Amplification of Positivity Treatment for Anxiety and Depression: A Randomized Experimental Therapeutics Trial Targeting Social Reward Sensitivity to Enhance Social Connectedness

The study addresses persistent social disconnection in individuals with anxiety and depressive disorders, a domain that often remains impaired despite standard treatments. Theoretical and empirical work implicates hyporesponsivity of the positive valence system (PVS)—including low positive affect, reduced approach motivation, and diminished mesolimbic (striatal) activation during reward processing—as a contributor to social anhedonia and impaired connectedness. Prior research indicates that positive valence processes (e.g., approach motivation, positive affect) uniquely predict social connectedness beyond negative valence processes. The authors hypothesized that enhancing sensitivity to social rewards via a targeted psychosocial intervention (Amplification of Positivity; AMP) would engage neural mechanisms of reward (particularly striatal responses during social reward anticipation) and improve social connectedness. A secondary question examined whether a briefer 5-session AMP dose could engage targets as well as, or more efficiently than, a 10-session dose.

Evidence shows the striatum reliably engages during anticipation and receipt of social rewards across contexts (e.g., self-disclosure, shared experiences, viewing smiling faces). Depression and several anxiety disorders (notably social anxiety disorder and PTSD) feature blunted positive affect, decreased approach behavior, and reduced mesolimbic activation during reward processing, including social reward anticipation. Social anhedonia correlates with greater social impairment beyond symptom severity. First-line treatments often fail to remediate positive valence deficits, potentially explaining persistent social dysfunction. AMP was developed to directly target positive valence via positive activity interventions (savoring, gratitude, acts of kindness). In nonclinical samples, such activities engage striatal circuitry and promote dyadic approach behavior and reciprocity (e.g., gratitude). Preliminary trials in clinical samples reported large increases in positive affect and social connectedness after 10 sessions of AMP versus waitlist, with session-by-session gains in connectedness accounted for by increases in positive affect. The current trial, framed within NIMH’s experimental therapeutics paradigm, prioritizes establishing target engagement (striatal responsivity to social reward) before efficacy testing.

Design: Mechanism-focused, 3-arm, parallel-group randomized controlled trial comparing two doses of AMP (5 vs 10 sessions) to a waitlist (WL) control. Primary aim: AMP (doses combined) vs WL on neural target engagement; secondary aim: dose comparison (5 vs 10 sessions). Participants: N=68 adults with clinically elevated anxiety and/or depression and at least moderate social impairment. Randomization 1:1:1 to AMP-5 (n=23), AMP-10 (n=22), or WL (n=23), stratified by sex assigned at birth and baseline social connectedness (SCSR ≥60 vs ≤59), using randomly permuted blocks. Blinding: Experimental personnel (confederates, MRI operators) blinded to assignment; participants and clinicians unblinded. Intervention (AMP): Manualized clinician-delivered program targeting three core elements: (1) increasing exposure/responsivity to positive events (savoring, reminiscing, positive event disclosure; activity scheduling), (2) practicing gratitude (reflection and expression), and (3) engaging in kind/generous acts. The first 4 strategies were common to both doses; the 10-session protocol included additional activities (e.g., active/constructive responding, make someone else happier). Post-session supports for the 5-session arm included a 30-minute phone review at week 6 and weekly encouragement emails (weeks 7–10). WL: Pre- and post-assessments over 10 weeks; offered AMP after postassessment (data not analyzed). Procedures: Baseline self-report surveys, social affiliation task, and a separate fMRI session administering the Social Incentive Delay (SID) task. Postassessment occurred after treatment completion or ~10 weeks post-baseline for WL. Compensation provided; IRB-approved. Outcomes: Primary target engagement—fMRI striatal activation during anticipation of social reward vs implicit baseline on SID task (regions within Harvard-Oxford striatal mask: caudate, putamen, nucleus accumbens). Secondary target engagement—responses during an 18-minute standardized dyadic social affiliation task with a trained same-sex confederate: self-reported positive affect postconversation; positive facial expressions during listening; affiliative/approach behavior; desire for future interaction; respiratory sinus arrhythmia (RSA) reactivity. Exploratory outcomes—social connectedness (NIH Toolbox Friendship, Loneliness; Social Connectedness Scale–Revised [SCSR]); additional measures of social functioning, symptoms (anxiety, depression), affect, functioning, and well-being. Nonspecific treatment factors: credibility/expectancy, working alliance, homework completion; adherence monitored. Statistical analysis: Modified intent-to-treat including all randomized with baseline and ≥1 postbaseline measure. Primary fMRI analysis used AFNI (3dLME) linear mixed-effects model with group (AMP vs WL) × time (pre, post) on SID anticipation contrasts (any social reward vs implicit baseline). Cluster correction within striatal mask via permutation (AFNI 3dClustSim; voxelwise p=.005, clusterwise p=.05). Four fMRI datasets removed for poor quality (blinded). Secondary and exploratory outcomes analyzed with linear mixed-effects models including treatment arm, visit, and interaction; baseline demographics considered as covariates if imbalanced and related (not needed here). Effect sizes: Cohen’s d for change score differences with 95% CI for AMP (combined) vs WL (primary) and AMP-5 vs AMP-10 (dose comparison). Sample size planning targeted 60 evaluable participants (80% power, alpha .05, 15% attrition).

Primary neural target engagement: AMP (combined doses) produced significantly greater pre-to-post increases in striatal activation during anticipation of social rewards vs WL across multiple striatal regions: right putamen (largest cluster; d = 1.01, 95% CI 0.42–1.61), left caudate (d = 0.80, 95% CI 0.21–1.38), left putamen (two clusters; d = 0.76 and 0.68), right caudate (d = 0.73, 95% CI 0.15–1.31), and left nucleus accumbens (d = 1.18, 95% CI 0.57–1.79). Secondary target engagement (social affiliation task): Significant AMP vs WL increases in postconversation positive affect (d = 0.62, 95% CI 0.06–1.18) and social approach behaviors (d = 0.65, 95% CI 0.09–1.21). Non-significant group × time interactions for RSA reactivity (d = 0.50, 95% CI −0.06 to 1.05), positive facial expressions (d = −0.07, 95% CI −0.62 to 0.47), and desire for future interaction (d = −0.23, 95% CI −0.78 to 0.31). Exploratory social connectedness: AMP yielded larger increases in NIH Friendship (b = 3.22, SE 1.09, t57 = 2.95, p = .005; d = 0.80, 95% CI 0.24–1.37) and SCSR (b = 12.24, SE 2.98, t57 = 4.11, p < .001; d = 1.12, 95% CI 0.54–1.70), and larger decreases in Loneliness (b = −4.02, SE 1.00, t57 = −4.03, p < .001; d = −1.10, 95% CI −1.68 to −0.52). Other exploratory outcomes: AMP outperformed WL on measures of anxiety, depression, positive/negative affect, functional interference, and satisfaction with social roles/activities. Dose comparison: The 5-session AMP showed larger increases across all significant striatal clusters and social connectedness metrics compared to the 10-session dose (d range ~0.08–1.03 favoring 5-session). Completion/adherence: High completion (AMP 82%, WL 91%) and adherence; two mild adverse events in AMP-10 deemed unrelated.

Findings support the central hypothesis that directly targeting positive valence processes through AMP enhances neural responsivity to social reward, as evidenced by increased striatal activation during anticipation. This neural engagement aligns with theoretical models positing the striatum’s role in coding reward value and guiding motivated social approach, processes foundational to social connectedness. Behavioral and affective gains during controlled dyadic interaction (increased positive affect and approach behaviors) complement neural results, indicating AMP influences key proximal mechanisms of affiliation. Exploratory improvements in perceived friendship, belongingness, and reduced loneliness suggest potential downstream clinical relevance, despite the trial not being powered for efficacy. The unexpected dose effect—greater target engagement with 5 sessions—suggests parsimony and focused practice of core skills may suffice for target activation, potentially facilitating efficiency and scalability. Results highlight social reward sensitivity as a promising transdiagnostic target to remediate social disconnection in anxiety and depressive disorders.

This experimental therapeutics RCT demonstrates that AMP enhances striatal sensitivity to anticipated social rewards and improves proximal affiliative processes and perceived social connectedness in adults with anxiety and/or depression. A briefer 5-session regimen may be sufficient—and potentially more efficient—than a 10-session regimen for engaging neural targets and improving connectedness. Future research should replicate in larger samples against active controls, test whether changes in social reward responsivity mediate improvements in connectedness, refine dosing, and identify patient subgroups (e.g., high anhedonia) most likely to benefit. Expanding measurement to real-world social behavior and quality of life will clarify clinical utility.

- Small sample size and primary comparison to a waitlist control limit causal attribution beyond nonspecific factors; active control comparisons are needed. - Trial was mechanism-focused and not powered for clinical efficacy; exploratory clinical findings require confirmation. - Dose evaluation compared only two dosing levels and a specific operationalization of dose; optimal dosing and individual tailoring remain unknown. - Neural target measured using static smiling faces and a constrained laboratory interaction task, which may not fully capture dynamic real-world social cues and contexts. - Reliance on self-report for connectedness and symptom outcomes may introduce bias; ecological/behavioral measures are warranted. - Majority female sample; potential sex/gender differences in social reward sensitivity suggest the need for stratified analyses in larger samples. - Transdiagnostic sample may mask differential effects across diagnoses or symptom dimensions (e.g., anhedonia vs anxious arousal). - Four fMRI datasets were excluded for quality issues; although blinded, data loss may affect generalizability.