Treatment for Auditory Hallucinations in Schizophrenia Without Drugs?!?

A randomized controlled trial of rTMS in schizophrenia offers new hope

Hot off of Wednesday’s coverage of the novel oral schizophrenia treatment Cobenfy in this newsletter—hailed by my reader Chris Aiken, M.D., as “brief”—we have yet more good news. The serious exploration of transcranial magnetic stimulation for schizophrenia dates back to the late 1980s and early 1990s. I remember, when I was 20 years old, chatting with a fellow research subject at the New York State Psychiatric Institute, while we were both enrolled in clinical trials for the exploration of our illnesses. I was enrolled in a trial—that would involve a lumbar puncture from J. John Mann, M.D.—on bipolar disorder, while R. was there for his his schizophrenia.

For those new to the topic, a little bit of science review: transcranial magnetic stimulation is a technique that uses changing magnetic fields to induce neurons to fire in the brain. The firing pattern can either speed up or slow down neurons. We call these stimulation patterns “faciliatory” if they induce more firing or “inhibitory” if they reduce firing. TMS can be focal and stimulate just a tiny part of a circuit or, more broadly, with differently shaped magnetic coils.

A meta-analysis published in 2008¹ reviewed the data at the time, using inhibitory stimulation of the auditory cortex:

6… double-blind placebo-controlled and 4 crossover controlled trials, all randomized [were included] (n = 232). The primary outcome measure [reduced auditory hallucinations] …reached a significant Effect Size estimate (Hedges' g = 0.514; P = 0.001)

Thus, the evidence suggested even in 2008, this was a better treatment in schizophrenia with auditory hallucinations and most of the medicines we had.

There is more data today from a well-designed neuro navigated (via fMRI Using functional connectivity mapping) clinical trial out today, that demonstrates even more potent effects. The other major headline from this publication is a dose-response relationship between the magnetic field generated in the brain and the outcome of the treatment.

A brief aside, to explain auditory hallucinations. Everybody has a general sense with this might be like. You can imagine hearing voices. It's hard to imagine, if you haven't had psychotic symptoms, emotional experience of actually having those voices in your head. People have a relationship with their voices. They can be frightening, comforting, and some of this actually varies depending on where in the world grew up. In the US, most auditory nations tend to be persecutory in nature—the voices are saying negative or accusatory things. In other parts of the world, India for example, the content of those voices is reported as more positive. Auditory hallucinations are not unique to schizophrenia, they can actually happen across a range of individuals, many of whom don't have any psychiatric illness! About 5% of the population will hear voices at some point in life.

Having a voice in your head when you don't want it, and you don't have control of it, and it's saying things you'd rather not be hearing? This sucks. Taking medication that makes you gain massive amount of weight, as used to be the standard, until very recently, is an unappealing prospect that many people have simply had to live with. Being able to address the brain circuits that were causing this involuntary hearing of voices? That's something that gives people control back over their own minds. It has the potential a sense of agency to your experience of life.

Philosophical and explanatory points aside, let’s dig into this JAMA Open paper. The first thing I wanna emphasize is that journals are getting into memes, almost, in a big way. With the advent of Dr. Glaucomaflecken partnering with NEJM, and the explosion of decent infographics from major journals, we are getting closer to science being something anyone could understand.

Here is the infographic for this publication:

Who, where, what they did, what they measured, and what they found!

Patients with schizophrenia were randomized to either sham (with sensory stimulation to make it feel like you're getting TMS) or active transcranial magnetic stimulation.

As an aside, when I asked the creator of TMS for depression, Mark George, M.D., if he could tell the difference between modern sham and active TMS on his own head…he said, “Nope.” Unlike the recent psychedelic studies that rhyme with smokos, the blind is very much preserved in these trials! They tested this by asking trial participants to guess their assigned treatment condition:

Eighteen of 32 participants in the active rTMS group and 11 of 30 in the sham treatment group correctly guessed their actual group. There was no statistically significant difference between groups (χ² = 2.39; P = .12).²

This wouldn’t be a “The Frontier Psychiatrists” if I didn't look at table one for any given paper…

Astute table examiners will notice there isn’t a p-value listed for comparing the two groups—which would only tell us, “Are they different or not?” Instead, a t-test is reported, which was new to me. So, I learned something new! The t-test evaluates how different the two groups are.

Higher values of the t-score indicate that a large difference exists between the two sample sets. The smaller the t-value, the more similarity exists between the two sample sets.

Oh, gosh. Now, we all need to learn something new about statistics! But I got myself into this trouble by demanding my readers look at Table One, so I will hold myself to that standard. When we look at the differences between the intervention and experimental groups, we see T values between 1.08 and 1.74. They also use Mann-Whitney tests for illness, duration, olanzapine, and Hamilton, anxiety, and depression rating scales. I don’t know what those are either.

Frankly, I find this annoying— I am not a statistician. I should be a better one; if I’m producing science journalism, I should know more about statistics, which is why I'm looking these things up at this point. However, to anybody writing a paper, please, God, try to report your findings in the most straightforward, understandable way. So, for example, in this paper, I'm glad they gave me a tea value, but I would be a lot happier if they also gave me a P-value or indicated no significant difference between the groups. I am firmly of the opinion that what I'm likely determined now is that there is no significant difference between the groups, but I can't do that with just an eyeball; I have to go learn something about statistics to answer that question for myself, and thus for my readers.

I've learned thus far that when looking at a T-value, you assume a normal distribution in keeping with a null hypothesis. You need two additional pieces of information: the degree of freedom and the confidence interval you wish to use to answer the question, “Is this different or not.”

Assuming a normal distribution and no difference between the groups, which is the no hypothesis, the distribution looks like the above. And at two standard deviations, with a degrees of freedom know, you can get the following line, above which these groups would, in fact, be different:

But I have to find the degree of freedom to answer this question. FINE, Mom!

---

In fairness, my mom is a medical editor by training, so this is the kind of conversation I have with my real-life mother. Hi Mom! I appreciate all you've taught me about science, writing science, editing science, and trying to communicate clearly!

---

Or, I could read the text, which tells me:

There was no significant difference in baseline demographic and clinical variables between the active rTMS and sham treatment groups(Table 1).

Ok, so the randomization was successful. Moving on to Methods.

They gave each person an fMRI scan to target auditory circuits in the brain using functional connectivity brain mapping. This is at least superficially similar to the methodology I have described previously here, in PTSD, and here, in anxiety and OCD co-occurring with Depression.

It is a different methodology from SAINT, which uses a proprietary algorithm.

They then repeatedly stimulated this network using a figure-8 TMS coil, and I’m not going to spend time on those details other than to say it was over two weeks. The coil they used is focal; however, the authors modeled the induced e-field stimulated in each brain as an additional science flourish. This magnetic field is an electrically engaged area, a.k.a. e-field, that predicts some of the treatment response data:

It worked remarkably well.

How well is that, though? The effect size in this smallish trial ( which is still twice as large as the RCT that demonstrated SAINT neuromodulation was effective) demonstrated an effect size of Cohen d, 1.17-1.49.

Yes, that is more effective than any prior schizophrenia treatment when it comes to auditory hallucinations. You have been reading closely! It’s almost as if writing a whole book on crappy effect sizes of drugs (amazon affiliate link) paid off for you, dear readers.

It is also in the running for the safest treatment for this problem of all time:

There was no significant difference in the rates of adverse events between the active rTMS and sham treatment groups throughout the study (eTable 5 in Supplement 2). The most common adverse event was sleepiness. All adverse effects were tolerable and gradually disappeared on cessation of treatment. No serious adverse events were reported.

There you have it! fMRI-guided accelerated rTMS is in the running for the safest and most effective treatment for auditory hallucinations ever demonstrated. It works well for some, but on closer examination, it works okay for some and PHENOMENALLY well for others (the “responder” cohort).

I can’t wait to see this replicated!

1. Tranulis, C., Sepehry, A. A., Galinowski, A., & Stip, E. (2008). Should We Treat Auditory Hallucinations with Repetitive Transcranial Magnetic Stimulation? A Metaanalysis. The Canadian Journal of Psychiatry. https://doi.org/10.1177/070674370805300904

2. Hua Q, Wang L, He K, et al. Repetitive Transcranial Magnetic Stimulation for Auditory Verbal Hallucinations in Schizophrenia: A Randomized Clinical Trial. JAMA Netw Open. 2024;7(11):e2444215. doi:10.1001/jamanetworkopen.2024.44215