Placebos without deception reduce self-report and neural measures of emotional distress

Placebo interventions can be effective for clinical and nonclinical conditions, but their widespread use is limited by the belief that deception is necessary for efficacy. Researchers have begun testing whether benefits can be achieved without deception by educating participants about placebos and how they can work even when known to be inert, leveraging expectancy mechanisms. Prior non-deceptive placebo research shows benefits primarily via self-report, with few studies including objective behavioral or biological measures and none showing direct biological effects. The present work argues that prior failures to detect biological effects may reflect choosing outcomes that are not reliably modulated by suggestion-based deceptive placebos. The authors therefore examine a domain known to be responsive to placebo suggestion—emotional distress—testing whether non-deceptive placebos reduce self-reported distress (Experiment 1) and a neural biomarker of emotional reactivity, the late positive potential (LPP; Experiment 2). They hypothesized that non-deceptive placebos would reduce negative affect and decrease sustained LPP amplitude (1000–6000 ms), while remaining agnostic about effects on early LPP (400–1000 ms).

The paper reviews evidence that non-deceptive (open-label) placebos can alleviate symptoms in conditions such as irritable bowel syndrome, chronic low back pain, experimental pain, emotional distress, psychological well-being, and sleep quality. Among 26 published non-deceptive placebo studies, only eight included objective behavioral or biological measures; only one showed behavioral effects and none showed direct biological effects. The authors note that prior biological outcomes (e.g., wound healing, inflammatory skin reactions) are not consistently responsive to deceptive, suggestion-based placebos, reducing the likelihood of observing effects with non-deceptive placebos. In contrast, emotional distress reliably shows placebo effects across autonomic and neural indices in deceptive placebo studies, motivating the current tests with self-report and the LPP as an EEG biomarker indexing attention (early LPP) and appraisal/meaning-making (sustained LPP).

Design: Two randomized between-subjects experiments compared a non-deceptive placebo intervention to a control condition during an emotional picture viewing task.

Participants: Experiment 1 recruited 68 university participants; 6 were excluded, yielding n=62 (control n=33, Mage=18.61, 39.4% female; non-deceptive placebo n=29, Mage=18.76, 34.5% female). Experiment 2 recruited 218 right-handed female participants aged 18–30 from a different university; 20 were excluded (non-native English, software error, excessive artifacts), yielding n=198 (control n=99, Mage=19.92; non-deceptive placebo n=99, Mage=19.78). Ethical approval and informed consent were obtained in both studies.

Intervention and control: All participants received a saline nasal spray. Non-deceptive placebo group: read an article explaining placebo effects and that placebos can work without deception; were explicitly told the spray was an inert placebo that could reduce negative emotional reactions if they believed it would. Control group: read an article on neural processes of pain; were told the spray improved physiological recording quality. Articles were matched in length, structure, and valence content. Control participants were blind to being in a placebo study; in the non-deceptive placebo group, participants learned the spray was a placebo at administration.

Tasks: Experiment 1: One block of 40 images (30 negative high-arousal; 10 neutral) presented in randomized order. Trial: fixation (4000 ms), image (6000 ms), fixation (4000 ms), then affect rating (up to 5000 ms). Ratings from 1 (not at all negative) to 9 (very negative). Mean ratings computed for negative and neutral images. Experiment 2: Two blocks of 30 images each (15 negative, 15 neutral; 60 total), randomized. Trial: blank (500 ms), fixation (500 ms), image (6000 ms), relax prompt (4000 ms). No self-reports during EEG recording to avoid introspective confounds. Nasal spray administered before each block.

EEG acquisition and preprocessing (Experiment 2): 64-channel Biosemi ActiveTwo, 1024 Hz sampling; mastoid reference offline; band-pass filter 0.01–20 Hz; ocular artifact correction; artifact rejection thresholds: >50 µV step between samples, >400 µV within trial, and <0.5 µV variance within 100 ms windows; baseline −500 to 0 ms relative to picture onset subtracted. LPP computed over clusters: eight topographic regions (anterior/posterior × hemisphere × inferior/superior) using specified electrode groups. Early LPP: 400–1000 ms in 300 ms epochs; Sustained LPP: 1000–6000 ms in 1000 ms epochs. CPz analyses reported as corroboration.

Analytic strategy: Experiment 1: 2 (condition) × 2 (picture type) mixed ANOVA on self-report; follow-up independent t-tests per picture type; two-tailed p<0.05, partial eta squared and Cohen’s d reported. Experiment 2 (preregistered on AsPredicted): Mixed-factorial ANOVAs combining two samples. Sustained LPP: 2 (condition) × 2 (sample) × 2 (picture type) × 5 (time bins: 1000–6000 ms) × 2 (hemisphere) × 2 (anterior/posterior) × 2 (inferior/superior). Early LPP: 2 (condition) × 2 (sample) × 2 (picture type) × 2 (time bins: 400–700, 700–1000 ms) × 2 (laterality) × 2 (anterior/posterior) × 2 (inferior/superior). Greenhouse–Geisser corrections applied where appropriate. Focus on main effects of condition and interactions involving condition robust across samples; CPz corroborative analyses. No multiple comparison adjustments for follow-ups unless noted.

Experiment 1 (self-report): Mixed ANOVA showed main effects of condition, F(1,60)=7.34, p=0.009, η²=0.109 (non-deceptive placebo reported less distress), and picture type, F(1,60)=627.25, p<0.001, η²=0.913 (negative>neutral). Condition × picture type interaction: F(1,60)=12.41, p<0.001, η²=0.171. Follow-ups: Negative images: non-deceptive placebo < control, t(60)=3.94, p=0.0002, d=1.00. Neutral images: no difference, t(60)=−0.36, p=0.72, d=−0.09.

Experiment 2 (sustained LPP, 1000–6000 ms): Main effect of condition: F(1,194)=8.98, p=0.003, η²=0.044 (non-deceptive placebo reduced sustained LPP amplitude versus control). Condition × time interaction: F(1.62,314.94)=4.58, p=0.017, η²=0.023. In the non-deceptive placebo group, sustained LPP decreased over time, F(1.73,167.98)=6.38, p=0.003; linear contrast F(1,97)=6.83, p=0.01, η²=0.066. Control group showed no change over time, F(1.53,148.63)=0.41, p=0.61. Between-group effects by time bin: d≈0.21 (ns) at 1000–2000 ms; d≈0.35 (p<0.05) at 2000–3000 ms; d≈0.45 (p<0.01 or p<0.05) at 3000–4000 ms and 4000–5000 ms; d≈0.43 (p<0.01) at 5000–6000 ms. No significant interactions with picture type, indicating a general dampening across neutral and negative images. CPz analyses corroborated the main findings.

Experiment 2 (early LPP, 400–1000 ms): Although higher-order interactions reached significance, no consistent or reliable main effect of condition was detected across analyses or at CPz (all p>0.05), indicating no robust non-deceptive placebo effect on attentional allocation.

Expectations/beliefs: Exploratory correlations found no significant relationships between beliefs/expectations and self-report (Exp 1) or sustained LPP (Exp 2) outcomes.

Findings demonstrate that non-deceptive placebos reduce both subjective reports of emotional distress (for negative images) and an objective neural biomarker (sustained LPP) indexing appraisal and meaning-making stages of emotional processing. This supports the view that non-deceptive placebo effects are genuine psychobiological effects rather than mere response bias. The neural time course shows a gradual reduction beginning around 2000–3000 ms post-stimulus and plateauing around 3000–4000 ms, suggesting modulation of appraisal/meaning-making mechanisms rather than immediate attentional processes. An observed asymmetry across measures—self-report effects limited to negative stimuli versus neural effects present for both neutral and negative—may reflect differences in timing (online neural responses versus retrospective reports) and aligns with literature showing weak coherence among self-report, behavior, peripheral physiology, and neural measures. Translationally, non-deceptive placebos could help manage emotional components of pain-related and psychiatric conditions and serve as a low-cost regulation strategy potentially distinct from cognitively demanding reappraisal. The lack of robust associations between expectations and outcomes aligns with prior open-label placebo research and may reflect limited introspective access to expectations.

Across two experiments, non-deceptive placebos reduced subjective emotional distress to negative images and decreased sustained LPP amplitude during emotional picture viewing, indicating objective psychobiological effects on appraisal-related neural processes. These results establish that open-label placebo interventions can modulate neural markers relevant to emotion regulation. Future research should test generalizability across demographics (including males and more diverse populations), clinical and nonclinical domains, additional biomarkers, and further delineate the psychological mechanisms and the role of conscious versus unconscious appraisals and expectations.

Samples were predominantly college students with limited age and ethnic diversity; Experiment 2 included only female participants, limiting generalizability and preventing assessment of sex differences. The non-deceptive placebo group could not be fully blinded by design (though control participants were blind to the placebo study), which can introduce expectancy or demand effects despite mitigating steps. No consistent effects were observed on early LPP (attentional allocation). Exploratory analyses found no robust correlations between expectations/beliefs and outcomes, and the studies did not test clinical populations. Self-report in Experiment 2 was not collected to preserve neural signals, precluding direct within-experiment comparison of subjective and neural outcomes.