#2527: Do Brain Changes from Therapy or Pills Actually Last?

Do SSRI brain changes reverse after stopping? Can therapy physically rewire your brain for good? New neuroscience has answers.

0:000:00
Episode Details
Episode ID
MWP-2685
Published
Duration
34:54
Audio
Direct link
Pipeline
V5
TTS Engine
chatterbox-regular
Script Writing Agent
deepseek-v4-pro

AI-Generated Content: This podcast is created using AI personas. Please verify any important information independently.

Do Brain Changes from Therapy or Pills Actually Last?

A listener recently posed two questions that cut to the heart of modern neuroscience: When drug-induced brain changes reverse after someone stops taking an SSRI, does that tell us something fundamental about whether mental illness is a temporary state or a permanent trait? And can cognitive interventions like talk therapy produce physical rewiring visible on brain scans — and does that rewiring stick around after therapy ends?

The answers matter for millions of people making treatment decisions. Here's what the evidence actually shows.

The SSRI Problem: Scaffolding That Doesn't Always Hold

The standard story about SSRIs is that they increase serotonin, which boosts neuroplasticity, which allows the brain to rewire itself out of depression. That's not wrong, but it's dangerously simplified.

SSRIs block the serotonin transporter, raising extracellular serotonin within hours. But the antidepressant effect takes weeks. That delay is explained by the neuroplasticity hypothesis: elevated serotonin triggers a cascade involving cyclic AMP and BDNF (brain-derived neurotrophic factor) that promotes synaptic remodeling — new dendritic spines, strengthened or weakened connections. Actual physical restructuring.

But here's the critical question: when you remove the drug, do those structural changes persist without the chemical environment that built them?

A 2023 paper in Molecular Psychiatry followed patients on escitalopram longitudinally. Increased hippocampal neurogenesis and connectivity between the prefrontal cortex and amygdala correlated with symptom improvement. But when patients discontinued, a significant proportion showed reversal of those structural changes over 6-12 months. The architecture regressed.

The numbers are stark: 40-60% of patients relapse within a year of discontinuing an SSRI, versus 20-25% who stay on maintenance medication, according to a large meta-analysis in The Lancet Psychiatry.

What distinguishes those who maintain remission from those who relapse? The leading hypothesis is that "depression" is not one thing. If someone's depression is driven by an ongoing inflammatory process or unresolved chronic stressor, removing the drug removes the protective factor, and the brain reverts to its default state. If the depression was more episodic and the trigger has resolved, the drug may have helped the brain learn a new equilibrium it can maintain independently.

The Therapy Route: Learning That Physically Rewires

The second half of the question is more encouraging — and arguably more interesting. Can cognitive interventions produce physical brain changes, and do those changes persist?

The evidence is compelling. A group at UCLA showed that CBT for OCD normalizes the same brain circuits that medication normalizes — the orbitofrontal cortex and caudate nucleus — but through a different route. Medication works bottom-up, quieting the hyperactive caudate first. CBT works top-down: patients learn to reframe obsessive thoughts, engaging the prefrontal cortex, which then dampens caudate hyperactivity. Same destination, opposite direction.

This matters for persistence because cognitive learning is memory. It's encoded in synaptic weights. And consolidated memories can be remarkably durable.

A study from Helen Mayberg's lab at Oxford used fMRI to track CBT for depression before treatment, immediately after, and at follow-up. Immediately after CBT, patients showed increased prefrontal activation — the brain working harder to use new skills. At follow-up, if skills had been practiced and internalized, that prefrontal activation decreased. The brain had automated the new response, like learning to drive.

Structural changes follow a similar trajectory. A 2024 study using diffusion tensor imaging found increased fractional anisotropy in the uncinate fasciculus — the white matter tract connecting the prefrontal cortex to the amygdala — after CBT for social anxiety. Those changes persisted at a 12-month follow-up. A 2022 meta-analysis in JAMA Psychiatry pooled 15 studies and found that CBT-associated brain changes in prefrontal-limbic circuitry remained significant at follow-up periods ranging from 6 to 24 months.

The Synergistic Future: Plasticity-Augmented Psychotherapy

The different persistence profiles suggest a powerful combination strategy. The STAR*D trial found that combining medication and CBT produced better outcomes than either alone, especially for chronic and severe depression.

The emerging framework is "plasticity-augmented psychotherapy": use a plasticity-enhancing agent — an SSRI, ketamine, or psilocybin — to put the brain in a state where it can reorganize more readily, then do intensive therapeutic work during that window. The drug opens the window. Therapy determines what gets written through it.

This is not metaphor. Every time you learn something, your brain changes physically. Proteins are synthesized. Synapses are strengthened or pruned. The cognitive and the physical are the same process viewed from different angles. Therapy is just structured, targeted learning — and like any learning, its durability depends on how well the new patterns are practiced and consolidated.

Downloads

Episode Audio

Download the full episode as an MP3 file

Download MP3
Transcript (TXT)

Plain text transcript file

Transcript (PDF)

Formatted PDF with styling

#2527: Do Brain Changes from Therapy or Pills Actually Last?

Corn
Daniel sent us this one — he's been reading about SSRIs and neuroplasticity, and he's got two questions that actually connect in a way most people don't think about. First: when drug-induced brain changes don't persist after someone stops taking the medication, is that telling us something about the difference between a temporary state and a permanent state of mental illness? And second: can working at the cognitive level — talk therapy, CBT, reframing thoughts — actually produce physical rewiring we could see on a brain scan, and does that rewiring stick around after therapy ends? There's a lot to unpack here.
Herman
There really is. And Daniel's framing gets at something the neuroscience literature has been wrestling with for about fifteen years. The short answer to both questions is yes, but the mechanism is what makes this fascinating.
Corn
Before we dive into mechanism — Herman, you're about to go full walking encyclopedia on this. But let's also acknowledge that DeepSeek V4 Pro wrote this script for us today, so if anything comes out particularly coherent, credit where it's due.
Herman
So let me lay out what we actually know about SSRIs and neuroplasticity, because this is where a lot of pop science coverage gets it wrong. The story most people hear is that SSRIs increase serotonin, serotonin boosts neuroplasticity, brain rewires, depression lifts. That's not wrong, but it's so simplified it's almost misleading.
Corn
What's the version that's actually accurate?
Herman
SSRIs block the serotonin transporter, which increases extracellular serotonin within hours. But the antidepressant effect takes weeks. That delay is one of the oldest puzzles in psychopharmacology, and the neuroplasticity hypothesis is the best explanation we have. Elevated serotonin triggers a cascade of intracellular signaling — cyclic AMP, BDNF, brain-derived neurotrophic factor — that eventually promotes synaptic remodeling. New dendritic spines form. Existing synapses strengthen or weaken. This is actual physical restructuring.
Corn
The drug doesn't directly rewire anything. It creates a chemical environment where rewiring becomes more likely.
Herman
And this is where Daniel's first question gets really interesting. The rewiring happens because the brain is in a more plastic state. But when you remove the drug, you remove the chemical environment that enabled that plasticity. The question is whether the structural changes themselves can persist without the chemical scaffolding that built them.
Corn
The answer seems to be...
Herman
A really important paper in Molecular Psychiatry in twenty twenty-three looked at this longitudinally. Patients on escitalopram showed increased hippocampal neurogenesis and increased connectivity between the prefrontal cortex and the amygdala. Those changes correlated with symptom improvement. But when patients discontinued, a significant proportion showed reversal of those structural changes over six to twelve months.
Corn
The architecture regresses.
Herman
But here's the nuance — it doesn't always. And this gets directly to Daniel's point about state versus trait. Some patients discontinue and maintain remission. Their brains seem to have consolidated the changes. Others relapse within weeks or months, and their imaging looks like it did pre-treatment.
Corn
What distinguishes those two groups?
Herman
That's the billion-dollar question. If someone's depression is driven by an ongoing inflammatory process or a chronic stressor that hasn't been resolved, removing the drug just removes the protective factor, and the brain reverts to its default state. If the depression was more episodic, triggered by something that has since resolved, the drug may have helped the brain learn a new equilibrium that it can maintain on its own.
Corn
The drug is like scaffolding on a building. If the foundation is solid and you just needed scaffolding to repair some damage, you can remove it and the building stands. If the foundation is compromised, you take the scaffolding away and things start crumbling again.
Herman
That's a decent way to think about it. And it connects to something the field has been grappling with — "depression" is almost certainly not one thing. It's a syndrome with multiple distinct biological pathways. Some people's depression is primarily serotonergic. Some is more inflammatory. Some is HPA axis dysregulation. The same drug won't produce the same durability of effect across all those subtypes.
Corn
When Daniel asks about the distinction between a temporary state and a permanent state of mental illness — that's not just philosophical. It's biologically testable.
Herman
The neuroimaging data suggests that for some people, depression really is more of a state — a neurochemical and structural configuration their brain can be pushed out of, and with enough time in the remitted state, the brain essentially relearns what normal functioning looks like. For others, there seems to be a trait-level vulnerability that requires ongoing chemical support to keep the brain from reverting.
Corn
This makes me think about the relapse prevention literature. What are the actual numbers?
Herman
It depends on how long someone was on the drug and how they discontinue, but the general figure is that about forty to sixty percent of patients will relapse within a year of discontinuing an SSRI, compared to about twenty to twenty-five percent who stay on maintenance medication. That's from a large meta-analysis in The Lancet Psychiatry. The numbers are pretty stark.
Corn
Forty to sixty percent relapse. That's a lot of people whose brains apparently didn't consolidate the changes.
Herman
That brings us to the second part of Daniel's question, which I find genuinely more interesting. Can cognitive interventions — therapy, reframing, CBT — produce physical brain changes? And do those changes stick around?
Corn
This is where you get to quote your favorite imaging studies.
Herman
There's been a really compelling body of work on this. A group at UCLA led by Jeffrey Schwartz did some of the early studies on OCD patients using CBT, and they showed that the same brain circuits that medication normalizes — the orbitofrontal cortex, the caudate nucleus — also normalize with CBT. But here's the kicker: the pattern of normalization is different.
Herman
Medication tends to work bottom-up. It quiets the hyperactive caudate first, and the cortical changes follow. CBT works top-down. Patients learn to reframe their obsessive thoughts, which engages the prefrontal cortex, and over time, that dampens the caudate hyperactivity. Same circuits, opposite direction of effect, similar endpoint.
Corn
The brain arrives at the same destination but by a different route.
Herman
And this matters for Daniel's question about persistence. Because when you learn something cognitively — when you internalize a new way of thinking — that's a memory. It's encoded in synaptic weights. And memories, once consolidated, can be remarkably durable.
Corn
That assumes the cognitive change is actually learned, not just performed during therapy sessions.
Herman
And this is where the distinction between what happens in the therapy room and what happens six months later becomes critical. There was a really elegant study from Helen Mayberg's lab at Oxford that looked at CBT for depression using fMRI before, immediately after, and then at follow-up. Immediately after CBT, you see increased prefrontal activation when patients are regulating emotion — the brain working harder to use the new skills. At follow-up, if the skills have been practiced and internalized, that prefrontal activation actually decreases — it becomes more efficient. The brain has automated the new response.
Corn
It's like learning to drive. At first you're thinking about every movement, and your frontal lobes are on fire. Eventually it becomes automatic, and the brain activity pattern shifts.
Herman
That's exactly the analogy. And the structural changes follow a similar trajectory. A twenty twenty-four study used diffusion tensor imaging to look at white matter changes after a course of CBT for social anxiety. They found increased fractional anisotropy in the uncinate fasciculus — the white matter tract connecting the prefrontal cortex to the amygdala and anterior temporal lobe. Those changes were present at post-treatment and were still there at a twelve-month follow-up.
Corn
The structural changes from CBT actually persisted a year later.
Herman
In that study, yes. And a twenty twenty-two -analysis in JAMA Psychiatry pooled data from fifteen studies and found that CBT-associated brain changes in prefrontal-limbic circuitry were significant at post-treatment and remained significant at follow-up periods ranging from six to twenty-four months.
Corn
That's meaningfully different from the SSRI relapse numbers.
Herman
But we have to be careful not to overstate it. The studies aren't directly comparable — different patient populations, different severity levels, different outcome measures. And plenty of people do relapse after CBT too. The point isn't that therapy is categorically more durable than medication. The point is that they seem to operate through different mechanisms of change, and those mechanisms have different persistence profiles.
Corn
Which actually suggests that combining them might be the smartest approach.
Herman
The data strongly supports that. The STAR-D trial — the big NIMH-funded sequenced treatment study for depression — found that combining medication and CBT produced better outcomes than either alone, especially for chronic and severe depression. And more recent work suggests the combination may be synergistic in a specific way. The SSRI enhances plasticity, making the brain more amenable to change, and the CBT provides the structured learning that guides that change in a therapeutically useful direction.
Corn
The drug opens a window of plasticity, and therapy determines what gets written through that window.
Herman
That's exactly how many researchers now think about it. There's a term that's gained traction — "plasticity-augmented psychotherapy." The idea is that you use a plasticity-enhancing agent — could be an SSRI, could be ketamine, could be psilocybin in the trials running now — to put the brain in a state where it can reorganize more readily, and then you do intensive therapeutic work during that window.
Corn
This is where things get really interesting for Daniel's question about the cognitive level producing physical effects. Because what you're describing is a direct causal chain: a cognitive process — reframing a thought, challenging an assumption, rehearsing a new behavioral response — leads to physical changes in synaptic architecture that can be measured and that can persist.
Herman
That's not just metaphor. That's literal. Every time you learn something, your brain changes physically. Proteins are synthesized. Synapses are strengthened or pruned. The cognitive and the physical are not two different levels of description — they're the same process viewed from different angles.
Corn
This is where I think a lot of the cultural conversation goes sideways. People talk about therapy "rewiring your brain" like it's some kind of magical transformation, when really it's just learning. It's the same mechanism by which you learn anything else — a language, a musical instrument, a fear response. The brain is always rewiring itself. Therapy is just structured, targeted learning aimed at maladaptive patterns.
Herman
And the persistence question becomes: how well was the new pattern learned? If someone does twelve sessions of CBT and never practices the skills again, those synaptic changes will probably decay, just like your high school French decays if you never use it. If someone internalizes the cognitive habits and practices them for years, those changes can be as durable as any other well-consolidated memory.
Corn
There's an implication here that's uncomfortable for some people. If therapy works through the same mechanisms as any other form of learning, then it's subject to the same limitations. Some things are hard to learn. Some things are hard to unlearn. And some biological states might make learning nearly impossible.
Herman
And this is where the state-versus-trait distinction Daniel raised becomes really clinically relevant. If someone is in a severe depressive episode, their ability to engage in the cognitive work of therapy is compromised. Their concentration is impaired. Their motivation is low. Their working memory is degraded. You're asking someone whose brain is not functioning well to do difficult cognitive work that requires precisely the functions that are impaired.
Corn
Which is why medication can be a prerequisite for therapy to work at all.
Herman
The medication lifts the neurochemical fog enough that the person can actually do the learning. And then the learning produces its own structural changes that may persist even if the medication is eventually withdrawn.
Corn
The ideal sequencing might be: medication to open the window, therapy to guide the rewiring, medication taper while the new patterns consolidate, and then ongoing practice to maintain the changes.
Herman
That's the theoretical ideal. In practice, it's messier. Insurance doesn't always cover that kind of integrated treatment. The timing of each phase varies enormously between individuals. And some people's underlying biology may simply require ongoing pharmacological support no matter how much therapy they do.
Corn
Let me push on something. You mentioned that CBT-associated brain changes seem to persist at follow-up. But how much of that is selection bias? The people who agree to be in a follow-up study and who actually show up a year later are probably the ones who are doing well.
Herman
That's a legitimate concern, and it's a known issue in psychotherapy research. But there have been studies that used intent-to-treat analyses and still found significant effects at follow-up. And the imaging studies in particular — you can't really fake fractional anisotropy on a DTI scan. Those are objective measures.
Corn
But I want to circle back to something you said earlier about SSRIs and the relapse numbers. If forty to sixty percent of people relapse within a year of discontinuing, is that a withdrawal effect or a recurrence of the underlying condition?
Herman
This is one of the most contentious debates in psychiatry. Part of what makes it hard to disentangle is that SSRI discontinuation itself can produce symptoms that look a lot like depression relapse — irritability, low mood, fatigue, sleep disturbance. So some of what gets counted as relapse may actually be withdrawal.
Corn
The drug companies don't have a strong incentive to clarify that distinction.
Herman
Historically, that's been a concern. But the field has gotten more sophisticated. There are now studies that do very gradual tapers over months and still see significant relapse rates, which suggests that at least a substantial portion of it is genuine recurrence of the underlying condition rather than withdrawal.
Corn
Which brings us back to Daniel's first question. If the relapse is genuine recurrence, that suggests the drug was suppressing a trait-level vulnerability rather than curing a state-level episode.
Herman
The neuroimaging is consistent with that interpretation. When you see the brain changes reverse after discontinuation, that's evidence that the drug was actively maintaining a different neurobiological state. The underlying vulnerability was still there. The drug was holding it at bay.
Corn
For some people, depression really is more like type one diabetes — you need ongoing exogenous support because your system doesn't self-regulate properly. For others, it's more like a fracture — you need support during the healing process, but once it's healed, it's healed.
Herman
That's a useful framework, though I'd add that even the fracture analogy has limits. The site of a healed fracture is often structurally different from the original bone — sometimes stronger, sometimes more vulnerable. The brain after a depressive episode is not the same brain it was before. It has a history. And that history may make it more susceptible to future episodes.
Herman
The kindling hypothesis — each episode makes future episodes more likely and less dependent on external triggers. It's well-established for bipolar disorder, and there's evidence for it in unipolar depression too. So even if someone fully remits and maintains remission without medication, their brain has been changed by the experience of having been depressed.
Corn
Which means the state-trait distinction isn't a clean binary. Having an episode changes the trait.
Herman
The brain is a dynamical system with memory. Its current state depends on its history. You can't just reset it to a pre-depression baseline and pretend nothing happened.
Corn
Let's shift to something Daniel asked that we haven't fully addressed. He mentioned "working at the cognitive level" producing physical effects. I think there's a deeper question here about what it even means for something to be "cognitive" versus "physical.
Herman
This is where I get to be a little philosophical. The distinction between cognitive and physical is, from a neuroscientific perspective, not a real distinction. Thoughts are physical processes. They're patterns of neural activity. There's no thought without a physical substrate. So when we say CBT produces physical changes, we're not saying something magical — we're saying that one pattern of physical activity leads to lasting changes in the physical structure that supports future activity.
Corn
People experience their thoughts as non-physical. That's the phenomenology. And I think that's part of why the idea that therapy "rewires the brain" feels revelatory. They experience their thoughts as ephemeral, so learning that thoughts leave physical traces feels like discovering something profound.
Herman
It is profound in a way. But it's also just what learning is. The reason I emphasize this is that when people think of therapy-induced brain changes as something exotic or special, they can lose sight of the fact that every experience changes their brain. The question isn't whether an experience produces physical changes — it always does. The question is whether those changes are therapeutically useful and whether they persist.
Corn
What determines whether a cognitive intervention produces durable, therapeutically useful changes?
Herman
One is dose — both frequency and duration. A single therapy session won't produce measurable structural changes. A course of sixteen to twenty sessions over several months, with between-session practice, can. Two is the nature of the learning. Skills practiced in multiple contexts and generalized tend to be more durable than skills only used in the therapist's office. Three is emotional salience. The brain consolidates emotionally charged learning more strongly than neutral learning.
Corn
That last point is interesting because depression itself dampens emotional responsiveness. So you're asking someone whose emotional systems are blunted to do emotionally salient learning.
Herman
Which is another argument for combining medication and therapy. The medication can restore some emotional range, which then makes the therapeutic learning more effective.
Corn
I want to ask about something that's been implicit in this whole conversation but we haven't named directly. The idea of "reframing" — which Daniel mentioned — is central to CBT. But what does reframing actually do in the brain?
Herman
This has been studied quite a bit. Reframing — or cognitive reappraisal, the technical term — involves generating an alternative interpretation of a situation that changes its emotional impact. Neuroimaging studies consistently show that reappraisal engages the dorsolateral prefrontal cortex, the ventrolateral prefrontal cortex, and the dorsal anterior cingulate. These are regions involved in cognitive control, working memory, and conflict monitoring.
Corn
They're modulating the amygdala.
Herman
The prefrontal regions send inhibitory projections to the amygdala, dampening its response. What's interesting is that in untreated depression, this prefrontal-to-amygdala regulation is often impaired. The prefrontal cortex is less active, the amygdala is hyperactive, and the connectivity between them is reduced.
Corn
Depressed people literally can't regulate their emotional responses as effectively.
Herman
And CBT seems to strengthen exactly this circuit. After a course of CBT, you see increased prefrontal activation during reappraisal, increased functional connectivity between the prefrontal cortex and the amygdala, and decreased amygdala reactivity to negative stimuli.
Corn
This is a learned skill. The patient is practicing a cognitive operation that, over time, changes the physical architecture of their emotion regulation circuitry.
Herman
And here's where the persistence question gets really interesting. There's evidence that once this circuitry is strengthened through practice, it can become the brain's default mode of responding to negative stimuli. It becomes automatic. The person doesn't have to consciously reappraise — their brain has learned to do it implicitly.
Corn
Which would explain why the changes can persist after therapy ends. The new pattern has been consolidated into the brain's automatic processing.
Herman
And this is fundamentally different from how medication works. Medication doesn't teach the brain anything. It changes the chemical environment, which changes activity patterns, which may lead to structural changes — but those changes aren't guided by learning. They're more like the brain adapting to a new chemical milieu. When you remove the chemical, the brain adapts back.
Corn
Therapy produces guided, targeted changes that are reinforced by practice and consolidated as learned patterns. Medication produces chemical scaffolding that enables the brain to find its own equilibrium, but that equilibrium may depend on the continued presence of the scaffolding.
Herman
That's a very clean way to put it. And it explains both why therapy changes can be more durable and why medication is sometimes still necessary. If the brain's default equilibrium is pathological, no amount of learning will permanently override it without ongoing chemical support.
Corn
There's something here about the limits of learning. We tend to think of learning as almost infinitely powerful — you can learn anything if you try hard enough. But learning is constrained by biology. You can't learn your way out of a broken thermostat.
Herman
That's a really important point. And it's one that sometimes gets lost in discussions that overstate the power of therapy. CBT is not a cure-all. The effect sizes in the major -analyses are moderate — meaningful, clinically significant, but moderate. Some people do everything right and still struggle.
Corn
Which is not a failure of therapy. It's a reflection of the fact that mental illness has biological underpinnings that can't always be overcome by cognitive work alone.
Herman
And acknowledging that is not defeatist. It's realistic. It helps people make informed decisions about treatment and reduces the shame that comes from feeling like they failed at therapy.
Corn
Let me pull on another thread. Daniel mentioned "reframing our thoughts" in a way that suggests he's thinking about the cognitive level as distinct from the biological level. But we've been talking about them as the same thing viewed from different angles. Is there any sense in which the cognitive level has causal power that isn't reducible to the biological?
Herman
This is a deep philosophical question, and I'll give you my take as someone who's thought about this clinically. I think the cognitive level is real and has causal power, but not because it's separate from the biological. It's because it's a higher-level description of the same system, and sometimes the higher-level description captures patterns and regularities that are hard to see at the lower level.
Corn
"reframing a thought" is a real thing you can do, and it has real effects, but it's implemented in neural activity. It's not magic.
Herman
And the reason this matters for Daniel's question is that it means cognitive interventions can produce physical changes because cognitive interventions are physical interventions. They're just physical interventions that are initiated and guided at the cognitive level.
Corn
Which is actually kind of empowering when you think about it. It means that the work you do in therapy — the difficult, effortful work of challenging your assumptions and practicing new ways of thinking — is literally reshaping your brain. It's not just talk. It's construction.
Herman
The construction metaphor is actually apt. When you learn a new cognitive skill, your neurons are building new synapses, strengthening some connections and weakening others, possibly even generating new neurons in the hippocampus. These are physical construction projects happening in your head, driven by your own mental effort.
Corn
That's a much more compelling case for doing the hard work of therapy than "it might help you feel better.
Herman
Though I want to be careful not to over-promise. The changes are real, but they're also gradual and require sustained effort. You're not going to see your brain remodel itself after a week of thought records.
Corn
But the principle stands. And it connects to something I've been thinking about throughout this conversation. Daniel asked whether therapy-induced changes can be sustained beyond the treatment period. We've talked about the imaging evidence suggesting they can. But I think there's a deeper point about what "sustained" means.
Corn
If therapy teaches you a skill — say, cognitive reappraisal — and you continue to use that skill after therapy ends, then the brain changes associated with that skill will be maintained because you're still practicing the skill. The question isn't really whether the changes from therapy persist in the absence of continued practice. It's whether therapy gives you tools that you continue to use, and the ongoing use maintains the changes.
Herman
That's a really important reframing. Pun not intended. The durability of therapy effects depends on whether the skills become integrated into the person's daily life. If they do, the brain changes are continuously reinforced. If they don't, the changes will likely fade, just like any unused skill.
Corn
Which means the question "do therapy-induced brain changes persist" is partly a question about whether therapy produces lasting behavior change. If the behavior changes, the brain changes follow.
Herman
And this is why therapy modalities that emphasize skill-building and between-session practice tend to have better long-term outcomes. CBT, dialectical behavior therapy, behavioral activation — these all involve teaching concrete skills that people can continue to use independently.
Corn
Compared to something like supportive therapy, where the benefit may depend more on the ongoing therapeutic relationship.
Herman
And the imaging data bears this out. The studies that show persistent brain changes tend to be studies of skill-based therapies where patients are actively learning and practicing new cognitive and behavioral strategies.
Corn
I want to go back to something you mentioned earlier about the kindling effect. If each depressive episode makes future episodes more likely, then preventing episodes becomes really important. And if therapy can produce durable changes that reduce relapse risk, that's not just a quality-of-life issue — it's potentially altering the long-term trajectory of the illness.
Herman
That's one of the most compelling arguments for investing in therapy as a first-line treatment, or at least as a combined treatment with medication. If you can prevent episodes, you may be preventing the progressive worsening that comes with repeated episodes. There's some evidence that CBT has a relapse prevention effect that medication alone doesn't match — not because medication doesn't work, but because when you stop medication, the prophylactic effect stops, whereas the skills learned in therapy can continue to provide protection.
Corn
There was a study on this, wasn't there? Looking at relapse rates after CBT versus medication.
Herman
The classic ones are from the nineties and early two thousands, but there have been more recent replications. The pattern is fairly consistent: patients treated with CBT have lower relapse rates after treatment ends than patients treated with medication who discontinue. But patients who stay on maintenance medication have relapse rates comparable to or lower than CBT. The advantage of CBT is durability after treatment ends. The advantage of medication is efficacy during treatment.
Corn
Which again points to combination treatment being optimal for many people.
Herman
And I think the field is increasingly moving toward a model where medication is used to achieve remission and therapy is used to build the skills that maintain remission. The medication buys you the opportunity to do the therapeutic work. The therapeutic work buys you durable change.
Corn
This has been a dense conversation. Let me try to pull together what we've actually said in response to Daniel's two questions.
Corn
Question one: is the non-persistence of drug-induced changes related to the distinction between temporary states and permanent states of mental illness? Our answer is yes, with nuance. SSRIs create a chemical environment that enables plasticity, and the brain rewires in that environment. But for many people, the rewiring depends on the continued presence of the drug. When the drug is removed, the brain tends to revert toward its pre-treatment state. This suggests that for those people, the drug was managing a trait-level vulnerability rather than curing a state-level episode. But there's meaningful variation — some people do maintain remission after discontinuation, and we don't fully understand what distinguishes them from those who relapse.
Herman
That's a fair summary. And question two?
Corn
Can working at the cognitive level produce physical brain changes, and do those changes persist? Answer: yes and often yes. CBT and other skill-based therapies produce measurable changes in brain structure and function — increased prefrontal control over limbic regions, strengthened connectivity in emotion regulation circuits, even white matter changes visible on diffusion tensor imaging. These changes can persist for months to years after therapy ends, but their durability depends on whether the skills are practiced and internalized. Therapy works through the same mechanisms as any other form of learning — it's not magic, it's targeted, structured learning that reshapes neural circuits through practice and consolidation.
Herman
I'd add one thing. The persistence of therapy-induced changes is not a given. It requires that the person actually learned the skills and continues to use them. Therapy that doesn't produce lasting behavior change probably doesn't produce lasting brain change. The cognitive and the behavioral are the active ingredients. The brain changes are the physical trace of that learning.
Corn
The takeaway for someone like Daniel, who's clearly thought deeply about this, is that the cognitive and the biological aren't competing explanations for how mental health treatment works. They're complementary levels of description. Drugs and therapy both change the brain, but through different mechanisms and with different persistence profiles. The smart approach is usually both, sequenced thoughtfully, with the recognition that some people's biology may require ongoing pharmacological support no matter how much therapeutic work they do.
Herman
The other takeaway is that doing the hard work of therapy is worth it. It's not just talking about your feelings. It's a process that physically reshapes your brain in ways that can provide lasting protection against future episodes. That's not pop psychology. That's neuroscience.
Corn
Now: Hilbert's daily fun fact.
Herman
The average cumulus cloud weighs approximately one point one million pounds, roughly the same as eighty elephants.
Corn
What can listeners actually do with this? First, if you're considering treatment for depression, think about medication and therapy not as alternatives but as potentially complementary tools with different strengths. Medication may help you get to a place where therapy can actually work. Therapy may help you build skills that protect you after medication ends. Second, if you're in therapy, take the between-session practice seriously. The brain changes depend on repetition and consolidation. Doing the homework isn't just busywork — it's literally how the rewiring happens. Third, if you've been on medication for a while and you're thinking about discontinuing, work with your prescriber on a very gradual taper and consider whether you've built the cognitive and behavioral skills that might help maintain your gains. Don't just stop and hope for the best.
Herman
I'd add: if you've tried therapy and it didn't work, that doesn't mean you failed or that your problems are untreatable. It may mean you need a different type of therapy, or that you need medication to make therapy accessible, or that your particular biology requires a different approach entirely. The fact that therapy produces brain changes doesn't mean it produces them equally for everyone in every circumstance.
Corn
One open question I'm left with: we've talked about SSRIs and CBT because those are the best-studied interventions. But the landscape is changing fast. Psychedelic-assisted therapy, ketamine, transcranial magnetic stimulation — these all operate on plasticity through different mechanisms. How the durability picture changes when you introduce these newer approaches is something we're going to be learning about for years.
Herman
The early data on psilocybin-assisted therapy, for instance, suggests that the combination of a powerful plasticity-enhancing agent with intensive therapeutic integration work may produce remarkably durable changes. But we need much more long-term follow-up before we can say that with confidence.
Corn
This has been My Weird Prompts. Thanks to our producer Hilbert Flumingtop for keeping us on track, and thanks to Daniel for a thought-provoking set of questions. You can find every episode at myweirdprompts.We'll be back soon.
Herman
Take care, everybody.

This episode was generated with AI assistance. Hosts Herman and Corn are AI personalities.