Plain wording

1. Being the decision-maker is unpleasant in itself, whether or not the partner ends up worse off.

   slug alt-agency-aversion-not-guiltrole hypothesis

2. The drop in happiness after a partner's bad outcome is disappointment at the participant's own outcome, not guilt.

   slug alt-guilt-effect-driven-by-own-outcomerole hypothesis

3. The imaging pipeline fails to recover known effects, so the Social-Partner differences cannot be credited to the manipulation.

   slug alt-imaging-contrast-invalidrole hypothesis

4. The model-based fMRI approach recovers no known neural signals, so regressors from the computational model cannot be trusted.

   slug alt-model-based-glm-invalidrole hypothesis

5. Participants chose inattentively or at random, so their choices say nothing about what they preferred or felt.

   slug alt-participants-insensitive-to-valuerole hypothesis

6. Choosing for someone else changes how much risk people accept, and that, not guilt, explains the differences between conditions.

   slug alt-social-context-shifts-risk-attituderole hypothesis

7. On one fit measure, the Responsibility Redux model fitted momentary happiness best in both studies.

   slug among-computational-models-fitted-momentaryrole methodologicalpanel table1

8. The anterior insula carries the guilt effect, responding more when the participant is responsible for the partner's bad outcome.

   slug anterior-insula-neural-substrate-guiltrole hypothesis

9. The anterior insula should respond more in the Social than the Partner condition, and most strongly after low outcomes.

   slug anterior-insula-tracks-guilt-insularole prediction

10. The left superior temporal sulcus tracks a partner's surprising outcomes less when the participant did not cause them.

    slug authors-suggest-left-sts-regionrole interpretation

11. The larger drop in happiness after a partner's bad outcome that the participant caused fits the idea of simple guilt.

    slug behavioural-guilt-effect-larger-happinessrole interpretation

12. The ventral striatum on both sides was more active when participants chose the risky option, in every condition, as found before.

    slug bilateral-ventral-striatum-more-activerole controlpanel fig4a

13. In both studies participants felt worse after a partner's bad outcome they had chosen, which the authors read as guilt.

    slug both-studies-participants-felt-worserole synthesis

14. Connectivity from the left insula and left superior temporal sulcus to prefrontal cortex should depend on condition and choice.

    slug connectivity-between-guilt-responsibility-related-outcome-phrole prediction

15. The right inferior frontal gyrus, in its coupling with the left anterior insula, is sensitive to guilt during social choices.

    slug connectivity-between-left-anterior-insularole interpretation

16. Deciding for both players rather than alone engaged the precuneus, left temporo-parietal junction and medial prefrontal cortex.

    slug decisions-social-compared-solo-conditionrole empiricalpanel fig4b

17. The gap between low and high outcome responses was larger in the Social condition in both insulae and middle temporal cortex.

    slug difference-response-between-low-highrole empiricalpanel app1table10

18. Individual neural guilt responses matched a previously published guilt brain signature, converging with earlier work.

    slug dot-products-between-individual-neuralrole control

19. At outcome both insulae responded more, and right middle temporal cortex less, to low outcomes in Social than Partner trials.

    slug during-outcome-phase-responses-lowrole empiricalpanel app1table9

20. Lottery rather than safe outcomes engaged the anterior insula, ventral striatum, dmPFC, right STS and frontoparietal cortex.

    slug during-receipt-lottery-versus-saferole empiricalpanel fig4d

21. On each trial participants chose between a safe and a risky option, for themselves, for both players, or the partner chose.

    slug each-trial-participants-chose-betweenrole scope

22. The findings rest on two samples of healthy adults, one behavioural and one scanned, with all brain results from the scanner.

    slug findings-rest-two-samples-healthyrole scope

23. Coupling between the guilt- and responsibility-related regions and prefrontal cortex changes with who decides and what is chosen.

    slug functional-connectivity-between-guilt-responsibility-relatedrole hypothesis

24. Left anterior insula coupling with the right inferior frontal gyrus was strongest for risky self choices and safe joint choices.

    slug functional-connectivity-between-left-anteriorrole empiricalpanel fig5

25. The guilt effect held whether the participant's own lottery outcome was high or low.

    slug guilt-effect-occurred-whether-participantrole control

26. The partner's choices were simulated by an algorithm always taking the highest expected value, the same for every participant.

    slug hold-partner-behaviour-constant-acrossrole methodological

27. The neural guilt signature did not track how strongly individual participants showed the behavioural guilt effect.

    slug individual-grbs-dot-product-values-notrole control

28. The insula responded more to a partner's low outcome when the participant had chosen, mirroring the behavioural guilt effect.

    slug insula-rois-responded-more-lowrole empiricalpanel fig4e

29. A left inferior frontal gyrus cluster showed the opposite coupling pattern with the left STS seed, but did not survive correction.

    slug left-ifg-cluster-showed-oppositerole empiricalpanel fig5s1

30. A smaller left inferior frontal gyrus cluster showed the same insula coupling pattern as the right, but did not survive correction.

    slug left-ifg-insula-connectivity-subthresholdrole empiricalpanel fig5

31. The left superior temporal sulcus tracked the participant's reward predictions from both participant and partner choices.

    slug left-superior-temporal-sulcus-clusterrole empiricalpanel fig4i

32. The Responsibility model fitted the happiness data better than every other model, including Responsibility Redux.

    slug likelihood-ratio-test-showed-responsibilityrole empiricalpanel table1

33. The happiness model with all three two-way interactions fitted better than simpler models and no worse than the fullest one.

    slug linear-mixed-model-containing-allrole methodologicalpanel app1table2

34. Being the decision-maker may itself cost happiness, a weight of responsibility.

    slug lower-happiness-when-participant-decision-makerrole interpretation

35. As a check, ventral striatum activity on both sides rose with the rewards participants expected from their chosen option.

    slug manipulation-check-bilateral-ventral-striatumrole controlpanel fig4g

36. A small left anterior insula cluster responded more to a partner's low outcome after the participant's own choice.

    slug mass-univariate-voxel-wise-analysis-found-smallrole empiricalpanel fig4f

37. Choosing for both players raised the rate of risky choices in the first study but not in the second.

    slug mixed-effects-regressions-choices-social-conditionrole empiricalpanel app1table1

38. A model-based analysis used the fitted model's reward and prediction-error terms to locate the brain regions reflecting them.

    slug model-based-glm-entered-best-fitting-computationalrole methodological

39. Model selection compared the Responsibility model against each rival model in turn, alongside the usual fit criteria.

    slug model-selection-among-happiness-modelsrole methodological

40. Momentary happiness was fitted with five computational models sharing decaying terms for rewards, expectations and surprises.

    slug momentary-happiness-modelled-five-computationalrole methodological

41. Regions sensitive to risky outcomes should respond more to a partner's reward surprises when the participant caused them.

    slug neural-substrate-tracks-participant-responsibility-2role prediction

42. A brain region tracks responsibility, representing a partner's reward surprises more strongly when the participant chose.

    slug neural-substrate-tracks-participant-responsibilityrole hypothesis

43. Sensitivity to the value difference between the options did not change across conditions in either study.

    slug no-significant-interaction-between-differencerole control

44. A left superior temporal sulcus cluster responded more to the partner's reward surprises after the participant's own choices.

    slug one-cluster-left-sts-respondedrole empiricalpanel fig4h

45. Only the precuneus and temporo-parietal junction were most active when participants chose the lottery for both players.

    slug only-precuneus-tpj-showed-positiverole empiricalpanel fig4c

46. Fitting the happiness model to simulated data recovered its parameters reliably, showing the estimates are stable.

    slug parameter-recovery-procedure-synthetic-data-generatedrole methodologicalpanel fig3s1

47. Participants were less happy when they were the one deciding, whatever the outcome.

    slug participant-happiness-lower-when-participantrole empirical

48. Among the influences on momentary happiness is the reward the partner received on the current trial.

    slug participant-momentary-happiness-varied-rewards-2role empiricalpanel fig3b,fig3f

49. Among the influences on momentary happiness is the reward the participant received on the current trial.

    slug participant-momentary-happiness-varied-rewardsrole empiricalpanel fig3a,fig3e

50. One measure of risk taking, how often the lottery was chosen, was higher deciding alone than for both, in the first study only.

    slug participants-chose-risky-option-lotteryrole empiricalpanel fig2a,fig2d

51. Participants' own reward surprises shaped their happiness more than their partner's did.

    slug participants-own-reward-prediction-errorsrole empirical

52. Participants chose the lottery more often the more its expected value exceeded the safe option.

    slug participants-probability-choosing-risky-optionrole controlpanel fig2a,fig2d

53. Risk preferences were much the same whether deciding alone or for both players, with only a hint of extra caution for both.

    slug participants-showed-very-similar-riskrole synthesis

54. A second risk measure, the fitted risk-aversion parameter, was slightly higher deciding for both than alone in Study 1 only.

    slug participants-slightly-more-risk-averserole empiricalpanel fig2c,fig2f

55. The partner's reward surprises that followed the participant's own choices helped explain the participant's momentary happiness.

    slug partner-reward-prediction-errors-resultingrole empirical

56. A pre-task icebreaker left participants rating their partners highly on every likeability measure in both studies.

    slug pre-task-icebreaker-succeeded-establishing-positiverole controlpanel app1table11

57. Earlier connectivity work linked social choice, guilt compensation and responsibility aversion to distinct network interactions.

    slug prior-functional-connectivity-work-shownrole literature-context

58. Earlier work links the anterior insula to guilt.

    slug prior-literature-documents-association-betweenrole literature-context

59. Responsibility made no difference to happiness after good outcomes for the partner, bounding the effect to bad ones.

    slug responsibility-choices-not-influence-happinessrole control

60. On explained variance, the Responsibility model beat every rival except the Guilt-envy model in the first study.

    slug responsibility-model-yielded-higher-valuesrole empiricalpanel table1

60 of 71 shown. The table has all of them, and so does the file.