Why Antidepressants Make You Feel Worse Before They Make You Feel Better

Anyone who has started taking an SSRI knows the cruel little paradox of the first few weeks. You swallow the pill meant to lift your mood, and at first your mood does the opposite. The anxiety sharpens. Sleep frays. And then, somewhere around week three or four, if you stick with it, the fog begins to clear. For decades doctors have warned patients about this rough patch without being able to say, in any real molecular detail, why it happens. Now a team in Stockholm has gone looking inside the brain cells the drugs actually target, and found something rather neat: two populations of neurons being shoved in opposite directions by the very same drug.

The work, published in Molecular Psychiatry, focused on fluoxetine, better known to most of us as Prozac. SSRIs are among the most prescribed medications on the planet; in Sweden alone more than one in ten people are on one.

And yet, as Iskra Pollak Dorocic of Stockholm University puts it, we know far less than you might expect about what they get up to once they reach the brain. “Yet we still understand surprisingly little about what these drugs actually do in the brain. Our study set out to map the gene-expression changes SSRIs induce in their primary target, the brain’s serotonin neurons”, she says. That target is a small, unglamorous knot of tissue in the brainstem called the dorsal raphe nucleus, home to most of the serotonin neurons that wire up the forebrain.

One Label, Six Kinds of Cell

The old picture treated those neurons as a single uniform crowd. They aren’t.

Using a technique called spatial transcriptomics, which reads out which genes are switched on while keeping track of exactly where each cell sits in the tissue, the researchers pulled the dorsal raphe apart into six distinct serotonergic subpopulations, each with its own molecular signature and its own little neighbourhood. “Rather than treating the serotonin system as a single uniform population, we used spatial transcriptomics to read out gene activity at high resolution and map different types of serotonin neurons in the same brain area. That allowed us to see that these neurons are far more diverse than a single label suggests, and importantly that they do not all respond to the drug in the same way”, says Pollak Dorocic.

That last point is where it gets interesting. Because when the team dosed mice with fluoxetine, either a single hit or 22 days of it, and then mapped the changes, two neuropeptides told opposing stories.

The first was prodynorphin, or Pdyn, made by serotonin neurons clustered along the midline of the dorsal raphe. After a single dose, Pdyn expression jumped. Now, prodynorphin is the source of dynorphin, a molecule that activates the brain’s kappa opioid receptors, and kappa signalling has a reputation: it tends to produce dysphoria, the flat, low, slightly miserable state that is more or less the opposite of what an antidepressant is supposed to deliver. So here is the drug, on day one, cranking up a system linked to bad moods and, in some earlier work, to stress and even raised suicide risk early in treatment. The kicker is that this spike faded with time. After three weeks of dosing, Pdyn levels had dropped back down.

The second neuropeptide ran the opposite way. Thyrotropin-releasing hormone, or TRH, sits out in the lateral wings of the same nucleus, and it barely budged after a single dose. Only after chronic treatment did it climb, robustly. TRH is best known as a hormone of the thyroid axis, but it also turns up in mood-related circuits, and mice engineered to lack its receptors tend toward depression- and anxiety-like behaviour. In other words, the cells that might be doing the antidepressant’s actual therapeutic work only really get going after weeks, which is, give or take, exactly the lag patients experience.

A Mirror of the Clinic

“We found that two distinct serotonin neuron populations are pushed in opposite directions by the same drug, one early and transiently, and one slowly over weeks. That mirrors the clinical picture, where unpleasant effects often come first and relief comes later, and it gives us concrete molecular candidates to interrogate next”, says Pollak Dorocic.

It is worth keeping the caveats in view. These were healthy mice, not depressed ones, so the study captures the drug’s baseline fingerprint rather than how it behaves in a brain already altered by illness. The method, for all its spatial cleverness, still reads several cells at once rather than one at a time, so some of the finer detail is necessarily blurred. The findings also slot alongside, rather than overturn, the older textbook story about serotonin autoreceptors slowly desensitising; one gene the team tracked, Htr1a, followed that familiar up-then-down pattern as expected.

Still, what makes the neuropeptide result appealing is how cleanly it maps onto something patients have described for years. Two molecular switches, one flipping on early and briefly, one switching on late and staying on, between them sketching a plausible mechanism for both the early misery and the eventual relief. If that holds up, it points toward a tantalising possibility: a future antidepressant that leaves the TRH machinery alone to do its slow good work while sidestepping the dynorphin surge that makes the opening weeks such a slog. Better targets, in other words, and perhaps fewer reasons for people to give up before the drug has had its chance.

Whether any of this translates from mouse midbrain to human pharmacy shelf is, of course, the long game, and the team is careful not to oversell it. But for a side effect that has been shrugged off as a mystery for the better part of forty years, having two named molecules to chase feels like progress.

Read the full study in Molecular Psychiatry: Effects of SSRIs on the spatial transcriptome of dorsal raphe serotonin neurons.

---

Frequently Asked Questions

Why do SSRIs take weeks to work when they raise serotonin within hours?

Boosting serotonin is only the first step. This study suggests the therapeutic benefit may ride on a slow-building neuropeptide called TRH, which only ramps up after weeks of treatment, while a separate, fast-acting system tied to low mood spikes first and then fades. The mismatch in timing between these two systems could be a big part of why relief is delayed.

Is it true that antidepressants can make you feel worse at first?

For many people, yes, the early weeks bring heightened anxiety or a dip in mood before things improve. The Stockholm team found that a single dose of fluoxetine ramps up prodynorphin, which feeds into a brain system linked to dysphoria and stress. That early surge subsided with continued treatment, which may explain why the rough patch tends to pass.

How can one drug push brain cells in two opposite directions?

Because the serotonin neurons it targets are not all the same. Using spatial mapping, the researchers showed the dorsal raphe contains at least six distinct subpopulations, and two of them responded to fluoxetine in opposite ways and on opposite timescales. It is a reminder that a single drug acting on a single brain region can still produce wildly different effects cell by cell.

Could this lead to antidepressants with fewer side effects?

That is the hope, though it remains some way off. By identifying specific molecules behind the early adverse phase and the later therapeutic phase, the work hands drug developers concrete targets to aim at or avoid. A treatment that preserved the beneficial pathway while dampening the one tied to early distress is, for now, a promising idea rather than a product.