CoChat

A mechanistic account of serotonin’s impact on mood

Open access

The DOI matches its Crossref record; 57 of the 59 references it cites were found in a registry or an index.

Abstract

Selective serotonin reuptake inhibitors (SSRIs) constitute a first-line antidepressant intervention, though the precise cognitive and computational mechanisms that explain treatment response remain elusive. Using week-long SSRI treatment in healthy volunteer participants, we show serotonin enhances the impact of experimentally induced positive affect on learning of novel, and reconsolidation of previously learned, reward associations. Computational modelling indicated these effects are best accounted for by a boost in subjective reward perception during learning, following a positive, but not negative, mood induction. Thus, instead of influencing affect or reward sensitivity directly, SSRIs might amplify an interaction between the two, giving rise to a delayed mood response. We suggest this modulation of affect-learning dynamics may explain the evolution of a gradual mood improvement seen with these agents and provides a novel candidate mechanism for the unfolding of serotonin’s antidepressant effects over time.

Overview

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1038/s41467-020-16090-2

Cites

57 of 57 references verified0 citation errors · 0 not checkable

Statistics

Figures

Fig. 1 Experimental task and study design. a Trial structure. On every trial, subjects had 3 s to make a choice between two presented images. The chosen image was highlighted for 0.7 s. Subsequently, an outcome, either reward (gold coin), no reward (red cross), or no feedback (box), was shown above the image for 1.5 s. After a 2 s fixation cross period, the next trial started. b Different outcome types. On learning trials, subjects encountered either a monetary reward (gold coin, £0.03), or no reward (red cross) outcome, depending on the chosen image’s associated reward probability. Here, each image was associated with either a fixed high (70%) or a fixed low reward probability (30%). On test and interleaved maintenance trials, outcomes were not revealed. Instead, they were ‘hidden in a box’ visible to subjects. Subjects were instructed that images still led to reward outcomes, depending on reward probabilities. However, to prevent new learning at this stage, feedback was not shown. c Mood rating. Multiple times per block (three times during blocks of 48 trials, four times during blocks of 72 trials; fixed intervals), subjects were asked to report their current emotional state by responding to the question ‘How do you feel right now?’. The rating was provided without time restriction by moving a red cursor that always started at the midpoint along a scale ranging from ‘very unhappy’ to ‘very happy’. d Study design. Subjects were randomly allocated to take a daily dose of 20 mg citalopram or placebo for 7 days and participated in four session over the week: a pre-drug session I on day 1, and three sessions on the drug, on days 1, 4, and 7 (sessions II, III and IV). To modulate subject’s affective state between learning blocks, a Wheel of Fortune (WoF) draw was held in session II (day 1) and IV (day 7). Cf. Supplementary Fig. 1 for details on session structure and image appearance across the entire study.
Fig. 1 Experimental task and study design. a Trial structure. On every trial, subjects had 3 s to make a choice between two presented images. The chosen image was highlighted for 0.7 s. Subsequently, an outcome, either reward (gold coin), no reward (red cross), or no feedback (box), was shown above the image for 1.5 s. After a 2 s fixation cross period, the next trial started. b Different outcome types. On learning trials, subjects encountered either a monetary reward (gold coin, £0.03), or no reward (red cross) outcome, depending on the chosen image’s associated reward probability. Here, each image was associated with either a fixed high (70%) or a fixed low reward probability (30%). On test and interleaved maintenance trials, outcomes were not revealed. Instead, they were ‘hidden in a box’ visible to subjects. Subjects were instructed that images still led to reward outcomes, depending on reward probabilities. However, to prevent new learning at this stage, feedback was not shown. c Mood rating. Multiple times per block (three times during blocks of 48 trials, four times during blocks of 72 trials; fixed intervals), subjects were asked to report their current emotional state by responding to the question ‘How do you feel right now?’. The rating was provided without time restriction by moving a red cursor that always started at the midpoint along a scale ranging from ‘very unhappy’ to ‘very happy’. d Study design. Subjects were randomly allocated to take a daily dose of 20 mg citalopram or placebo for 7 days and participated in four session over the week: a pre-drug session I on day 1, and three sessions on the drug, on days 1, 4, and 7 (sessions II, III and IV). To modulate subject’s affective state between learning blocks, a Wheel of Fortune (WoF) draw was held in session II (day 1) and IV (day 7). Cf. Supplementary Fig. 1 for details on session structure and image appearance across the entire study.CC BY · PubMed Central
Fig. 2 No drug effects on reward learning and maintenance of learned associations. Correct choices for images learned during a the pre-drug session and b over the course of 1-week drug treatment. Task performance was quantified as proportion of choices of the image associated with a high reward probability on trials comparing images of different reward probability, i.e., high vs. low. Learning was assessed during a test block on every session, comparing images from blocks 1 and 2, encountered during preceding learning. Maintenance was assessed during interleaved trials involving familiar images on subsequent sessions, where images learned about in session I re-appeared in session II, and images from session II in session III, etc. Notably, both learning and maintenance were assessed on choices between the same image pairs, with the time-lag between learning and maintenance test constituting the only difference. Results show that learning performance was stable across sessions, and unaffected by SSRI treatment (pre-drug (n = 63): F1,61 = 0.5, p = 0.457; sessions II−IV (n = 62): drug: F1,60 = 0.6, p = 0.428; drug × session: F2,120 = 0.3, p = 0.717). Moreover, performance decreased more for 3-day (II to III and III to IV) as compared to 3-h time-lag (I to II, time: F2,120 = 6.8, p = 0.002), but performance decline was unaffected by drug manipulation drug: (F1,60 = 0.03, p = 0.858; drug × time: F2,120 = 0.005, p = 0.995). n.s. = not significant (no difference across drug groups). Error bars indicate SEM.
Fig. 2 No drug effects on reward learning and maintenance of learned associations. Correct choices for images learned during a the pre-drug session and b over the course of 1-week drug treatment. Task performance was quantified as proportion of choices of the image associated with a high reward probability on trials comparing images of different reward probability, i.e., high vs. low. Learning was assessed during a test block on every session, comparing images from blocks 1 and 2, encountered during preceding learning. Maintenance was assessed during interleaved trials involving familiar images on subsequent sessions, where images learned about in session I re-appeared in session II, and images from session II in session III, etc. Notably, both learning and maintenance were assessed on choices between the same image pairs, with the time-lag between learning and maintenance test constituting the only difference. Results show that learning performance was stable across sessions, and unaffected by SSRI treatment (pre-drug (n = 63): F1,61 = 0.5, p = 0.457; sessions II−IV (n = 62): drug: F1,60 = 0.6, p = 0.428; drug × session: F2,120 = 0.3, p = 0.717). Moreover, performance decreased more for 3-day (II to III and III to IV) as compared to 3-h time-lag (I to II, time: F2,120 = 6.8, p = 0.002), but performance decline was unaffected by drug manipulation drug: (F1,60 = 0.03, p = 0.858; drug × time: F2,120 = 0.005, p = 0.995). n.s. = not significant (no difference across drug groups). Error bars indicate SEM.CC BY · PubMed Central
Fig. 3 SSRIs induce an impact of positive mood on learning. a Each session comprised two learning blocks, where subjects encountered four novel images, two with a high (70%) and two with a low reward probability (30%). Feedback enabled subjects to learn about images’ reward probabilities through trial-and-error. Additionally, sessions II and IV comprised a WoF draw between learning blocks, where each subject either won (yellow) or lost (red) a large monetary amount (£7−£9). After the second learning block, subjects performed a test choice block, where they chose between images from the two preceding learning blocks. To prevent further learning, outcomes were not revealed in these test blocks, and subjects were instructed to rely solely on what they had learned previously about the images. b Proportion of test block choices between images encountered during preceding learning with feedback in different blocks, i.e., pre-WoF or post-WoF. This analysis compared choices between images with equal reward probabilities, i.e., high vs. high and low vs. low, and preference is computed as difference between percent choices of post-WoF images and pre-WoF images, as a function of WoF outcome. Participants showed a preference for images encountered in better mood, favouring post-WoF images after a winning draw, and favouring pre-WoF images after a losing draw. The SSRI group showed a significantly positive mood bias (F1,32 = 12.1, p = 0.001), and it was enhanced as compared to placebo subjects (F1,60 = 5.5, p = 0.021, n = 62). **p < 0.01, *p < 0.05. Error bars indicate SEM. c Interaction between affect and reward sensitivity. Rewards impact on emotional state, which, in turn, impacts on perception of subsequent rewards. In this bidirectional interaction, SSRIs enhance the impact of positive affect on reward sensitivity. This serotonergic effect can engender a positive feedback loop.
Fig. 3 SSRIs induce an impact of positive mood on learning. a Each session comprised two learning blocks, where subjects encountered four novel images, two with a high (70%) and two with a low reward probability (30%). Feedback enabled subjects to learn about images’ reward probabilities through trial-and-error. Additionally, sessions II and IV comprised a WoF draw between learning blocks, where each subject either won (yellow) or lost (red) a large monetary amount (£7−£9). After the second learning block, subjects performed a test choice block, where they chose between images from the two preceding learning blocks. To prevent further learning, outcomes were not revealed in these test blocks, and subjects were instructed to rely solely on what they had learned previously about the images. b Proportion of test block choices between images encountered during preceding learning with feedback in different blocks, i.e., pre-WoF or post-WoF. This analysis compared choices between images with equal reward probabilities, i.e., high vs. high and low vs. low, and preference is computed as difference between percent choices of post-WoF images and pre-WoF images, as a function of WoF outcome. Participants showed a preference for images encountered in better mood, favouring post-WoF images after a winning draw, and favouring pre-WoF images after a losing draw. The SSRI group showed a significantly positive mood bias (F1,32 = 12.1, p = 0.001), and it was enhanced as compared to placebo subjects (F1,60 = 5.5, p = 0.021, n = 62). **p < 0.01, *p < 0.05. Error bars indicate SEM. c Interaction between affect and reward sensitivity. Rewards impact on emotional state, which, in turn, impacts on perception of subsequent rewards. In this bidirectional interaction, SSRIs enhance the impact of positive affect on reward sensitivity. This serotonergic effect can engender a positive feedback loop.CC BY · PubMed Central
Fig. 4 SSRIs modulate the impact of mood on reconsolidation. a In addition to novel images, each block comprised interleaved trials involving choices between familiar images from previous sessions. Outcomes were not revealed on such trials to prevent further learning, and subjects had to rely on previously acquired knowledge when making choices. On session IV, a subset of images re-appeared as interleaved trials either before (pre-WoF), or after (post-WoF), the mood manipulation. The final test block comprised an additional set of trials comparing these images (re-appeared pre-WoF vs. re-appeared post-WoF). By design, such images only differed in the mood state under which subjects retrieved learned reward associations from memory when choosing. Note these images had initially been learned in the same block, unbiased by mood during session III (cf. Methods for details). b Proportion of test block choices between images that had been encountered without feedback during preceding reconsolidation in different blocks, i.e., pre-WoF or post-WoF. This analysis involved choices between images with objectively similar reward probabilities, i.e., high vs. high and low vs. low, and preference is computed as difference between percent choices of post-WoF images and pre-WoF images, as a function of WoF outcome. Contrasting a mood bias on learning, subjects preferred images that re-appeared post-WoF, irrespective of whether the WoF was won or lost. However, this effect was differentially modulated by the drug (drug × valence: F1,60 = 4.8, p = 0.032, n = 64), with a significant preference for images encountered after winning F1,15 = 6.4, p = 0.023) but not losing the draw in the SSRI group, and a preference for images encountered after losing (F1,15 = 10.1, p = 0.006) but not winning the draw in the placebo group. **p < 0.01, *p < 0.05. Error bars indicate SEM.
Fig. 4 SSRIs modulate the impact of mood on reconsolidation. a In addition to novel images, each block comprised interleaved trials involving choices between familiar images from previous sessions. Outcomes were not revealed on such trials to prevent further learning, and subjects had to rely on previously acquired knowledge when making choices. On session IV, a subset of images re-appeared as interleaved trials either before (pre-WoF), or after (post-WoF), the mood manipulation. The final test block comprised an additional set of trials comparing these images (re-appeared pre-WoF vs. re-appeared post-WoF). By design, such images only differed in the mood state under which subjects retrieved learned reward associations from memory when choosing. Note these images had initially been learned in the same block, unbiased by mood during session III (cf. Methods for details). b Proportion of test block choices between images that had been encountered without feedback during preceding reconsolidation in different blocks, i.e., pre-WoF or post-WoF. This analysis involved choices between images with objectively similar reward probabilities, i.e., high vs. high and low vs. low, and preference is computed as difference between percent choices of post-WoF images and pre-WoF images, as a function of WoF outcome. Contrasting a mood bias on learning, subjects preferred images that re-appeared post-WoF, irrespective of whether the WoF was won or lost. However, this effect was differentially modulated by the drug (drug × valence: F1,60 = 4.8, p = 0.032, n = 64), with a significant preference for images encountered after winning F1,15 = 6.4, p = 0.023) but not losing the draw in the SSRI group, and a preference for images encountered after losing (F1,15 = 10.1, p = 0.006) but not winning the draw in the placebo group. **p < 0.01, *p < 0.05. Error bars indicate SEM.CC BY · PubMed Central
Fig. 5 SSRIs induce mood effects via affect-learning interaction. a Mean mood ratings during the first block of every session, i.e., prior to any mood manipulation. Mood self-reports did not differ between groups in the pre-drug session (session I), or during the 1-week drug period (sessions II−IV). Note that there were similarly no effects on affective state questionnaires, as repeatedly assessed over the course of the study (cf. Supplementary Table 1). b Immediate mood responsivity, as a function of WoF outcome, i.e., mean mood change from block 1 to block 2. WoF outcome profoundly affected emotional state, such that subjects reported to feel happier in the post-WoF block (block 2) after a win and less happy after a loss. However, immediate mood responsivity was unaffected by SSRI treatment. c Delayed mood responsivity, as a function of WoF outcome, i.e., mean mood change from block 1 to block 3. Subjects showed an ongoing mood response during test block (block 3), with elevated mood after a winning draw, and decreased mood after a losing draw. However, delayed mood responsivity was modulated by treatment, such that there was a greater positive mood effect in the SSRI group, and a greater negative mood effect in the placebo group (F1,62 = 5.6, p = 0.020, n = 64). d Interaction between affect and reward sensitivity. Rewards impact on emotional state, which, in turn, impacts on perception of subsequent rewards. Biased perception of reward, a downstream effect of a serotonergic impact on affect-learning dynamics (cf. Fig. 3b, c), can result in a delayed mood impact of the WoF that was only evident after subjects had played an additional block in which they received more rewards. e Relationship between positive mood perception bias and delayed positive mood response. Subjects with a high positive mood bias parameter, derived from the computational model, showed a strong delayed positive mood response (F1,30 = 14.6, p < 0.001), an effect not present in subjects with a low mood bias (division using a median split across the whole sample), and that significantly differed between the two groups (F1,60 = 9.9, p = 0.002, n = 62). Note that this positive mood bias parameter acts via boosting subjectively perceived reward during learning after winning the WoF, an effect significantly enhanced in the SSRI as compared to the placebo group. ***p < 0.001, **p < 0.01, *p < 0.05, n.s. = not significant (no difference across drug groups). Error bars indicate SEM.
Fig. 5 SSRIs induce mood effects via affect-learning interaction. a Mean mood ratings during the first block of every session, i.e., prior to any mood manipulation. Mood self-reports did not differ between groups in the pre-drug session (session I), or during the 1-week drug period (sessions II−IV). Note that there were similarly no effects on affective state questionnaires, as repeatedly assessed over the course of the study (cf. Supplementary Table 1). b Immediate mood responsivity, as a function of WoF outcome, i.e., mean mood change from block 1 to block 2. WoF outcome profoundly affected emotional state, such that subjects reported to feel happier in the post-WoF block (block 2) after a win and less happy after a loss. However, immediate mood responsivity was unaffected by SSRI treatment. c Delayed mood responsivity, as a function of WoF outcome, i.e., mean mood change from block 1 to block 3. Subjects showed an ongoing mood response during test block (block 3), with elevated mood after a winning draw, and decreased mood after a losing draw. However, delayed mood responsivity was modulated by treatment, such that there was a greater positive mood effect in the SSRI group, and a greater negative mood effect in the placebo group (F1,62 = 5.6, p = 0.020, n = 64). d Interaction between affect and reward sensitivity. Rewards impact on emotional state, which, in turn, impacts on perception of subsequent rewards. Biased perception of reward, a downstream effect of a serotonergic impact on affect-learning dynamics (cf. Fig. 3b, c), can result in a delayed mood impact of the WoF that was only evident after subjects had played an additional block in which they received more rewards. e Relationship between positive mood perception bias and delayed positive mood response. Subjects with a high positive mood bias parameter, derived from the computational model, showed a strong delayed positive mood response (F1,30 = 14.6, p < 0.001), an effect not present in subjects with a low mood bias (division using a median split across the whole sample), and that significantly differed between the two groups (F1,60 = 9.9, p = 0.002, n = 62). Note that this positive mood bias parameter acts via boosting subjectively perceived reward during learning after winning the WoF, an effect significantly enhanced in the SSRI as compared to the placebo group. ***p < 0.001, **p < 0.01, *p < 0.05, n.s. = not significant (no difference across drug groups). Error bars indicate SEM.CC BY · PubMed Central

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Show all 57 references

Research works printed but not linked: 2

  1. Frijda, N. H. in Handbook of Emotions (eds Lewis, M. & Haviland, J. M.) 381–403 (Guilford Press, 1993).
  2. Bishop, C. M. Pattern Recognition and Machine Learning (Springer, 2006).

Cited by

See all 80

Metadata

Authors
Michely, J., Eldar, E., Martin, I. M., & Dolan, R. J.
Affiliations
  • Wellcome Centre for Human Neuroimaging (GB)
  • Hebrew University of Jerusalem (IL)
Corresponding author
Michely, Jochen
Venue
Nature Communications
Published
2020-05-11
Volume
11
Issue
1
Pages
2335-2335
ISSN
2041-1723 (ISSN-L)
OpenAlex
W3023864941
Cited by
72
References
59
Citations by year
2026: 7 · 2025: 8 · 2024: 21 · 2023: 11 · 2022: 10 · 2021: 12 · 2020: 3
FWCI
4.13
Type
journal article
Language
en
Open access
Licence
Locations
  • Nature Communications (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • RePEc: Research Papers in Economics (submittedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
  • UCL Discovery (University College London) (submittedVersion)
  • MPG.PuRe (Max Planck Society) (submittedVersion)
Funders
Deutsche Forschungsgemeinschaft (MI 2158/1-1)
Topics
  • Life Sciences › Neuroscience › Cellular and Molecular Neuroscience › Neurotransmitter Receptor Influence on Behavior
  • Health Sciences › Medicine › Pharmacology (medical) › Treatment of Major Depression
  • Life Sciences › Neuroscience › Cognitive Neuroscience › Neural and Behavioral Psychology Studies
Keywords
serotonin, mood, selective serotonin reuptake inhibitors, antidepressant mechanism, reward learning, memory reconsolidation, positive affect, computational modelling, mood induction, treatment response
MeSH terms
Adult, Affect, Female, Humans, Learning, Male, Reward, Serotonin, Task Performance and Analysis, Selective Serotonin Reuptake Inhibitors, Young Adult

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: A mechanistic account of serotonin’s impact on mood. https://app.cochat.ai/papers/10.1038/s41467-020-16090-2
The paper
Michely, J., Eldar, E., Martin, I. M., & Dolan, R. J. (2020). A mechanistic account of serotonin’s impact on mood. Nature Communications. https://doi.org/10.1038/s41467-020-16090-2

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.