Pre‐planned reappraisal via implementation intention down‐regulates emotional experiences of youth with self‐reported depressive symptoms Authors: Shengdong Chen, Li Luo, Xiling Zhang, Jiemin Yang, Tao Wang, Jiajin Yuan Published: 2026 ABSTRACT ABSTRACTBackgroundCognitive reappraisal is an emotion‐regulation strategy that involves altering the emotional impact of a situation by changing its appraisal. However, individuals with depressive symptoms often struggle to deploy momentary cognitive reappraisal during emotional encounters due to impaired cognitive control. Pre‐planned reappraisal formed through implementation intentions, that is, a strategy rehearsed in advance to activate automatically upon emotional triggers, may offer a more efficient alternative.AimsThis study examined whether pre‐planned cognitive reappraisal via implementation intentions more effectively reduces negative emotional experience in youth with depressive symptoms than momentary reappraisal, and whether these effects persist after short‐term re‐exposure.MethodsThis randomised, single‐blind and three‐arm experiment was conducted from 10 December 2021 to 19 December 2021 at Southwest University in China. Following screening with the Beck Depression Inventory‐II (BDI‐II), 64 undergraduate students with depressive symptoms (score ≥ 20) and 64 without depressive symptoms (score ≤ 10) were enrolled. Participants were randomly allocated to one of three experimental conditions: passive watching of emotional stimuli, momentary reappraisal (reinterpreting the meaning of a sad face upon presentation) or pre‐planned reappraisal (applying a pre‐trained reappraisal framework before stimulus presentation). The formal experiment included an emotion elicitation and regulation task, a 20‐min interval during which participants completed questionnaires and a re‐exposure phase in which all groups passively viewed novel sad faces. Subjective emotional experiences were measured using the nine‐point self‐assessment manikin (SAM).ResultsPre‐planned reappraisal produced significantly less negative valence and lower arousal ratings for sad faces than passive watching in both participants with and without depressive symptoms. Momentary reappraisal only reduced negative emotional ratings in participants without depressive symptoms, but not those with depressive symptoms. During re‐exposure, only prior pre‐planned reappraisal yielded sustained reductions in negative emotional ratings. Participants with depressive symptoms also reported greater perceived benefit from pre‐planned than from momentary reappraisal.ConclusionsCompared with momentary reappraisal, pre‐planned cognitive reappraisal via implementation intentions produced stronger and more short‐term sustained effects among youth with depressive symptoms, supporting its potential use in low‐cost self‐regulation interventions and clinical cognitive therapies.Keywords: cognitive reappraisal, depressive symptoms, emotional experience, pre‐planned emotion regulation, sad INTRODUCTION Individuals with depressive symptoms are characterised by cognitive control deficits, which broadly range from decreased attention and working memory to impaired cognitive flexibility. Compared with individuals without depressive symptoms, they exhibit a stronger attentional bias towards negative information. 1 They also find it more difficult to remove irrelevant negative material from working memory 2 and switch between different experimental tasks. 3 Consequently, people with depressive symptoms have difficulties in momentary emotion regulation, which requires a high degree of cognitive control. 4 For example, people with depressive symptoms fail to use momentary emotion regulation strategies to suppress their amygdala responses to emotional stimuli. 5 Automatic emotion regulation refers to changes in emotional responses that can be triggered by environmental cues with little deliberate effort at the time emotion is experienced. 4 Importantly, in commonly used automatic emotion regulation paradigms, this automaticity applies primarily to the execution phase rather than the entire regulation process. Specifically, individuals first invest cognitive effort during the preparation phase to establish regulatory goals or strategies. Once these have been established, encountering the relevant cue automatically initiates the planned regulatory response, thereby reducing the need for deliberate control during execution. 6 , 7 , 8 , 9 Neuroimaging studies have shown that automatic emotion regulation attenuated amygdala responses without commensurate increases in activation of cognitive‐control regions, 7 , 8 and event‐related potential (ERP) evidence further indicates reduced online control‐related activity (e.g., frontal late positive potential) during implementation–intention‐based reappraisal relative to momentary reappraisal. 9 Accordingly, we here use the term ‘pre‐planned emotion regulation’ to emphasise the preparatory prospective nature of the automatic emotion regulation. Compared with momentary emotion regulation, pre‐planned emotion regulation involves relatively reduced online control demands during emotion elicitation, which may be especially relevant for individuals with depressive symptoms who often show deficits in cognitive control. 1 , 2 , 3 , 4 , 5 , 10 However, few studies have determined whether pre‐planned emotion regulation is effective for people with depressive symptoms. There was a behavioural study of pre‐planned attentional distraction in people with depressive symptoms. 11 In this study, an implicit distraction goal was primed using a scrambled‐sentence task, after which the goal‐priming effect was evaluated in a frustrating situation. It was found that pre‐planned attentional distraction down‐regulated negative emotional experiences in individuals with depressive symptoms and enhanced their positive emotional experiences. However, implicit emotion regulation focuses on modulating emotional responses without informing participants of the actual experimental goal, that is, emotion regulation. Therefore, the implicit form of pre‐planned emotion regulation may not be suitable for common intrapersonal emotion regulation, in which people are generally aware of their emotion‐regulatory aims. 12 Moreover, although attentional distraction has been shown to have a good short‐time regulatory effect, participants subjected to distraction (but not reappraisal) exhibited higher negative emotional responses than those in the passive watching condition upon re‐exposure to the same negative situations. 13 Pre‐planned reappraisal strategies formed through implementation intentions may offer a more efficient and sustainable choice for people with depressive symptoms. Cognitive reappraisal is an emotion‐regulation strategy in which a person modifies the meaning they assign to an event, thereby altering its emotional impact, 14 and it is one of the core strategies used in cognitive behavioural therapy (CBT). Cognitive reappraisal has been shown to effectively regulate subjective, physiological and behavioural emotional responses, 15 , 16 while producing fewer maladaptive emotional or social outcomes than those using attentional distraction (e.g., shifting attention away from the emotional situation) 13 or expressive suppression (e.g., concealing emotional expressions). 17 , 18 Pre‐planned reappraisal through implementation intentions means that a person decides in advance how they will reinterpret an emotional situation, and they link that reinterpretation to a specific cue using an if–then plan. In other words, individuals practice the reappraisal before encountering emotional stimuli so that the planned interpretation is automatically activated when the cue appears. Similar to the implicit emotion‐regulation paradigms, the pre‐planned reappraisal paradigm also involves a practice phase and an implementation phase. However, unlike the implicit emotion‐regulation paradigms, goal intention (i.e., ‘I want to stay calm’) and implementation intentions (i.e., ‘If I see a negative image, then I will reinterpret it as staged or less threatening.’) are consciously and explicitly formed during the practice phase. 19 , 20 Therefore, implementation intentions allow individuals to consciously establish their own emotion‐regulation goals and strategies before encountering emotional stimuli. Furthermore, emotion regulation strategies based on implementation intentions have been shown to effectively reduce participants' subjective experiences and physiological responses to negative stimuli without increasing their cognitive load. 9 , 21 , 22 Therefore, our aim was to test the immediate and lasting regulatory effects of pre‐planned reappraisal based on implementation intention using a two‐phase experimental design. In the first phase (Phase 1), participants with or without depressive symptoms were subjected to one of the three conditions: the passive‐watching, momentary or pre‐planned reappraisal tasks. Operationally, the key distinction between momentary and pre‐planned reappraisal is when the reappraisal is initiated. In the momentary reappraisal condition, participants received an explicit instruction to implement the reappraisal strategy while viewing the emotional stimuli. In contrast, in the pre‐planned reappraisal condition, participants formed and rehearsed a goal intention and an if–then plan during the preparation phase, but then later received only a neutral viewing instruction during the presentation of emotional stimuli. Consequently, during the emotional presentation, reappraisal in the pre‐planned reappraisal condition was self‐initiated through the cue‐triggered enactment of the previously formed plan. Twenty minutes later, all three groups were subjected to the emotional re‐exposure task (Phase 2), in which they only passively watched emotional or neutral pictures. We hypothesised that in Phase 1, pre‐planned reappraisal based on implementation intention would effectively reduce sadness experiences in both participants with and without depressive symptoms. In contrast, momentary cognitive reappraisal would be effective only in participants without depressive symptoms but not in those with depressive symptoms. Furthermore, pre‐planned reappraisal has been reported to produce more lasting emotion regulation effects than momentary reappraisal at both behavioural and physiological levels. 9 , 23 Accordingly, we also hypothesised that pre‐planned reappraisal, but not momentary reappraisal, would generate lasting regulatory effects in participants with depressive symptoms during Phase 2. METHODS Participants and selection criteria This randomised single‐blind and three‐arm experiment was conducted from 10 December 2021 to 19 December 2021 at Southwest University in China. Participants were recruited via the university campus BBS (an online bulletin board system) and printed recruitment flyers, both of which included Quick‐Response (QR) codes that linked to the online screening survey. A total of 640 undergraduate students scanned the QR code and completed a baseline demographic information form and the Beck Depression Inventory‐II (BDI‐II). 24 The BDI‐II required participants to self‐report their emotional feelings within the most recent month. After excluding questionnaires with incomplete information, 627 valid questionnaires were obtained. The following participants were included: participants with scores of 20 or above were classified into the depressive symptoms group, and those with scores of 10 or below were classified into the control group without depressive symptoms. Participants with the following characteristics were excluded: a family or personal history of mental illness, alcohol or substance addiction, organic brain disorder, poor current physical health (e.g., caught a cold) or recent experiences of traumatic events (e.g., death of relatives). Those who had previously participated in similar experiments of emotion regulation were also excluded. Before the formal experiment, all participants provided written informed consent and were informed that they would view both emotional and neutral facial stimuli and that they had the right to withdraw from the experiment at any time. Sample size estimation For each group, we set the desired sample size to 20 with reference to two recent behavioural studies that used BDI‐II to prescreen participants with depressive symptoms. 11 , 25 Given that some participants could not arrange their time schedules to attend the formal experiment, the final sample comprised 128 participants (64 with depressive symptoms and 64 without); all were right‐handed with normal or corrected vision (mean (standard deviation (SD)) age = 19.86 (1.08); males/females = 18/110), with a 1:1:1 allocation ratio to three groups. Randomisation The 128 participants were then randomly allocated to one of the three different self‐regulation conditions using a pseudorandom number sequence (figure 1): passive‐watching (n = 43, n control = 22 and n depression = 21), momentary reappraisal (n = 43, n control = 22 and n depression = 21) and pre‐planned reappraisal (n = 42, n control = 20, n depression = 22). There was no significant difference in the ratio of men to women between experimental groups (χ 2 (5, N = 128) = 1.04 and p = 0.96). Participants were blind to the study hypotheses and other intervention conditions, and outcome assessors and data analysts were blinded to group allocation until the completion of data analyses. Consolidated Standards of Reporting Trials flowchart of the study. MATERIALS Sad faces were selected as experimental stimuli, because sadness is the affective state most chronically experienced and least effectively regulated by individuals with depressive symptoms, 5 , 25 thus representing the most ecologically valid negative emotion for investigating emotion regulation deficits in this population. Sad stimuli, in turn, preferentially capture the attention of depressed individuals and elicit significantly amplified subjective and neural responses relative to healthy controls. 25 , 26 Therefore, the use of sad faces strengthened the internal validity of our experimental design and enabled more reliable detection of group differences in emotion regulation efficacy. Specifically, 60 face pictures were selected from the Chinese facial affective picture system (CFAPS), 27 including 30 sad and 30 neutral faces. Because CFAPS does not provide valence and arousal scores for these pictures, we derived these scores from a pilot study involving 20 healthy undergraduates. We used the nine‐point self‐assessment manikin (SAM) scales to measure participants' negative emotional experiences (valence 1–9, very pleasant to very unpleasant; arousal 1–9, very calm to very excited). SAM is a classic pictorial assessment tool suitable for people with different cognitive and cultural levels. 28 The pilot study confirmed that sad faces (Valence/Arousal, 6.96/6.05 (0.49/0.99)) were significantly more unpleasant and had stronger arousals than neutral faces (Valence/Arousal, 5.18/4.23 (0.34/1.17)), t valence = 14.62, t arousal = 10.41, both p < 0.001. Design and procedure We employed a 3 × 2 × 2 mixed experimental design. Instruction type (passive watching, momentary reappraisal and pre‐planned reappraisal) and group (participants with vs. without depressive symptoms) were between‐subject factors, whereas picture type (sad or neutral) was a within‐subject factor. The experiment consisted of two phases, each including 15 sad and 15 neutral face pictures. All stimuli were presented using E‐prime 2.0 on a 19.5‐inch colour Liquid Crystal Display (LCD) Screen. Before the main tasks, participants completed baseline questionnaires assessing emotion‐ and regulation‐related characteristics. State and trait anxiety were assessed using the Spielberger state–trait anxiety inventory (STAI), which consists of a 20‐item state anxiety inventory (SAI) and a 20‐item trait anxiety inventory (TAI) rated on 4‐point Likert scales. 29 Habitual use of reappraisal strategy was assessed with the cognitive reappraisal subscale of the emotion regulation questionnaire (ERQ). 30 Participants were then randomly assigned to one of three instruction conditions and first completed two practice trials (one sad and one neutral): (1) The passive watching group was instructed to view each picture naturally, allowing emotions to arise spontaneously; (2) the momentary reappraisal group was instructed to reinterpret the faces from a director's perspective to reduce negative affect, that is, treating them as emotions portrayed by actors rather than real emotions; and (3) the pre‐planned reappraisal group was instructed to form a goal intention (‘I will not feel sad’) and an implementation intention (‘If I see a sad face, then I will view it from a director's perspective’). Participants silently rehearsed these statements until they reported readiness. Potential harms were defined as any increase in acute emotional distress during or after the task (e.g., crying, early termination or verbal reports of strong discomfort). These were assessed non‐systematically through experimenters' observations during the tasks and participants' self‐reports after the experiment. No adverse events were reported. In Phase 1, the passive watching and pre‐planned reappraisal groups received identical neutral instructions (‘Please view neutral and sad faces carefully and allow emotions to arise naturally’), while the momentary reappraisal group continued with their specific instruction. Each trial began with a fixation cross (3–7 s, mean 4 s), followed by a sad or neutral face (4 s), after which participants rated their emotional responses using the SAM. After Phase 1, participants completed personality questionnaires (∼20 min). Participants who finished the questionnaires earlier rested quietly with their eyes closed until 20 minutes had elapsed, ensuring that Phase 2 began 20 minutes after the completion of Phase 1 for all participants. In Phase 2, all participants received the same instructions as those given to the passive watching group in Phase 1. To minimise habituation, the stimuli used during Phase 2 were not shown in the previous phase. Primary outcomes The primary outcome was participants' subjective emotional experience during the picture‐viewing tasks. After each trial, participants rated emotional valence and arousal using the nine‐point SAM. Secondary outcomes Two post‐task questions were administered at the end of Phase 2 to assess participants' perceived regulation processes. Both questions were rated on a nine‐point agreement scale to indicate how much participants agreed with each description of their task experience (1 = strongly disagree and 9 = strongly agree). The first question assessed perceived practice efficacy by asking participants, ‘I think the practice I received reduced my negative emotional experience’. The second question assessed the use of the perspective‐taking reappraisal by asking, ‘During the task, I viewed the pictures from the perspective of a director’. In addition, we administered a single non‐strategy‐specific engagement item (‘I tried to reduce my negative emotional experience’) to index perceived emotion‐regulation implementation during the task; because this item may reflect multiple regulation strategies beyond reappraisal, it was treated as exploratory and is reported in the Supporting Information S1. Statistical analysis All statistical analyses were performed using SPSS version 26.0 (IBM Corp). Mixed‐design ANOVAs were conducted with instruction type (passive watching, momentary reappraisal and pre‐planned reappraisal) and group (participants with depressive symptoms vs. controls) as between‐subject factors, and image type (sad or neutral) as a within‐subject factor. The primary outcomes were subjective ratings of emotional valence and arousal, and the secondary outcomes were scores on two post‐task questions assessing perceived practice efficacy and use of perspective‐taking. Where significant main effects or interactions were observed, Bonferroni‐corrected post hoc comparisons were conducted. Partial eta squared (η p 2) was reported as a measure of effect size. We used p < 0.05 as the threshold for statistical significance; results with 0.05 < p < 0.1 are described as trend‐level and interpreted with caution. RESULTS Baseline characteristics The two‐way ANOVAs with group and instruction type as factors revealed no significant interaction effects between group and instruction type (Fs(2, 122) = 0.13–1.32) and no significant main effects of instruction type (F(2, 122) = 0.11–2.39 and all p > 0.05) on the scores of Spielberger STAI, 29 positive and negative affect schedule (PANAS), 31 reappraisal subscale of ERQ, 30 BDI‐II and age, suggesting that there were no significant differences in emotional baselines among the three groups. There was a significant main effect of group on the SAI scores (F(1, 122) = 219.07, p < 0.001 and η p 2 = 0.64), TAI (F(1, 122) = 202.56, p < 0.001 and η p 2 = 0.62), the PANAS positive affect scores (F(1, 122) = 40.99, p < 0.001 and η p 2 = 0.25), the PANAS negative affect scores (F(1, 122) = 83.33, p < 0.001 and η p 2 = 0.41), the BDI‐II (F(1, 122) = 709.30, p < 0.001 and η p 2 = 0.85) and the ERQ reappraisal subscale (F(1, 122) = 20.29, p < 0.001 and η p 2 = 0.16). Specifically, compared with those without depressive symptoms, participants with depressive symptoms scored higher on STAI, the negative affect subscale of PANAS and the reappraisal subscale of ERQ, and they scored lower on the positive affect subscale of PANAS (table 1). TABLE 1Baseline demographic and emotional characteristics by groupItemsDepressive symptom groupControl groupPWMRPRPWMRPRAge19.71 (1.12)20.29 (0.72)20.18 (0.96)19.73 (1.12)19.64 (1.39)19.61 (0.95)Sex (M/F)3/183/183/192/203/194/16BDI‐II24.14 (3.54)26.81 (6.99)26.95 (6.14)4.00 (3.04)4.50 (3.14)3.60 (3.42)TAI52.33 (6.33)55.29 (5.51)54.77 (7.49)35.14 (6.56)37.14 (7.98)38.55 (6.47)SAI51.48 (8.04)53.29 (6.56)53.55 (8.50)31.77 (5.60)35.32 (6.89)34.10 (7.69)ERQ reappraisal15.67 (4.95)14.52 (3.76)14.86 (4.78)11.68 (3.49)12.00 (4.65)10.65 (4.44)PANAS positive affect24.24 (6.83)23.33 (7.29)23.27 (6.33)32.59 (7.83)30.18 (5.18)30.90 (6.56)PANAS negative affect26.05 (8.63)24.95 (8.49)25.45 (7.15)13.82 (4.34)15.14 (4.45)15.65 (4.90) Note: Values represent mean (SD).Abbreviations: BDI‐II, Beck Depression Inventory‐II; ERQ, emotion regulation questionnaire; MR, momentary reappraisal; PANAS, positive and negative affect schedule; PR, pre‐planned reappraisal; PW, passive watching; SAI, state anxiety inventory; TAI, trait anxiety inventory. Baseline demographic and emotional characteristics by group Note: Values represent mean (SD). Abbreviations: BDI‐II, Beck Depression Inventory‐II; ERQ, emotion regulation questionnaire; MR, momentary reappraisal; PANAS, positive and negative affect schedule; PR, pre‐planned reappraisal; PW, passive watching; SAI, state anxiety inventory; TAI, trait anxiety inventory. Short‐time regulatory effects of pre‐planned and momentary reappraisal in Phase 1 Arousal ratings. The mixed ANOVAs showed significant main effects of image type (F(1, 122) = 79.92, p < 0.001 and η p 2 = 0.40), group (F(1, 122) = 9.14, p = 0.003 and η p 2 = 0.07) and instruction type (F(2, 122) = 6.42, p = 0.002 and η p 2 = 0.10). Post hoc pairwise comparisons revealed that sad faces (4.84 (1.63)) elicited higher arousal ratings than neutral faces (4.22 (1.24), t(127) = 8.94 and p < 0.001); participants with depressive symptoms (4.84 (1.00)) reported higher arousal than the control participants (4.23 (1.43), t(126) = 3.02 and p = 0.003). For instruction type, both pre‐planned reappraisal (4.08 (1.26), t(83) = 3.58 and p < 0.001) and momentary reappraisal (4.51 (1.30), t(84) = 1.91 and p = 0.055, trend‐level) led to lower arousal ratings than passive watching (5.00 (1.09)). None of the two‐way interactions was significant. However, the image type × group × instruction type interaction was significant, F(2, 122) = 3.74, p = 0.027 and η p 2 = 0.06, whereas, again, all two‐way interaction effects were non‐significant (all p > 0.05). Further simple effect analysis showed that among participants with depressive symptoms, only pre‐planned reappraisal (4.59 (1.14), t(41) = 2.54 and p = 0.012), but not momentary reappraisal (5.25 (1.12), t(40) = 0.92 and p = 0.360), led to significantly lower arousal ratings of sad pictures than passive watching (5.63 (0.83)). In contrast, for participants without depressive symptoms, both pre‐planned reappraisal (4.06 (1.55), t(40) = 2.73 and p = 0.007) and momentary reappraisal (4.35 (1.63), t(42) = 2.04 and p = 0.043) led to significantly lower arousal ratings of sad pictures than passive watching (5.18 (1.52)) (figure 2A). For neutral pictures, arousal ratings did not differ across instruction conditions overall (p > 0.05), except that among participants without depressive symptoms, pre‐planned reappraisal yielded lower arousal ratings for neutral pictures than passive watching (t(40) = 2.35 and p = 0.020). Emotional arousal and valence ratings for sad faces by group (depressive symptoms vs. controls) and condition. Error bars represent standard error; *p < 0.05; **p < 0.001. MR, momentary reappraisal; PR, pre‐planned reappraisal; PW, passive watching. Valence ratings. The mixed ANOVAs also showed significant main effects of image type (F(1, 122) = 70.53, p < 0.001 and η p 2 = 0.37), group (F(1, 122) = 4.15, p = 0.044 and η p 2 = 0.03) and instruction type (F(2, 122) = 9.33, p < 0.001 and η p 2 = 0.13). Planned pair‐wise comparisons showed that participants with depressive symptoms (5.53 (0.68)) reported significantly more negative valence ratings than the control participants (5.29 (0.74), t(126) = 2.04 and p = 0.044); sad pictures (5.78 (1.01)) were scored significantly more negative than neutral pictures (5.04 (0.57), t(127) = 8.40, p < 0.001) and both pre‐planned reappraisal (5.14 (0.64), t(83) = 3.82 and p < 0.001) and momentary reappraisal (5.33 (0.70), t(84) = 2.92 and p = 0.004) led to significantly less negative valence ratings than passive watching (5.75 (0.68)). The image type × instruction type interaction was significant, F(2, 122) = 4.33, p = 0.015 and η p 2 = 0.07. Simple effect analyses showed that, for sad faces, both pre‐planned reappraisal (5.37 (0.95), t(83) = 3.91 and p < 0.001) and momentary reappraisal (5.66 (1.11), t(84) = 2.78 and p = 0.006) led to significantly less negative valence ratings than passive watching (6.29 (1.05)) (figure 2B). For neutral faces, pre‐planned reappraisal (4.92 (0.59), t(83) = 2.53 and p = 0.012) also yielded significantly fewer negative ratings than passive watching (5.21 (0.59)). The image type × group × instruction type interaction was not significant (F(2, 122) = 0.94, p = 0.393). Lasting effects of pre‐planned and momentary reappraisal in Phase 2 To examine the short‐term sustained emotion‐regulatory effects in Phase 2 (i.e., a 20‐min delayed re‐exposure phase following the initial emotion regulation task), emotional arousal and valence ratings were analysed using three‐way mixed ANOVAs with group, instruction history and image type as factors. Arousal ratings. The mixed ANOVA showed significant main effects of image type (F(1, 122) = 89.52, p < 0.001 and η p 2 = 0.42), group (F(1, 122) = 12.24, p = 0.001 and η p 2 = 0.09) and instruction history (F(2, 122) = 5.85, p = 0.004 and η p 2 = 0.09). The image type × group × instruction history interaction was significant (F(2, 122) = 3.93, p = 0.022 and η p 2 = 0.06). Simple effect analysis showed that for the control participants, only pre‐planned reappraisal (3.76 (1.87), t(40) = 3.02, p = 0.003), but not momentary reappraisal (4.49 (1.69), t(42) = 1.56 and p = 0.12), led to significantly lower arousal of sad pictures than passive watching (5.21 (1.64)). For the participants with depressive symptoms, pre‐planned reappraisal (4.58 (1.28), t(41) = 1.85 and p = 0.067), but not momentary reappraisal (5.65 (1.42), t(40) = 0.44 and p = 0.66), led to lower arousal of sad pictures than passive watching (5.44 (1.17)) (figure 2C). For neutral pictures, pre‐planned reappraisal also yielded lower neutral‐picture arousal than passive watching (t(40) = 2.58, p = 0.010) and momentary reappraisal (t(40) = 2.34 and p = 0.020) among participants without depressive symptoms, with no other significant differences (see the Supporting Information S1 for details). Valence ratings. The mixed ANOVA showed significant main effects of image type (F(1, 122) = 133.00, p < 0.001 and η p 2 = 0.52), group (F(1, 122) = 5.35, p = 0.022, η p 2 = 0.04) and instruction history (F(2, 122) = 5.74, p = 0.004 and η p 2 = 0.09). The image type × instruction history interaction was significant, F(2, 122) = 9.08, p < 0.001 and η p 2 = 0.13. Simple effect analysis revealed that pre‐planned reappraisal (5.46 (1.10), t(83/83) = 3.76/3.57 and all p < 0.001) generated significantly less negative valence ratings of sad pictures than both momentary reappraisal (6.28 (0.91)) and passive watching (6.30 (1.08)), with no other significant differences (figure 2D). Sensitivity analysis controlling for trait reappraisal Participants with higher depressive symptoms reported higher trait reappraisal (Emotion Regulation Questionnaire‐Chinese Revised Version [ERQ‐CR]) scores than those without symptoms, which could potentially shape our findings. To address this issue, we repeated the above repeated‐measures ANOVAs with ERQ‐CR entered as a covariate (ANCOVA) as a robustness check. The overall result pattern for instruction‐related effects and key interactions remained the same (please see the Supporting Information S1 for full ANCOVA outputs). Emotion regulation questions Two‐way ANOVAs with instruction type and group as between‐subject factors were conducted on scores from the two post‐task questions. Planned comparisons were then performed in accordance with our hypotheses. Question 1 (perceived practice efficacy). The two‐way ANOVA showed a non‐significant main effect of group (F(1, 122) = 1.29, p = 0.260) and a significant main effect of instruction type (F(2, 122) = 6.26, p = 0.003 and η p 2 = 0.09). The group × instruction interaction was trend‐level, F(2, 122) = 2.64, p = 0.076 and η p 2 = 0.04. Planned sample contrasts revealed that for participants with depressive symptoms, the pre‐planned reappraisal group (6.59 (0.31)) reported a significantly higher degree of benefit than the passive watching (4.05 (0.30), t(41) = 4.21 and p < 0.001) and momentary reappraisal groups (5.14 (0.31), t(41) = 2.01 and p = 0.046) (figure 3A). There was no significant difference among the three control groups in the degree of benefit from the practice (all p > 0.5). Agreement ratings for the two post‐task self‐report questions. (A) Interaction effect between instruction type and group on perceived practice efficacy (Question 1: ‘I think the practice I received reduced my negative emotional experience’). (B) Main effect of instruction type on perspective‐taking reappraisal use (Question 2: ‘During the task, I viewed the pictures from the perspective of a director’). (C) Main effect of group on perspective‐taking reappraisal use. Error bars represent standard error; *mean p < 0.05; **mean p < 0.001; ns, not significant. MR, momentary reappraisal; PR, pre‐planned reappraisal; PW, passive watching. Question 2 (the use of the perspective‐taking reappraisal). The two‐way ANOVA showed a significant main effect of group (F(1, 122) = 4.07, p = 0.046 and η p 2 = 0.03) and a significant main effect of instruction type (F(2, 122) = 13.58, p < 0.001 and η p 2 = 0.18). The group × instruction interaction was non‐significant, F(2, 122) = 0.36 and p = 0.70. Participants using both pre‐planned reappraisal (5.89 (0.33), t(83) = 3.68 and p < 0.001) and momentary reappraisal (4.96 (0.33), t(84) = 3.12 and p = 0.002) reported a significantly higher level of using perspective reappraisal than participants in the passive watching condition (3.43 (0.34); figure 3B). Participants with depressive symptoms reported a significantly lower level (4.37 (0.28)) of using perspective reappraisal than control participants (5.15 (0.28), t(126) = 2.02 and p = 0.046; figure 3C). DISCUSSION Main findings We here tested whether pre‐planned emotion regulation based on implementation intention yields more robust and sustained shifts in subjective valence and arousal ratings than momentary emotion regulation among youth with depressive symptoms. Consistent with previous studies, 11 , 32 participants with depressive symptoms exhibited a higher sensitivity towards negative information and more dysfunctional emotion regulation than the control participants. Specifically, participants with depressive symptoms reported higher levels of state and trait anxiety, depression, negative affect and lower positive affect, as well as experiencing more negative emotions than control participants in both Phase 1 and 2. Participants with depressive symptoms also scored lower in implementing reappraisal than control participants on the post‐experimental emotion‐regulation questions. In Phase 1, both pre‐planned and momentary reappraisal yielded less negative valence ratings of sad pictures than passive viewing in both participants with and without depressive symptoms, consistent with previous studies. 33 , 34 Among participants without depressive symptoms, both pre‐planned and momentary reappraisal also effectively reduced arousal ratings of sad pictures. Among participants with depressive symptoms, only those using pre‐planned reappraisal reported significantly decreased arousal ratings of sad pictures. In Phase 2, the regulatory effects of momentary reappraisal were no longer evident, 13 whereas the effects associated with pre‐planned reappraisal remained observable overall. However, among the depressive‐symptom group, the sad‐face arousal difference between the pre‐planned reappraisal and passive viewing was reduced to trend‐level significance (p = 0.067). In addition, instruction‐related differences were not confined to sad faces. Pre‐planned reappraisal yielded lower negative valence ratings for neutral pictures than passive viewing in Phase 1, but not in Phase 2. However, neutral‐face arousal showed no consistent instruction‐related differences, with only a limited effect observed among control participants in follow‐up comparisons. Therefore, the regulatory advantage of pre‐planned reappraisal appears to be most robust for sad faces rather than for neutral faces. This finding suggests that the advantage of pre‐planned reappraisal may be especially pronounced when individuals are less able to rely on intact online cognitive control. More precisely, the pre‐planned reappraisal effects should be interpreted in a phase‐specific manner. That is, implementation intentions may involve deliberate goal formation and rehearsal during preparation. 10 , 21 , 22 However, given that participants in the pre‐planned reappraisal received only neutral instructions during picture viewing, pre‐planned reappraisal while viewing the pictures had to be self‐initiated via the cue‐triggered enactment of a pre‐specified response. This phase‐specific feature may be particularly helpful when online cognitive control is compromised. Our two post‐experimental emotion‐regulation questions showed three converging observations. First, participants with depressive symptoms reported greater overall benefit from pre‐planned reappraisal than from momentary reappraisal or passive watching (Q1), consistent with previous findings. 35 Second, both pre‐planned and momentary reappraisal conditions led participants to report more frequent use of perspective‐taking reappraisal than passive watching did (Q2). Third, overall, participants with depressive symptoms reported a lower frequency of perspective taking than control participants. Together, Q1 may capture participants' global, subjective evaluation of the extent of the benefit they received from the pre‐planned training, whereas Q2 may index how that benefit was related to spontaneous engagement in the process of perspective‐taking reappraisal. In brief, pre‐planned reappraisal was associated with greater perceived benefit, especially among participants with depressive symptoms, and with more frequent self‐reported perspective taking. Consistent with this interpretation, additional analyses showed that the use of perspective‐taking reappraisal (Q2) positively predicted the perceived benefit (Q1) in both groups, with a stronger association observed in the depressive‐symptom group (see Supporting Information S1 for details). Limitations Several limitations should be acknowledged. First, our primary outcomes were based on self‐reported valence and arousal ratings. Although we took steps to reduce demand characteristics, subjective ratings cannot directly index physiological reactivity and cognitive effort. Future work should incorporate multi‐method measures. Second, the present sample was relatively modest by modern standards, which limits power for detecting smaller effects and may reduce the precision of effect‐size estimates. Therefore, non‐significant findings and trend‐level effects should be interpreted cautiously. Third, our assessment of sustained regulatory effects was limited to a 20‐min interval after the initial regulation task, and the exact rest duration was not precisely controlled, precluding further analysis of how rest length might affect Phase 2 outcomes. Importantly, the short follow‐up window means the observed effects only reflect short‐term sustained benefits and cannot be generalised to long‐term regulatory effects. Fourth, participants were young college students, and future studies should evaluate whether implementation–intention‐based reappraisal is similarly effective in older adults or clinical populations with broader cognitive variability. Fifth, we did not include objective measures of participant expectancy, perceived task difficulty, or experimenter demand, which are potential confounders in self‐report emotion regulation studies. Although our single‐blind design and identical neutral instructions for the pre‐planned reappraisal and passive watching groups help mitigate these concerns, we cannot fully rule out the possibility that participants in the reappraisal conditions may have adjusted their ratings to comply with perceived experimenter expectations. Future studies should include objective measures to totally exclude the problem of demand characteristics. Finally, this study focused exclusively on the down‐regulation of sadness using facial stimuli, which limits the generalisability of our findings. Our results may not directly extend to other negative emotions commonly experienced by individuals with depression, such as anger, fear or stress‐related responses, nor to other types of emotional stimuli (e.g., autobiographical memories). Future research should test the efficacy of pre‐planned reappraisal across a broader range of emotions and stimulus types to establish its generalisability as a low‐burden emotion regulation strategy. Implications Pre‐planned reappraisal based on implementation intention was associated with more robust changes in self‐reported valence and arousal ratings than momentary reappraisal, with a clear advantage observed for arousal responses to sad faces among participants with depressive symptoms, and it was also perceived as more beneficial by this group. These findings suggest that pre‐planned reappraisal based on implementation intention may be a promising low‐burden strategy for supporting emotion regulation when online cognitive control is constrained. Future research could test whether integrating pre‐planned reappraisal into CBT protocols 36 can enhance treatment outcomes for depression, ideally in clinically diagnosed samples with multi‐method measures and longer follow‐up periods. 37 AUTHOR CONTRIBUTIONS Conception and design were carried out by Shengdong Chen and Jiajin Yuan. Data collection and analysis were performed by Li Luo and Xiling Zhang. The first draft of the manuscript was written by Shengdong Chen and Li Luo. Jiemin Yang, Jiajin Yuan and Tao Wang reviewed and revised the first draft. All authors read and approved the final manuscript. FUNDING This study was supported by the Sichuan Natural Science Foundation Joint Fund for Science & Technology and Education (2025NSFSC2112), the Humanities and Social Sciences Research Project Funded by the Ministry of Education (24XJA190003) and the National Natural Science Foundation of China (NSFC31671164). CONFLICT OF INTEREST STATEMENT The authors declare no conflicts of interest. ETHICS STATEMENT This study was approved by the Ethics Committee of Southwest University (H190019) and conducted in accordance with the Declaration of Helsinki. CONSENT Informed consent was obtained from all participants. Supporting information Supporting Information S1 ACKNOWLEDGEMENTS We acknowledge participants' time and contributions to this study. Biographies Shengdong Chen obtained his PhD in Applied Psychology from Southwest University in China in 2020. He has been working as a lecturer in the School of Psychology, Qufu Normal University, China since 2021. His main research focuses on the emotion‐language interaction, its neural underpinnings, and its applications in child development. Jiajin Yuan obtained his PhD in Psychology from Southwest University, China, in 2009. He completed postdoctoral training in the Department of Psychology at the University of Chicago in the USA from 2011 to 2012. He is now a Professor of Psychology, Vice Dean, and PhD Supervisor at the Institute of Brain and Psychological Sciences, Sichuan Normal University in China, where he has worked since 2020. He is also Executive Director of the Sichuan Provincial Key Laboratory of Psychology and Behavior in Discipline Inspection and Supervision. His current main research focuses on emotion and health psychology, with particular interests in the cognitive and neural mechanisms of emotion regulation, stress and affective disorders, and the links between emotion, cognition, and mental health. Contributor Information Jiemin Yang, Email: yangjiemin85@126.com. Jiajin Yuan, Email: yuanjiajin168@sicnu.edu.cn, Email: yuanjiajin168@126.com. DATA AVAILABILITY STATEMENT The data of this study are available from the corresponding author upon reasonable request. REFERENCES Associated Data Supplementary Materials Supporting Information S1 Data Availability Statement The data of this study are available from the corresponding author upon reasonable request.