Giles W. Story, PhD, from the University College London

Giles W. Story, PhD

Delay discounting did not predict relapse after antidepressant discontinuation in patients with remitted 

 despite a modest association with residual depressive symptoms, new data show.

, of University College London, noted that the findings highlight the lack of reliable tools for predicting MDD relapse and emphasize the importance of individualized, shared decision-making when counseling patients.

The multi-site longitudinal study, led by Story and Doron Elad, PhD, of Technion-Israel Institute of Technology, enrolled 97 patients with remitted MDD receiving antidepressants and 54 matched healthy controls across Zurich and Berlin, with participants completing validated delay discounting tasks at baseline and after randomization to continuation or discontinuation. During the 6-month follow-up, 35% of patients relapsed. Patients with remitted MDD showed steeper baseline delay discounting than controls (Cohen’s d = 0.34), and a modest link to subthreshold depressive symptoms, but neither baseline discounting nor post-discontinuation changes predicted relapse.

During the interview, Story also addressed study limitations, including the relatively short follow-up period for a chronic, relapsing illness, and suggested future research may benefit from examining more dynamic, stress-responsive behavioral markers rather than relying on static, single-time-point measures.

 What should clinicians take away from the study when counseling patients who are considering continuing antidepressant medication?

 The decision to stop or come to the new antidepressant is a difficult one. It's one that has to be made on an individual basis, and [in] discussion with the clinician. In some ways, our study contributes to the uncertainty around that decision.

We know…that continuing an antidepressant actually reduces your risk of relapse. Amongst people…that stop [a] antidepressant, around 40% of those will go on to relapse. Those [who] continue it, around 20%...relapse. We know that [taking] an antidepressant reduces the risk of relapse, but this needs to be weighed against side effects and tolerability.

And of course, because we don't know for any given person whether or not stopping the medication is going to be the thing that makes the difference. It’s something that has to be thought about in context…where that person is in their life, or other treatment options. They might have [to] start some talking therapy…My advice would be that's something to think through and communicate with [a] clinician.

 Can you discuss the limitations of the study?

 It was a fairly short follow-up time. Patients were followed up for 6 months… whereas depression is a chronic illness that can relapse over some years. It may be that if you had a very long follow-up, you'd start to fix some of these effects. That said, we had a reasonable number relapsing even in [a] not short time.

There was some dropout. Some people were lost [to] follow up, [which could] potentially… mean we missed an effect. Let's say, for example, that more impulsive people were more likely to relapse and they're more likely to drop out. That might mean that we can miss [a] bunch of important signals.

 Were there any subgroups in whom delay discounting may still have clinical relevance?

 In this particular study, we didn't do much on the way of subgroup analysis, but we do know from other work that discounting is quite strongly associated with addiction dependence.

This is an area where discounting can have some migrants. [It] is probably the strongest of all clinical syndromes that are linked to discounting. And there’s actually some evidence of prospective prediction as well. For example…a study that measured people's discounting as they went into a bar…found that people [who] were more impatient for reward as they were entering the bar became more drunk than higher blood patients.

 How might future studies build on this work to identify more reliable behavior or cognitive prediction predictors of relapse?

 I think we need to move to more dynamic forms of prediction. Relapse is a complex thing and depends on that interaction between a person, their social world, and the outside world. It might be more fruitful to start to look at stress test markers. For example…how somebody's discounting changed in response to a change in their circumstances might be more powerful than discounting as a standalone.

Elad D, Story GW, Berwian IM, Stephan KE, Walter H, Huys QJM. Delay discounting correlates with depression but does not predict relapse after antidepressant discontinuation. Mol Psychiatry. Published online January 8, 2026. 

Story G. Delay Discounting vs Traditional Predictors of MDD Relapse, With Giles Story, PhD. HCPLive. Published on February 1, 2026. Accessed on February 4, 2026. 

https://www.hcplive.com/view/delay-discounting-vs-traditional-predictors-mdd-relapse-giles-story-phd

Story G. Delay Discounting Offers No Predictive Value for MDD Relapse, With Giles Story, PhD. HCPLive. Published on February 4, 2026. Accessed on February 4, 2026. 

https://www.hcplive.com/view/delay-discounting-offers-no-predictive-value-mdd-relapse-giles-story-phd

Articles

New study shows baseline impulsivity links to residual symptoms but not relapse, highlighting the need for individualized, clinician-guided decisions.

Delay Discounting Fails to Predict MDD Relapse, Q&A With Giles Story, PhD

Delay Discounting Offers No Predictive Value for MDD Relapse, With Giles Story, PhD

 who show greater impatience for delayed rewards are more likely to report residual depressive symptoms, but that behavioral trait does not predict relapse after stopping antidepressant therapy, according to new longitudinal data.

The findings suggest caution in using delay discounting as a clinical tool for guiding antidepressant discontinuation. Although this behavioral measure reflects known links between depression, impulsivity, and serotonergic function, investigators found it provided minimal predictive value in determining which patients are most likely to relapse after stopping medication.

“The bottom line is probably discounting is not going to be used like a lab test,” Giles W. Story, PhD, from the University College London, told 

Rather, clinicians use this study to guide conversations with patients about how depression may be associated with a focus on the present and challenges in planning for the future.

“We know avoidance is a big part of what maintains depression, and through that, people miss out on opportunities for longer-term rewards,” he added.

In a multi-site study led by Doron Elad, PhD, and Story, the study enrolled 97 patients with remitted MDD receiving antidepressant medication and 54 matched healthy controls across sites in Zurich and Berlin. Participants completed validated delay discounting tasks at baseline (MA1) and again after randomization to either antidepressant discontinuation or continuation, with 6 months of follow-up for relapse. Overall, 35% of patients relapsed during follow-up, while 65% remained well.

At baseline, patients with remitted MDD demonstrated significantly steeper delay discounting than controls (Cohen’s d = 0.34), indicating a greater tendency to favor immediate rewards. Delay discounting showed a modest association with depressive symptom severity, even when patients’ symptoms remained below the clinical threshold for relapse. For instance, patients were indifferent between receiving €66 immediately or €75 after a delay, whereas controls were indifferent at approximately €70.

Neither baseline discounting nor changes following antidepressant discontinuation predicted relapse during follow-up. Discontinuation led to only small increases in depressive symptoms, which may have been insufficient to shift discounting behavior.

“The effect that we saw after discontinuation of symptoms was quite small, so people's depressive symptoms did worsen slightly when they stopped medication, but these are people that already have quite low levels of symptoms, and they got just a tiny bit worse,” Story said. “It's probable that discount is probably just not sensitive enough to pick up that small change in symptoms.”

The results also challenge assumptions about behavioral markers as standalone predictors of relapse. Although residual symptoms predict relapse risk overall, they do not reliably predict whether stopping medication increases that risk.

Unexpectedly, higher baseline discounting was associated with greater symptom improvement over time.

“We thought that higher discounts at baseline might be associated with worsening of depression. The fact that we found the opposite strikes a bit of a note of caution,” Story said. “[It] could be a false positive occasionally…it's possible that there's another reason for…why discounting might be increasing depression. On the one hand… maybe that people are not willing to delay gratification, and that's actually part of something that maintains depression…On the other hand, it could be that becoming a bit more impulsive when you're depressed is actually adaptive, and it means that you can get a bit more immediate reward that boosts your mood a bit and carries you through.”

Videos

Delay discounting links to residual depression symptoms but does not predict relapse after stopping antidepressant treatment in patients with remitted MDD.

 Delay Discounting vs Traditional Predictors of MDD Relapse, With Giles Story, PhD

New evidence suggests that delay discounting, which is linked to depression and serotonin function, offers little predictive value for identifying patients at risk of relapse after stopping antidepressant treatment. While delay discounting was elevated in remitted 

and correlated with residual depressive symptoms, it neither changed following antidepressant discontinuation nor predicted relapse over 6 months of follow-up.

“We know that people with a history of depression tend to be more impatient for rewards,” Giles W. Story, PhD, from the University College London, told 

 “And then secondly, [there is] the…separate line of evidence that links depression to changes in serotonin function.”

Giles said the most reliable evidence comes from animal studies showing that serotonergic injections cause animals to be more impatient to get rewards. He predicted that delay discounting could reflect vulnerability to future depression and predict relapse, with greater baseline impatience associated with a greater likelihood of relapse. He also predicted that delay discounting might be a marker of discontinuation.

“In broad brush terms, the idea here is that some people who stop their antidepressants might experience a reduction in serotonin,” Story said.

In a multi-site longitudinal study, investigators examined whether delay discounting, a behavioral measure of impatience for future rewards, could predict relapse after stopping antidepressant medication in patients with remitted MDD. The study, led by Doron Elad, PhD, from Technion-Israel Institute of Technology, and Giles W. Story, PhD, enrolled 97 patients with remitted MDD receiving antidepressant medication (77% female; average age, 34.78 years) and 54 matched healthy controls (70% female; average age, 33.52 years). The study included participants from both Zurich (MDD: 71; controls: 32) and Berlin (MDD: 26; controls: 22).

Using 2 validated delay discounting tasks, investigators measured participants’ preference for immediate versus delayed monetary rewards, with steeper discounting indicating more impatience to wait for symptom improvement. Patients were assessed before and after discontinuation and followed for 6 months to track relapse.

Story explained that delay discounting provides a snapshot of how individuals value the future, offering a behavioral marker that differs conceptually from traditional clinical predictors such as the number of prior episodes or residual symptoms.

“You can think of it as a way [of capturing a] particular point in time somebody's making decisions,” Story said. “In some ways, you could say that's more objective. It's harder to second-guess.”

At baseline, patients with remitted MDD demonstrated significantly steeper discounting than healthy controls (Cohen’s d = 0.34), indicating reduced willingness to wait for future rewards. For example, patients were willing to accept approximately €66 immediately instead of €75 after several weeks, compared with €70 among controls. Baseline discounting was also modestly correlated with depressive symptom severity, even below the clinical threshold for relapse.

Contrary to the study’s primary hypotheses, neither baseline discounting nor early changes following antidepressant discontinuation predicted subsequent relapse. Discontinuation itself did not significantly alter discounting behavior, despite measurable increases in depressive symptoms among those who stopped medication earlier.

The findings suggest that while delay discounting is elevated in remitted depression and reflects residual symptom burden, it does not function as a clinically meaningful predictor of relapse after antidepressant discontinuation.

Elad D, Story GW, Berwian IM, Stephan KE, Walter H, Huys QJM. Delay discounting correlates with depression but does not predict relapse after antidepressant discontinuation. 

Although patients with remitted MDD showed greater impulsivity than controls, a study found no link between delay discounting and relapse risk after stopping antidepressants.