EEG Alpha/Theta Neurofeedback to Reduce Trait Anxiety

Since the late 1960s neurofeedback treatment (NFT) has been used to treat adult individuals with anxiety disorders. Yet most related, evidence-based research studies were conducted between the mid-1970s and late 1990s. Therefore, NFT as an efficacious treatment for anxiety problems remains unclear. The literature research discloses that of few studies, most used sample sizes 10 subjects or less per experimental or control group and results have been mixed: Orne & Paskewitz (1974) and Watson & Herder (1980) found no significant effect sizes for anxiety symptom reduction and Egner, Strawson & Gruzelier (2002) did not find a significant difference in subjective activation level between alpha/theta (A/T) NFT and the sham group. On the other hand, Hardt & Kamiya (1978), Plotkin & Rice (1981), Sargunaraj, Kumaraiah, Mishra, & Kumar (1987), Rice, Blanchard & Purcell (1993), Vanathy, Sharma, & Kumar (1998), Sarkar, Rathee, & Neera (1999), Eismont Lutsyuk ,& Pavlenko (2011), and Gruzelier, Thompson, Redding, Brandt & Steffert (2013) did find a significant anxiety symptom reduction through NFT. The U.S. National Institutes of Health's National Center for Complementary and Integrative Health (NCCIH) at this time does not endorse NFT as an efficacious treatment for anxiety problems.

General hypothesis: A ten-session NFT protocol of up-training alpha (8-11 Hz) and theta (5-7.5 Hz) frequency bands lowers the subjective experience of trait anxiety and elevates mean theta and lowers alpha amplitudes in EEG recordings for individuals with moderate to high trait anxiety (while individuals' subjective experience of trait anxiety and theta/alpha ratios in the sham NFT condition will not change).

Control for treatment expectations and self-efficacy

• There will be no significant differences between the treatment and the sham group in treatment expectations / self efficacy as measured by the Stanford Expectations of Treatment Scale (SETS) and the Rotter Locus of Control Scale.
• There will be no significant differences between the treatment and the sham treatment group pre-and post treatment satisfaction as measured by assessments as measured by the SETS modified outcome scale.

Treatment effect hypotheses:

The treatment group will on average:

• show higher mean theta and higher alpha amplitudes within and between sessions than the sham group. (The mean theta/alpha ratios within a session will be calculated by adding the theta/alpha ratio of the beginning and end of the first 8-minute NFT block and the end of the second and third NFT block and dividing these four ratios by four).
• have a higher mean linear increase of theta amplitude and higher linear decrease in Alpha between the NFT session than individuals from the sham group (→ learning curves).
• feel subjectively more deactivated and calm (as assessed by Thayer's ActivationDeactivation Adjective Checklist (AD-ACL)) at the end of each session than the sham group.
• feel subjectively more deactivated and calm the more sessions of NFT they have had while the sham group will not.
• show more decreased self-rated anxiety symptoms (as measured by the the State Trait Anxiety Inventory (STAI), Beck Anxiety Inventory (BAI) and Generalized Anxiety Item Scale (GAD-7) than the sham group.
• show a higher mean linear increase of theta amplitude and lower mean alpha amplitude within and across sessions will be correlated to a higher score in deactivation and relaxation in the treatment group but not in the sham group.
• show a higher mean linear increase of theta amplitude and lower mean alpha amplitude across sessions will be correlated to a lower pre-post treatment score in the STAI, BAI, and GAD-7 inventories in the treatment group but not in the sham group, meaning=> a higher negative correlation between the STAI, BAI, and GAD-7 change scores and the mean linear increase of theta amplitude and lower mean alpha amplitude across sessions is to be expected in the treatment group than in the change group.
• have a higher alpha and theta distribution on the mini-QEEG, particularly in the Pz region.

Adults with no prior NFT experience who scored moderate to high on the State Anxiety Inventory Questionnaires (STAI-Y; scores: >66%), low on the Beck Depression Questionnaire (BDI-II; scores: <19), who may or may not have been diagnosed with an anxiety disorder, but have no other serious co-morbid mental health or neurological conditions, such as Major Depressive Disorder, Bipolar Disorder, Schizophrenia, Substance Use Disorder, epilepsy or tic disorders will be included in the study.

After pre-screening, participants will be assigned to the experimental or sham NFT conditions by randomized MATLAB protocol (seven random zero/one vectors of four integers length, equally made up of zeros and ones, to represent experimental (digit 1) and control group participants (digit 0)).

Experimental and sham group will be rewarded for uptraining the power of particular wave frequencies using auditory, visual, and number rewards. The two reward sounds (low pan flute sound and steel drum sounds) will be playing as sustained MIDI sounds with amplitude modulation and will depend on the power of the respective wave frequencies. The reward criterion will be set at 250ms with a refractory period of 200 ms between rewards. Participants will also be shown a flying dragon looped video (Vulcan flying) which becomes brighter in color when alpha and/or theta power is up as a visual reward and see their reward percentage as a number on the screen beside the looped video. Both groups will be trained in three segments of eight minutes of NFT with an up to 90s break (depending on the participants' preferences) between each of the segments.

The experimental A/T NFT group will receive ten sessions of A/T NFT - specifically up-training alpha (8-11 Hz, low pan flute sound) and theta (5-7 Hz, steel drum sound). The sham NFT group will receive ten sessions of two alternating NFT protocols. In odd-numbered sessions (1,3,5,7,9) lobeta (13-16 Hz pan flute sound) and hibeta (16-22 Hz, steel drum sound) will be uptrained and alpha (8-11 Hz), theta (5-7.5 Hz) and hibeta (23-30 Hz) are inhibited. In even numbered sessions (2,4,6,8,10) beta (13-16 Hz, pan flute sound) and hibeta (16-22 Hz, steel drum) are down-trained and alpha (8-11 Hz), theta (5-7.5 Hz) and hibeta (23-30 Hz) are inhibited.

The BrainAvatar® 4.0 suite (Brainmaster Technologies, Inc; http://www.brainmaster.com), will be used for all NFT. Pz will be selected as the EEG recording site, with A1 as the reference and A2 as the ground electrodes. The sampling rate will be set to the standard 256 Hz per second at a resolution of <0.01 Hz, and 10,000 amplification. Impedance will be set at < 10 kOhms. Filters used will be a band pass filter (80Hz high and 60 low pass) and an A/D low pass filter (allowing 0-30 Hz activity). Smoothing time will be 0.5s at 3 dB. The artifact rejection threshold will be set at 200 microV.

Statistical analysis:

The overall design of the proposed study calls for a mixed ANOVA Treatment condition (A/T vs. sham) serves as the between-subjects, and pre/post levels of anxiety, as the within-subjects factor. Treatment effectiveness (across both groups) would be reflected in a significant within-subjects effect and the efficacy of NFT specifically would be reflected in an interaction whereby significantly more improvement (i.e. lowered trait anxiety) would be found in the A/T in contrast to the Sham-Control condition. In a mixed analysis of this kind, a total of (N = 27) participants is required to detect moderate sized (f = .25) interaction effects and within-subjects effects with adequate power (.80).

Specifically:

ANOVAs

• pre-and post-treatment (before & after 10 sessions of neurofeedback) mean theta/alpha ratios between sham and treatment group
• pre-and post-session (10 sessions) mean theta/alpha ratios within individual & between sessions comparing sham & treatment group
• pre-and post-session (10 sessions) Activation Deactivation Adjective Checklist (AD-ACL) scores within individual
• pre-and post-session (10 sessions) AD-ACL scores between sessions
• pre- and post treatment STAI anxiety scores between treatment and sham group
• pre- and post treatment BAI anxiety scores between treatment and sham group
• pre- and post treatment GAD-7 anxiety scores between treatment and sham group
• pre-and post treatment mini-QEEG patterns between treatment and sham group

Correlations / regression analyses:

• individual learning curves for A/T treatment condition over ten sessions: mean theta/alpha ratios for individuals between sessions
• correlation between AD-ACL pre- and post treatment scores and theta/alpha ratio pre-and post treatment
• correlation between SETS scores pre- and post treatment scores and theta/alpha ratio pre-and post treatment
• correlation between STAI pre-and post treatment scores and theta/alpha ratio pre-and post treatment
• correlation between BAI pre-and post treatment scores and theta/alpha pre-and post treatment scores
• correlation between GAD-7 pre-and post treatment scores and theta/alpha pre-and post treatment scores
• correlation between Rotter Locus of Control scores and theta/alpha ratio pre-and post treatment

Qualitative Analysis:

• which mental strategies were used during the NFT training and as how effective different strategies were perceived at by the individual subjects
• what specific changes (if any) in thoughts, feelings and behavior in everyday life the individual subjects perceived

Inclusion criteria

• adults, age 18 and above
• scores of >66% on the State Trait Anxiety Inventory (STAI)

Exclusion criteria

• scores >19 on the Beck Depression Inventory (BDI-II)
• scores > 9 on the Patient Health Questionnaire (PHQ-9)
• comorbid diagnoses of: Major Depression, Bipolar Disorder, Schizophrenia, Substance Use Disorder, Tic disorders, Seizure Disorders