:32:PM | ||
| EEG-Files | 04/25/2024 02:26:PM | ||
| ICAN-MetaData | 03/29/2024 10:39:AM |
Project Citation:
Arnold, L Gene, and Debeus, Roger. Double-Blind 2-Site Randomized Clinical Trial of Neurofeedback for ADHD . Ann Arbor, MI: Inter-university Consortium for Political and Social Research [distributor], 2024-04-26. https://doi.org/10.3886/E198003V1
Project Description
Summary:
View help for Summary
Importance: The first large
double-blind randomized controlled trial of neurofeedback (EEG biofeedback) for
attention-deficit/hyperactivity disorder (ADHD).
Objective: Determine how much of reported neurofeedback benefit is specific to deliberate continuous brain-wave-contingent reinforcement and how enduring the effect is.
Design: Double-blind trial with randomization 3:2 to active neurofeedback vs. a yoked control of equal duration/intensity/appearance. Assessments at baseline, mid-treatment, treatment-end, and 6-, 13-, and 25-month follow-ups. Treatment was completed June 2018 and 13-month follow-ups June 2019.
Setting: Two sites: An Ohio university center and a North Carolina neurofeedback clinic.
Participants: Children age 7-10 with moderate/severe ADHD and EEG theta-beta power ratio>4.5. Mainly middle/upper class; 78% boys; 36% inattentive, 64% combined. Of 329 screened, 305 passed categorical and dimensional ADHD criteria; 235 had Theta-Beta Ratio.>4.5; 144 completely eligible, randomized; 2 withdrew early, leaving 142 as modified ITT population: 84 active, 58 controls.
Interventions: Neurofeedback (38 sessions at 3X/week) to reduce theta and increase beta power, using systematic reinforcement (game- and money-points) for meeting individualized manually set thresholds. The yoked-control reinforcements were based on pre-recorded EEGs with participant’s live artifacts superimposed to maintain blind. Conditions were programmed via internet by an off-site statistician-guided co-investigator.
Main Outcomes/Measures: Primary outcome was composited parent- and teacher-rated inattention at treatment end and follow-ups.
Results: Both neurofeedback and controls had large pre-post treatment-end benefits (d=1.5, p<0.001) for parent/teacher-rated inattention but were not significantly different. Hyperactive-impulsive symptoms and impairment showed similar results. Responders (CGI-I=1-2) were 61% of NF and 53% of Controls. Adverse events were proportional. 13-month follow-up found further nonsignificant improvement of inattention from treatment end for NF (d=0.1), but not controls (d= -0.07), with remission rates 39.7% for NF vs. 18.5% for Controls (p=.011).
Conclusions: This study failed to support a significant specific effect of deliberate continuous EEG-power-contingent reinforcement post-treatment but demonstrated large improvements in inattention for both neurofeedback and control. After 13 months, benefit was maintained, with small further improvement for neurofeedback, which had twice the remission rate of controls. These results suggest strong non-specific lasting effects and the possibility of delayed specific effect of TBR neurofeedback, being checked in 25-month follow-ups.
Objective: Determine how much of reported neurofeedback benefit is specific to deliberate continuous brain-wave-contingent reinforcement and how enduring the effect is.
Design: Double-blind trial with randomization 3:2 to active neurofeedback vs. a yoked control of equal duration/intensity/appearance. Assessments at baseline, mid-treatment, treatment-end, and 6-, 13-, and 25-month follow-ups. Treatment was completed June 2018 and 13-month follow-ups June 2019.
Setting: Two sites: An Ohio university center and a North Carolina neurofeedback clinic.
Participants: Children age 7-10 with moderate/severe ADHD and EEG theta-beta power ratio>4.5. Mainly middle/upper class; 78% boys; 36% inattentive, 64% combined. Of 329 screened, 305 passed categorical and dimensional ADHD criteria; 235 had Theta-Beta Ratio.>4.5; 144 completely eligible, randomized; 2 withdrew early, leaving 142 as modified ITT population: 84 active, 58 controls.
Interventions: Neurofeedback (38 sessions at 3X/week) to reduce theta and increase beta power, using systematic reinforcement (game- and money-points) for meeting individualized manually set thresholds. The yoked-control reinforcements were based on pre-recorded EEGs with participant’s live artifacts superimposed to maintain blind. Conditions were programmed via internet by an off-site statistician-guided co-investigator.
Main Outcomes/Measures: Primary outcome was composited parent- and teacher-rated inattention at treatment end and follow-ups.
Results: Both neurofeedback and controls had large pre-post treatment-end benefits (d=1.5, p<0.001) for parent/teacher-rated inattention but were not significantly different. Hyperactive-impulsive symptoms and impairment showed similar results. Responders (CGI-I=1-2) were 61% of NF and 53% of Controls. Adverse events were proportional. 13-month follow-up found further nonsignificant improvement of inattention from treatment end for NF (d=0.1), but not controls (d= -0.07), with remission rates 39.7% for NF vs. 18.5% for Controls (p=.011).
Conclusions: This study failed to support a significant specific effect of deliberate continuous EEG-power-contingent reinforcement post-treatment but demonstrated large improvements in inattention for both neurofeedback and control. After 13 months, benefit was maintained, with small further improvement for neurofeedback, which had twice the remission rate of controls. These results suggest strong non-specific lasting effects and the possibility of delayed specific effect of TBR neurofeedback, being checked in 25-month follow-ups.
Funding Sources:
View help for Funding Sources
United States Department of Health and Human Services. National Institutes of Health. National Institute of Mental Health (NCT02251743)
Scope of Project
Subject Terms:
View help for Subject Terms
children
Geographic Coverage:
View help for Geographic Coverage
Central Ohio and Ashville, North Carolina
Time Period(s):
View help for Time Period(s)
10/1/2017 – 9/30/2021
Universe:
View help for Universe
children under the age of 12 who have been diagnosed with ADHD.
Data Type(s):
View help for Data Type(s)
clinical data;
experimental data;
observational data
Collection Notes:
View help for Collection Notes
The design
was a 2-site, parallel-group, double-blind randomized comparison of active NF
to “sham” NF (Control) for up to 38 treatments in a 14-week period, with 6-, 13-,
and 25-month FU, in children with ADHD and TBR>4.5. At each of two sites, consented eligible
children were randomized to NF vs. Control in a 3:2 ratio, stratified by site and current ADHD
medication. Randomization was supervised by the study biostatistician
and implemented remotely in the software
by an unblinded co-investigator who had no contact with participants. The
follow-ups were considered an integral part of the RCT because enduring benefit
was important to justify the expense and effort of labor-intensive NF.
Treatment consisted of downtraining theta power and uptraining beta power for 38 active NF treatments vs. 38 control treatments. The yoked-control condition used pre-recorded EEGs instead of the live EEG to determine rewards. Treatment was conducted with EEGer NF software (EEG Software LLC, Northridge, CA), using a ProComp Infiniti (Though Technology, Toronto) or Atlantis (BrainMaster, Bedford, OH) amplifier. EEGer was configured for training four EEG parameters: preventing high overall amplitudes and gross movement (1-60 Hz), reducing theta (4-8 Hz), increasing beta (13-21 Hz), and reducing muscle movements (EMG, 45-60 Hz). Reinforcement consisted of points that could be cashed for money. Each session, the child participated in 5 training periods (blocks). Each period lasted 5 minutes at the beginning and gradually increased to 9 minutes per period in later sessions. Details of the reinforcement algorithm can be found in the supplementary materials.
Mid-treatment Decision about Continuing: Review of prior studies found that participants who did not show at least 10% improvement in parent/teacher ratings by 19 treatments did not respond after 38 treatments 17,22. Therefore, those who did not show a 10% improvement in the average of parent and teacher ratings at the midpoint assessment exited neurofeedback per protocol –in a blinded fashion, irrespective of assigned group- but were retained for assessments.
Measures The standardized diagnostic interview at screening and follow-ups was the Children’s Interview for Psychiatric Syndromes-child (ChIPS) and -parent (P-ChIPS) versions 23. Repeated outcome measures (unmedicated at major assessments) included: Parent-rated scales: Conners -3rd Edition: Long Version (C-3:P, 24) and Functional Assessment Checklist (FAC; 25). Teacher-rated scales: Teacher versions of the C-3 (C-3T) and FAC for Teachers (FACT; 26). We recruited from August through February so the child could have the same teacher for baseline, mid-treatment, and end-of-treatment assessments in the 3-month treatment. DSM-5 ADHD symptoms from the parent and teacher C-3 were used for the severity inclusion criterion (1.5 SD above norm). Clinician & staff ratings: Clinical Global Impression (CGI). Adverse effects and changes in concomitant treatments: Adverse Events Tracking Form and Concomitant Treatment/Education Services Form. Blinding Checks. Consumer Satisfaction and Blinding Questionnaire (parent and child) and trainer Blinding Questionnaire measured parent, child, and trainer blinding to assigned treatment. Clinical response and remission: Response was defined as CGI-Improvement (CGI-I) rating of 1 or 2. Remission (loss of diagnostic severity) was defined in two ways: as a CGI-severity rating of 1 or 2, and an ADHD symptom rating item mean <1.003,27.
Treatment consisted of downtraining theta power and uptraining beta power for 38 active NF treatments vs. 38 control treatments. The yoked-control condition used pre-recorded EEGs instead of the live EEG to determine rewards. Treatment was conducted with EEGer NF software (EEG Software LLC, Northridge, CA), using a ProComp Infiniti (Though Technology, Toronto) or Atlantis (BrainMaster, Bedford, OH) amplifier. EEGer was configured for training four EEG parameters: preventing high overall amplitudes and gross movement (1-60 Hz), reducing theta (4-8 Hz), increasing beta (13-21 Hz), and reducing muscle movements (EMG, 45-60 Hz). Reinforcement consisted of points that could be cashed for money. Each session, the child participated in 5 training periods (blocks). Each period lasted 5 minutes at the beginning and gradually increased to 9 minutes per period in later sessions. Details of the reinforcement algorithm can be found in the supplementary materials.
Mid-treatment Decision about Continuing: Review of prior studies found that participants who did not show at least 10% improvement in parent/teacher ratings by 19 treatments did not respond after 38 treatments 17,22. Therefore, those who did not show a 10% improvement in the average of parent and teacher ratings at the midpoint assessment exited neurofeedback per protocol –in a blinded fashion, irrespective of assigned group- but were retained for assessments.
Measures The standardized diagnostic interview at screening and follow-ups was the Children’s Interview for Psychiatric Syndromes-child (ChIPS) and -parent (P-ChIPS) versions 23. Repeated outcome measures (unmedicated at major assessments) included: Parent-rated scales: Conners -3rd Edition: Long Version (C-3:P, 24) and Functional Assessment Checklist (FAC; 25). Teacher-rated scales: Teacher versions of the C-3 (C-3T) and FAC for Teachers (FACT; 26). We recruited from August through February so the child could have the same teacher for baseline, mid-treatment, and end-of-treatment assessments in the 3-month treatment. DSM-5 ADHD symptoms from the parent and teacher C-3 were used for the severity inclusion criterion (1.5 SD above norm). Clinician & staff ratings: Clinical Global Impression (CGI). Adverse effects and changes in concomitant treatments: Adverse Events Tracking Form and Concomitant Treatment/Education Services Form. Blinding Checks. Consumer Satisfaction and Blinding Questionnaire (parent and child) and trainer Blinding Questionnaire measured parent, child, and trainer blinding to assigned treatment. Clinical response and remission: Response was defined as CGI-Improvement (CGI-I) rating of 1 or 2. Remission (loss of diagnostic severity) was defined in two ways: as a CGI-severity rating of 1 or 2, and an ADHD symptom rating item mean <1.003,27.
Methodology
Collection Mode(s):
View help for Collection Mode(s)
coded video observation;
cognitive assessment test;
computer-assisted self interview (CASI);
face-to-face interview;
on-site questionnaire
Scales:
View help for Scales
- CGI-I (improvement) -- Clinical Global Impressions Scale
- CGI-S (severity) -- Clinical Global Impressions Scale
- CHIPS -- Coping Health Inventory for Parents
- CHIPS Interview (Performed by Clinician)
- Conners - Parent
- Conners -Teacher
- CSHQ -- Children's Sleep Habits Questionnaire -- Parent
- FAC - Parent --Functional Assessment Checklist, Parent Completed
- FAC - Teacher -- Functional Assessment Checklist, Teacher Completed
- NORS -- Neurofeedback Observer Rating Scale
- PPDS -- Modified Petersen Pubertal Development Scale
- SBQ -- Suicide Behaviors Questionnaire
- SCT-23 -- Sluggish Cognitive Tempo
- TMT -- Time Math Test
- WASI -- Wechsler Abbresviate Scale Intelligence
- WIAT -- Wechsler Individual Achevement TEs
- WJ -- Woodcock-Johnson Cognitive Abilities test
Unit(s) of Observation:
View help for Unit(s) of Observation
Individuals
Related Publications
Published Versions
Report a Problem
Found a serious problem with the data, such as disclosure risk or copyrighted content? Let us know.
This material is distributed exactly as it arrived from the data depositor. ICPSR has not checked or processed this material. Users should consult the investigator(s) if further information is desired.