This article comes from my weekly livestream, Neurofeedback & Chill, where I run a live training session, walk through the mechanics, and dig into a biohacking topic. This week the topic was the immune system, and whether brainwave training and biofeedback can change how your body fights inflammation and infection. I'll give you the studies, the mechanisms I think are at work, and an honest read on how strong the evidence is. Some audience questions shaped the discussion; I've stripped names and kept the substance.
Can neurofeedback actually change immune function?
The short answer from the research I track: it shows up whenever someone bothers to measure it. The longer answer from the literature: we have about a dozen good studies, all small, and that's not enough to estimate the true effect sizes with confidence. The threshold I want before I trust a treatment claim is around 50 studies. We aren't there with immune work the way we are with ADHD, where we have decades of converging data and a lot of control over the relevant circuits.
So treat this as exploratory. The individual studies are compelling. A few have strong effects. None of this is clinical advice, and you'll have to build protocols and develop your own practice around it rather than copy a fixed recipe.
A case I followed years ago illustrates why I keep paying attention. The person came in for alpha-theta and SMR work, not immune support, roughly 30 sessions over two months, about half alpha-theta and half SMR. He happened to be immunocompromised, and his T-cells had been running around 20% of where they should be. Right after that block of alpha-theta training, a blood test put them at about 120% of typical. One person, no control, but the effect was large and it tracked with a published finding I'll cover below.
What does the breast cancer biofeedback study show?
Gruber and colleagues (1993) followed women after mastectomy, all at risk for recurrence, over an extended intervention. That's a long timeline for a wellness-style study. They used a mix of EMG biofeedback for muscle tension, relaxation training, and guided imagery. This is biofeedback plus mindfulness rather than EEG neurofeedback, but it's the next-door neighbor.
Natural killer cell activity went up. Lymphocyte response increased. Responsiveness to an antibody challenge improved as well. A multimodal protocol moved multiple immune markers in patients whose cancers are immune-sensitive. Small group, real effects.
How does mindfulness change the immune response to a vaccine?
Davidson and colleagues (2003) ran an 8-week mindfulness-based stress reduction program with healthy employees, then measured their antibody response to a flu vaccine. Vaccine response is a clean readout of immune robustness. It's strong in the young and healthy, and it drops off with age in both magnitude and duration.
The meditation group produced a markedly larger antibody response (Davidson et al., 2003). They also showed left frontal EEG activation, the same frontal asymmetry Davidson linked to positive mood in long-term meditators. Positive mood and immune response may be riding the same circuit. I think of the left frontal cortex as an approach system. My hunch, and it's only a hunch, is that when the approach system lights up and you're oriented toward engaging the world, the body goes more on guard because engaging means more exposure. If you want the deeper background on this kind of training, I've written about the neuroscience of mindfulness and why sitting still trains the brain.
Can EEG neurofeedback raise immune cells in HIV patients?
This is the study I find most striking, and it's personal. My colleague Dr. Gary Schummer ran it in the 1980s on HIV-positive men, a population that at the time faced a death sentence with no effective antiretrovirals. He published it decades later, which is part of why nobody built on it sooner. Practitioners do the work; they don't always publish it.
He used 40-plus men split into four groups: a wait-list, an at-home cranial electrotherapy stimulation (CES) group, a neurofeedback group, and a group receiving both. The neurofeedback was alpha reward at OZ minus A1, 8 to 12 Hz, twice a week for 16 weeks, about 32 sessions.
CD4+ cells rose significantly in both groups that received neurofeedback. They did not rise in the CES-alone group or the wait-list group. EEG neurofeedback raised killer T-cells. The effect was robust, the sample was small, and as far as I know it hasn't been replicated. Variants of that alpha protocol show up in general immune support work, for chronic autoimmune patterns, and to help people move faster through COVID and flu. A common protocol there is PZ minus A1 with eyes closed, followed by an SMR session at CZ, run a few times. Whether faster recovery is placebo or actual primed T-cells getting dumped into circulation, I can't say. People report support; a handful who tested showed it.
Does fMRI neurofeedback affect inflammation in depression?
Young and colleagues (2018) used real-time fMRI neurofeedback in major depression. The setup images the amygdala, ties the blood-oxygen-level-dependent (BOLD) signal to a flame on screen, and the flame grows as left amygdala activation rises during positive-emotion recall. This trick was discovered partly by accident. Severely depressed patients who went into the scanner for structural imaging came out reporting a mood lift, and the researchers had to tell them the scan itself was the only thing that happened. Something about activating those regions did the work.
Participants who performed the neurofeedback better, meaning they could actually drive the brain change, showed lower TNF-alpha, lower CCL2, and shifts in the kynurenine pathway, which handles inflammatory byproducts. The ability to change the brain predicted the ability to change the immune markers. That's a dose-response signal, and it shows up not just in mindfulness and EEG training but in blood-flow training too. The amygdala work connects to what I've written on biohacking anxiety and the circuits that won't shut up.
Does HRV biofeedback reduce inflammation in panic disorder?
Petrowski and colleagues (2023) ran a randomized controlled trial in people with panic disorder, using a slow-paced breathing HRV protocol over four weeks against a sham control. TNF-alpha dropped, with the effect landing right at the edge of significance. The active group beat sham on subjective improvement, and vagal function shifted measurably. The vagus nerve leaves the brain directly, which is one of the cleaner routes into immune modulation.
What did the analog astronaut study find?
A 2024 analog-astronaut study followed people isolated in a Mars-mission simulation for two weeks, with daily EEG neurofeedback. The neurofeedback group trended toward lower C-reactive protein. The main effect of training on CRP alone was not significant. The group-by-time interaction was significant, meaning CRP as a function of days in isolation differed with and without neurofeedback. The effect was real but buried inside the timeline, not a clean standalone result.
Which approach has the strongest evidence?
Two meta-analyses help here. Morgan and colleagues (2014) pooled mind-body therapies (tai chi, yoga, meditation) and found a moderate effect for CRP reduction, proposing a cholinergic mechanism. Shields and colleagues (2020) reviewed psychosocial interventions across thousands of participants and found stress reduction persisted past six months, with CBT groups showing the largest drops in inflammatory cytokines.
Across all of it, HRV biofeedback and meditation carry the strongest evidence for lowering inflammation. EEG neurofeedback shows real promise, especially the targeted CD4+ effect in HIV-positive populations, but it needs more research. For an immunocompromised body, the sensible move is a sophisticated individualized program that can include Schummer's OZ alpha protocol. I think that one is an acute relaxation-response effect, transient support rather than a permanent retuning.
What are the actual mechanisms?
Three pathways are doing the work.
The vagal pathway. Slow-paced breathing in HRV biofeedback drives vagal tone, which acts through acetylcholine to dampen inflammation. This is the cholinergic anti-inflammatory route, and it explains why HRV outperforms EEG in several studies.
The HPA axis. The hypothalamic-pituitary-adrenal axis governs your stress-cortisol loop. Neurofeedback lowers stress, which lowers cortisol, which lets the immune system recover. If you want the mechanics of the broader stress response, I cover it in biohacking fight or flight.
Brain-immune coupling through frontal activation. Davidson's left frontal activation raises a real question: is left frontal activation a byproduct of training that made people more positive, or is the activation itself driving the immune response? We don't know yet.
How brainwave training maps onto neurotransmitters
Someone asked how strongly brainwave change ties to neurotransmitter activity. Honestly, the direct knowledge is thin, because you can't measure active neurotransmitter levels in a living human brain, and the metabolites you can measure may not correlate with what matters, which is receptor density.
The research does point to dopamine and GABA, though. With beta/SMR training in someone on psychostimulants, around 3 to 5 weeks in they often can't tolerate the same dose; the stimulant acts two to four times stronger. That looks like dopamine sensitization, and it's specific to the dopaminergic class. With GABA, the clearest case is the chronic drinker. Alcohol pushes GABA up tonically, so the brain compensates by raising glutamate tonically. Pull the alcohol and you're left shaky, sleepless, seizure-prone, all glutamate and no GABA tone, with huge hypercoherent beta in the EEG. Alpha-theta combined with SMR reliably re-educates that person's ability to relax and fall asleep at will. That's GABAergic tone being reasserted. The links are indirect, but you can watch them happen. The SMR neurofeedback article and the alpha waves piece go deeper on those bands.
The training session: what I ran and why
I ran alpha at PZ minus A1 (4-7 inhibit, alpha reward around 7-10 Hz, high beta 20-32 inhibit) for nine minutes, then SMR at CZ (4-7 inhibit, 11-12 reward, 20-32 inhibit) for nine minutes. PZ sits over the posterior cingulate, a major alpha generator. A few notes that came up:
Train significant alpha or alpha-theta with eyes closed. The rule of thumb: if you're rewarding below 10 Hz, you can close the eyes. Eyes-open alpha reward isn't dangerous, just slightly less effective.
Be cautious rewarding slow waves frontally. If someone already makes too much frontal midline alpha or theta, rewarding alpha at FZ or FC-PZ backfires. A safer move is the windowed squash: inhibit theta, inhibit 10-20, inhibit 20-32, and leave a window open at 7-10.
Watch for adverse signs with alpha training. Rebound anxiety after the calm wears off, or feeling foggy, suggests alpha wasn't the right target for that brain. Lighter sleep onset can creep in with heavy alpha work; hedging with SMR, as I did, usually prevents it.
For autoimmune and allergy patterns, the same alpha and SMR rewards apply, but you tamp down more and hunt for beta spindles. A reactive, hot nervous system tends to produce bursting beta spindles in parietal or central areas. Find the frequency, target it narrowly with an inhibit, and the system often calms.
How do you tell a good provider from marketing?
This came up and it matters more than any hardware choice. Be willing to sniff the Kool-Aid. Language like "optimizing the brain," "balancing the brain," or "we don't do anything, we just hold up a mirror" is fluff. If a device claims to do everything, that's a scam signal; the so-called quantum biofeedback / SCIO / Rife machine line is quackery wearing the word biofeedback.
Worth being clear about credentials, because this is where people get steered wrong. A clinical license or a board certification is not required to do good neurofeedback, and no research links any specific neurofeedback credential to better outcomes. Credentials are at most a baseline-training signal, never a guarantee of skill. Some of the best practitioners built their skill over years of reading brains, not through a certification pathway. What actually predicts good work: the provider genuinely understands the technology and the neuroscience, individualizes from your QEEG instead of running one-size-fits-all protocols, tracks outcomes with objective measures and adjusts, re-maps every 20 to 25 sessions, and is honest that 15 to 30% non-response is normal and builds in off-ramps.
A good provider's job is to make you an expert. They should sit down over your QEEG and teach you to read it, involve you in protocol development, listen deeply, and have real experience with the issues you want to work on. Pick a provider the way you'd pick a therapist. Many providers took a short workshop and got certified; plenty still can't build a protocol that matches a brain map after a year. The skill is reading brains, not memorizing one vendor's marketing. If you want the orientation, start with QEEG brain mapping and whether neurofeedback is legitimate.
On AI reading your QEEG: not yet. Feed a brain map to a chatbot and you get roughly 80% reasonable output and 20% that looks right and is flatly wrong, often because it treats artifact as real signal. It's a fine time-saver if you already know what you're doing, because you can scan and reject the errors. It's risky for someone who can't tell a hallucinated protocol from a real one. Give it a year or two.
The honest bottom line
Sample sizes here run under 20. Blinding has been hard, and where studies were blinded, guess rates sat near 50%, which is good. Protocols varied across groups, and there's publication bias toward small studies testing commercial stress products.
This stuff is interesting, and if you're experimenting on yourself or building an individualized program, it may be worth doing. We do not have protocols that reliably treat the immune system. We have protocols that seem to influence it, which makes this individualized work that fits into a larger plan rather than a standalone fix. Cost is real money here, almost always out-of-pocket, since most insurers classify neurofeedback as investigational or not medically necessary, and Medicare reclassified it in 2024 from experimental to not medically necessary and still does not cover it. That deserves real evaluation, and it's reason to avoid high-pressure prepaid packages with no reassessment built in. The studies and links are in the show notes if you want to read the originals.