Introduction
Social anhedonia, reflecting reduced enjoyment from social interaction, is elevated in autistic youth (Chevallier et al., Reference Chevallier, Grèzes, Molesworth, Berthoz and Happé2012). It is associated with higher internalizing symptoms, such as depression and anxiety (Gadow & Garman, Reference Gadow and Garman2020), in this population and accounts for greater variance in depression and social anxiety than autism symptoms (Gerber et al., Reference Gerber, Griffin, Keifer, Lerner and McPartland2025). Prior longitudinal research indicates that social anhedonia appears to be relatively stable among individuals with schizophrenia, while it is state-dependent in those experiencing depression (Blanchard, Horan, & Brown, Reference Blanchard, Horan and Brown2001). However, no longitudinal work has been conducted with autistic children. Thus, whether social anhedonia in autistic children is stable across time – and how this may impact internalizing symptoms – remains unknown.
Social anhedonia as a transdiagnostic factor
Social interaction is one of the most basic and fundamental human needs (Baumeister & Leary, Reference Baumeister and Leary1995; Van Lange & Columbus, Reference Van Lange and Columbus2021), serving as protection against physical and mental health risks associated with isolation and loneliness (Holt-Lunstad et al., Reference Holt-Lunstad, Smith, Baker, Harris and Stephenson2015; Holt-Lunstad, Smith, & Layton, Reference Holt-Lunstad, Smith and Layton2010; House, Landis, & Umberson, Reference House, Landis and Umberson1988). Social anhedonia is a clinical characteristic defined by a reduction in pleasure from interpersonal interaction (Meehl, Reference Meehl1975; Myerson, Reference Myerson1922; Ribot, Reference Ribot1897). It is an important transdiagnostic marker implicated in a wide range of psychiatric and neurodevelopmental conditions, including autism, depression, schizophrenia, and eating disorders (Barkus & Badcock, Reference Barkus and Badcock2019; Gandhi, Mote, & Fulford, Reference Gandhi, Mote and Fulford2022). Symptoms of social anhedonia are linked with altered neural response to social reward (Healey et al., Reference Healey, Morgan, Musselman, Olino and Forbes2014; Wang, Li, Nie, & Zheng, Reference Wang, Li, Nie and Zheng2020; Zhang et al., Reference Zhang, Yang, Wang, Wang, Yang, Cheung, Martin and Chan2020) and have been associated with increased risk for mood disorders and suicidality in clinical populations (Blanchard et al., Reference Blanchard, Collins, Aghevli, Leung and Cohen2009; Harrison, Mountford, & Tchanturia, Reference Harrison, Mountford and Tchanturia2014; Sagud et al., Reference Sagud, Tudor, Šimunić, Jezernik, Madžarac, Jakšić, Mihaljević Peleš, Vuksan-Ćusa, Šimunović Filipčić, Stefanović, Kosanović Rajačić, Kudlek Mikulić and Pivac2021; Tchanturia et al., Reference Tchanturia, Davies, Harrison, Fox, Treasure and Schmidt2012; Yang et al., Reference Yang, Liu, Wang, Liu and Harrison2020).
Social anhedonia in the general population increases in prevalence from childhood into adolescence (Dodell-Feder & Germine, Reference Dodell-Feder and Germine2018; Yang, Guo, Harrison, & Liu, Reference Yang, Guo, Harrison and Liu2022). In individuals with schizophrenia, symptoms of social anhedonia remain relatively stable despite symptom improvement, indicating that it may be an ongoing characteristic in these individuals (Blanchard, Horan, & Brown, Reference Blanchard, Horan and Brown2001). In contrast, depressed individuals who experience symptom improvement report decreased social anhedonia (Blanchard, Horan, & Brown, Reference Blanchard, Horan and Brown2001). This is consistent with a recent meta-analysis indicating differences in the stability of social anhedonia between schizophrenia and depression (Gandhi, Mote, & Fulford, Reference Gandhi, Mote and Fulford2022). Overall, while social anhedonia is a transdiagnostic construct, it is possible that the mechanism underlying the development and maintenance of social anhedonia may vary across psychiatric conditions.
Social anhedonia in autistic children
Early characterizations of autism posited that social anhedonia was a hallmark characteristic of autism (Kanner, Reference Kanner1943), although it has not been defined as a core diagnostic feature. On average, autistic adolescents and adults report higher symptoms of social anhedonia compared with non-autistic individuals (Berthoz, Lalanne, Crane, & Hill, Reference Berthoz, Lalanne, Crane and Hill2013; Carré et al., Reference Carré, Chevallier, Robel, Barry, Maria, Pouga, Philippe, Pinabel and Berthoz2015; Chevallier et al., Reference Chevallier, Grèzes, Molesworth, Berthoz and Happé2012; Han, Tomarken, & Gotham, Reference Han, Tomarken and Gotham2019). Parents also report elevations in social anhedonia in autistic children as young as age 6 (Gadow & Garman, Reference Gadow and Garman2020; Gerber et al., Reference Gerber, Griffin, Keifer, Lerner and McPartland2025). However, interests in social interaction are known to vary across autistic individuals with some showing a high interest in interacting with others (e.g. Meng et al., Reference Meng, Xu, Song, Shou, Wang, Han, Han and Zhang2018; Neuhaus, Bernier, & Webb, Reference Neuhaus, Bernier and Webb2021; Scheeren, Koot, & Begeer, Reference Scheeren, Koot and Begeer2020; Zhao et al., Reference Zhao, Li, Zhang, Song, Zhu, Shou, Meng, Xu, Zhang and Kendrick2024). Symptoms of social anhedonia in autistic children are associated with heightened internalizing symptoms, such as depression and anxiety (Gadow & Garman, Reference Gadow and Garman2020), and account for greater variance in depression and social anxiety than autism symptoms (Gerber et al., Reference Gerber, Griffin, Keifer, Lerner and McPartland2025), highlighting its clinical importance. In contrast, symptoms of attention-deficit/hyperactivity disorder (ADHD) appear to be negatively correlated with social anhedonia in autistic youth (Gerber et al., Reference Gerber, Griffin, Keifer, Lerner and McPartland2025). Overall, social anhedonia represents an important transdiagnostic clinical characteristic for understanding symptom presentation and with potential relevance to clinical decision-making in autistic children.
Given its concurrent relationship with internalizing symptoms, it is important to understand whether social anhedonia has predictive clinical value in autism. However, no work has explored the variability of social anhedonia across time or the longitudinal relationships between social anhedonia and internalizing symptoms in autistic children. Theoretical models of the development of internalizing symptoms in autistic children posit bidirectional relationships between social withdrawal and internalizing symptoms over time (Wood & Gadow, Reference Wood and Gadow2010; Yarger & Redcay, Reference Yarger and Redcay2020). Early symptoms of social anhedonia could lead to social withdrawal and thus represent an important link to an increased risk of internalizing symptoms. Therefore, understanding the role of social anhedonia in the developmental course of internalizing symptoms in autistic children is essential for understanding relevant timing for potential therapeutic support.
Current study
This study utilized data obtained from a well-characterized sample of autistic children and their caregivers during in-person research visits across the course of 6 months. The objective of the current set of analyses was to assess the concurrent relationships between social anhedonia and co-occurring psychiatric symptoms, estimate the 6-month stability of social anhedonia in autistic children, and examine its longitudinal relationship with internalizing symptoms. Analyses were centered on the most common co-occurring psychiatric conditions in autistic children, including anxiety, depression, and ADHD (Kerns, Rast, & Shattuck, Reference Kerns, Rast and Shattuck2020; Simonoff et al., Reference Simonoff, Pickles, Charman, Chandler, Loucas and Baird2008).
We hypothesized that social anhedonia symptoms would be significantly associated with co-occurring psychiatric symptoms in autistic children after controlling for autism symptoms, age, sex, and intelligence quotient (IQ) (Hypothesis 1). Based on prior work, we specifically hypothesized that this would be the case for symptoms of depression, social anxiety, and ADHD (Gerber et al., Reference Gerber, Griffin, Keifer, Lerner and McPartland2025). We also examined the trajectory of social anhedonia symptoms in autistic children over the course of 6 months and whether this trajectory was moderated by age, sex, or IQ (Hypothesis 2). Lastly, we hypothesized that social anhedonia would explain unique variance in the development of co-occurring psychiatric symptoms, specifically depression and social anxiety, after accounting for autoregressive and concurrent relationships (Hypothesis 3).
Methods
Participants
Participants in this study were autistic children drawn from a larger study examining biomarkers in autism, the Autism Biomarkers Consortium for Clinical Trials (ABC-CT; McPartland et al., Reference McPartland, Bernier, Jeste, Dawson, Nelson, Chawarska, Earl, Faja, Johnson, Sikich, Brandt, Dziura, Rozenblit, Hellemann, Levin, Murias, Naples, Platt and Sabatos-DeVito2020). All participants were enrolled in the study between the ages of 6 and 11. Autistic children were required to meet the criteria for autism on standardized diagnostic assessments and have a composite IQ score between 60 and 150. Exclusion criteria included known genetic conditions and significant visual, auditory, or motor challenges that would limit study participation. The final sample included 276 autistic children between the ages of 6 and 11 (M = 8.60 years, SD = 1.65). See Table 1 for the sample demographics.
Table 1. Demographic characteristics of the sample
Procedure
Participants for ABC-CT were recruited from five research centers across the United States (Boston Children’s Hospital; Duke University; University of California, Los Angeles; University of Washington; and Yale University). Children completed diagnostic and cognitive assessments during in-person site visits. Their parents and caregivers completed paper questionnaires sent out in advance and collected during the in-person visit. Data were collected at three different timepoints, baseline, 6 weeks after baseline, and 6 months after baseline. Families were given a 2-week window before and after both follow-up visits to complete data collection. Clinicians were assigned to specific families and consistently administered diagnostic interviews for that family across timepoints whenever possible. Autism diagnosis required meeting criteria on gold-standard diagnostic assessments in addition to clinical judgment based on the DSM-5 criteria. Study procedures were approved by the Yale Institutional Review Board (IRB), which operated as a central IRB for this study. All parents and caregivers provided informed consent for study participation.
Measures
Autism diagnostic status and symptom severity
The Autism Diagnostic Observation Schedule, Second Edition (ADOS-2; Lord et al., Reference Lord, Rutter, DiLavore, Risi, Gotham and Bishop2012) is a standardized diagnostic tool that was used to confirm autism diagnosis for all participants. It consists of a semistructured set of social tasks designed to elicit social behaviors as well as any present restrictive and repetitive behaviors. The ADOS-2 yields a diagnostic classification and calibrated severity score, which can be compared across modules. Most participants completed Module 3 intended for youth with fluent speech (230; 83.3%); however, 40 participants completed Module 2 (14.5%) and 6 participants completed Module 1 (2.2%). Research reliable clinicians completed all ADOS-2 assessments, and 10% of recorded videos were co-scored for ongoing evaluation of reliability.
The Autism Diagnostic Interview–Revised (ADI-R; Lord, Rutter, & Le Couteur, Reference Lord, Rutter and Le Couteur1994) is a gold-standard, structured diagnostic parent interview designed to assess historical and current behaviors related to an autism diagnosis. It provides detailed information across three core diagnostic domains: Language/Communication, Reciprocal Social Interactions, and Repetitive Behaviors/Interests. All ADI-R interviews in this study were administered by research-reliable clinicians.
The Social Responsiveness Scales, Second Edition (SRS-2; Constantino & Gruber, Reference Constantino and Gruber2012) is a parent-report questionnaire measuring autism symptom severity. It consists of 65 items rated on a 4-point Likert scale. The SRS-2 provides a sex-normed composite T-score representing overall symptom severity, with higher T-scores reflecting greater symptom severity. All caregivers completed the school-age SRS-2 form. Reliability for the SRS-2 at baseline in the data was excellent (Cronbach’s α = .94).
Cognitive assessment
Cognitive ability and information processing were assessed using the Differential Ability Scales-II (DAS-II; Elliott, Reference Elliott2007). The DAS-II is a clinician-administered assessment that produces a verbal and nonverbal IQ, as well as an age-normed standardized IQ score. All participants completed the DAS-II school-age version core battery.
Co-occurring psychiatric symptoms
Co-occurring psychiatric symptoms were assessed with the Child and Adolescent Symptom Inventory, Fifth Edition (CASI-5; Gadow & Sprafkin, Reference Gadow and Sprafkin2012). The CASI-5 is a 173-item parent-report questionnaire in which parents are asked to rate their children’s current behavior on a 4-point Likert scale from never to very often. It has been widely used to assess symptoms of autism and co-occurring psychiatric conditions in autistic youth (e.g. Brenner et al., Reference Brenner, Pan, Mazefsky, Smith, Gabriels, Siegel, Erickson, Gabriels, Kaplan, Mazefsky, Morrow, Righi, Santangelo, Wink, Benevides, Beresford, Best, Bowen and Dechant2018; Duan et al., Reference Duan, Lee, Wolf, Naples and McPartland2022; Plesa Skwerer et al., Reference Plesa Skwerer, Joseph, Eggleston, Meyer and Tager-Flusberg2019). Consistent with our prior work, this study utilized the subscales measuring ADHD, social anxiety, generalized anxiety, separation anxiety, and depression symptoms (Gerber et al., Reference Gerber, Griffin, Keifer, Lerner and McPartland2025). The CASI-5 produces a T-score for each condition that is normed by age and sex. The ABC-CT study included 280 autistic children; however, four participants were missing CASI-5 data at baseline and were therefore excluded from further analyses. Reliability at baseline for each of the five CASI-5 subscales utilized for this study was good (all Cronbach’s α > .80).
Social anhedonia
Social anhedonia symptoms were measured using a set of six items drawn from the depression, schizoid personality, and autism subscales of the CASI-5. The subscale includes one item out of the seven from the depression subscale (shows little interest in or enjoyment of pleasurable activities). It also includes all three items from the schizoid personality subscale (prefers to be alone rather than with friends or family; shows little interest in having close relationships; is emotionally cold or indifferent to people) and 2 of the 16 items from the autism subscale (not interested in making friends; is unaware or takes no interest in other people’s feelings). Items from the social anxiety subscale and those related to social skills were specifically excluded from this subscale. The chosen items reflect observable symptoms of social anhedonia and contain content consistent with typical measures of social anhedonia, such as reduced interest in close relationships and a preference for being alone (see Gadow & Garman, Reference Gadow and Garman2020). These six items were summed to create a total social anhedonia symptom score as per Gadow and Garman (Reference Gadow and Garman2020), with higher total scores indicating greater symptom severity. Following cutoff procedures defined by Gadow and Garman (Reference Gadow and Garman2020), when caregivers endorsed two or more symptoms of social anhedonia (i.e. rated them as often or very often), children were classified as being above the cutoff for social anhedonia, while children with no symptoms were classified as not meeting criteria for social anhedonia. According to scoring procedures established by Gadow and Garman (Reference Gadow and Garman2020), children whose caregivers endorsed only one symptom did not receive a classification in order to increase differentiation across groups. Therefore, these children were not included when presenting the youth meeting the cutoff for social anhedonia. However, all other analyses used a dimensional approach utilizing the continuous social anhedonia total score. This subscale has been previously used to measure social anhedonia in autistic youth (Gadow & Garman, Reference Gadow and Garman2020; Gerber et al., Reference Gerber, Griffin, Keifer, Lerner and McPartland2025). Reliability for the social anhedonia subscale at baseline was good (Cronbach’s α = .80).
Data analytic plan
Means and standard deviations for key demographic and clinical variables were computed. Next, first-order correlations at baseline were calculated between each of these key variables. All analyses were completed in R version 4.3.2 (R Core Team, 2024). To examine levels of social anhedonia in autistic children, we ran further descriptive analysis regarding the total number of children meeting criteria and the distribution of social anhedonia scores across time. The relationships between social anhedonia with age and sex were also investigated.
To test whether social anhedonia symptoms had a significant association with key co-occurring psychiatric symptoms after controlling for autism symptoms, age, sex, and IQ (Hypothesis 1), a series of five separate multiple regressions were run. In each of the regression models, autism symptoms, age, sex, and IQ were entered as the predictors, while the dependent variable was one of the five key co-occurring psychiatric symptoms.
A multilevel modeling (MLM) framework was used to examine trajectories of social anhedonia symptoms across time (Hypothesis 2). MLM analyses were run using the lme4 package in R (Bates, Mächler, Bolker, & Walker, Reference Bates, Mächler, Bolker and Walker2015) and a Full Maximum Likelihood estimation procedure. Models were compared using log-likelihood ratio tests. All continuous predictors (i.e. age and IQ) were standardized. Time was modeled as an ordinal variable with three timepoints (baseline, 6 weeks, and 6 months). First, the need for random effects was tested, with future models retaining those effects. The next timepoint was added to the model predicting social anhedonia symptoms. Finally, models including the effects of age, sex, and IQ on trajectories of social anhedonia symptoms were tested.
Structural equation models were fit to the data to examine the longitudinal relationship between social anhedonia and co-occurring psychiatric symptoms. Specifically, cross-lagged panel models were used to estimate the autoregressive, concurrent, and cross-lagged relationships between social anhedonia and co-occurring psychiatric symptoms. Importantly, by accounting for concurrent relationships, cross-lagged coefficients in the model are independent of these associations and specifically represent the directional influence of variables over time. Only co-occurring conditions that were associated with social anhedonia at baseline after controlling for autism symptoms were investigated. Each co-occurring psychiatric symptom was examined in a separate model. All cross-lagged panel analyses were completed with the Lavaan package in R (Rosseel, Reference Rosseel2012). Power analysis was conducted using the semPower package in R (Moshagen & Bader, Reference Moshagen and Bader2024). Assuming autocorrelation coefficients of .8 across waves, a sample size of 276 was sufficient for detecting a cross-lagged effect size of .1 with a power of .8.
Results
Means and standard deviations for key variables, as well as zero-order correlations, are presented in Table 2. Social anhedonia total scores were associated with clinician-rated autism severity on the ADOS-2, as well as most parent-report variables. ADOS-2 symptom severity scores were also related to lower IQ scores, CASI-5 ADHD T-scores, CASI-5 generalized anxiety T-scores, and being male.
Table 2. Means, standard deviations, and correlations of demographic and clinical variables
Note: ADOS-2 CSS, Autism Diagnostic Observation Schedule, Second Edition Calibrated Severity Score; CASI-5, Child and Adolescent Symptom Inventory, Fifth Edition; SRS-2, Social Responsiveness Scale, Second Edition.
* p < 0.05.
** p < 0.01.
Social anhedonia in autistic children
Of the 276 autistic children in the sample, 52 children had a parent report endorsing only one symptom of social anhedonia and were not provided a qualitative classification (see the Methods section for details). There were 84 autistic children out of the remaining 224 that met the criteria for social anhedonia (38%). The distribution of social anhedonia scores for autistic children is shown in Figure 1. Social anhedonia was not associated with the participant age (b = 0.21, SE = 0.13, p = 0.103) or sex (t(120.59) = 1.16, p = .250). Further, we found no differences in social anhedonia across participant ethnicity (t(77.48) = 0.85, p = .400), participant race (F(5,270) = 0.89, p = .488), or family income (F(5,270) = 1.24, p = .287).
Figure 1. Distribution of Child and Adolescent Symptom Inventory, Fifth Edition social anhedonia scores for autistic children across all three timepoints of the study.
Social anhedonia and co-occurring symptoms
Results of linear regression models predicting co-occurring psychiatric symptoms are displayed in Table 3. All overall regression models were significant, and greater autism symptoms were associated with higher symptoms for all co-occurring psychiatric conditions. Greater social anhedonia symptoms were associated with higher symptoms of social anxiety and depression after accounting for autism symptoms. Further, greater social anhedonia symptoms were related to lower ADHD symptoms after accounting for autism symptoms.
Table 3. Results of linear regression models predicting co-occurring psychiatric symptoms
Note: Regression coefficients (β) with standard errors (SE). Significant predictors are in bold. The reference group for beta coefficients is in parentheses. ADHD, attention-deficit/hyperactivity disorder.
Trajectories of social anhedonia across 6 months
Overall, missing data were limited, and attrition rates were low. Of the 276 participants with baseline data, 271 (98.19%) had CASI-5 data at a 6-week follow-up and 257 (93.16%) had CASI-5 data at a 6-month follow-up.
The overall ICC for the model of time predicting changes in social anhedonia was .77 (see model 1 of Table 4). This indicates that 77% of the variance in social anhedonia in the model was associated with between-person differences, suggesting that social anhedonia symptoms were relatively stable at the within-person level. Figure 1 displays the distribution of social anhedonia total scores at each timepoint. There was a significant decline in social anhedonia symptoms between baseline and 6-week follow-up (B = −0.52, SE = 0.14, p < .001); however, there was no difference between baseline and 6-month follow-up (see model 1 of Table 4). The trajectory of social anhedonia symptoms over time was not impacted by age, sex, or IQ (see models 2–4 of Table 4).
Table 4. Results of multilevel models predicting social anhedonia symptoms over 6 months
Note: Standardized regression coefficients (β) with standard errors (SE). The reference group for beta coefficients is in parentheses.
* p < 0.05.
** p < 0.01.
*** p < 0.001.
Longitudinal relationships between social anhedonia and co-occurring symptoms
Standardized parameter estimates for each of the cross-lagged panel models are displayed in Figures 2– 4. Symptoms of social anhedonia consistently demonstrated significant autoregressive effects between prior timepoints across all models (all βs > 0.70, all ps < 0.001). In the model examining the relationship between social anhedonia and depression (Figure 2), there were significant autoregressive effects found between depression symptoms across all timepoints (all βs > 0.59, all ps < 0.001). There was also a concurrent positive relationship between social anhedonia and depression symptoms at baseline (cov = 22.99, SE = 3.80, p < 0.001), 6-week follow-up (cov = 3.87, SE = 1.31, p = 0.003), and 6-month follow-up (cov = 6.60, SE = 1.56, p < 0.001). Finally, there was a significant cross-lagged relationship between social anhedonia symptoms at baseline and depression symptoms at 6-week follow-up (β = 0.44, SE = 0.20, p = 0.029), as well as social anhedonia symptoms at 6-week follow-up and depression symptoms at 6-month follow-up (β = 0.63, SE = 0.22, p = 0.005), such that social anhedonia predicted increased depression. There was also a cross-lagged relationship between depression symptoms at 6-week follow-up and social anhedonia symptoms at 6-month follow-up (β = 0.03, SE = 0.01, p = 0.012), such that depression predicted increased social anhedonia. However, the cross-lagged relationship between depression symptoms at baseline and social anhedonia symptoms at 6-week follow-up (β = 0.00, SE = 0.01, p = 0.953) was not significant.
Figure 2. Results of a cross-lagged panel model representing the autoregressive, concurrent, and cross-lagged relationships between social anhedonia and depression. Note: Cross-lagged effects are bolded and highlighted in yellow. Social anhedonia and depression are concurrently related to each other at all three timepoints. In addition, they demonstrate a bidirectional relationship over time, such that each is significantly related to the other at future timepoints, with the exception of baseline depression and social anhedonia at 6 weeks.
In the model examining the relationship between social anhedonia and social anxiety (Figure 3), there were significant autoregressive effects found between social anxiety symptoms across all timepoints (all βs > 0.69, all ps < 0.001). There was also a concurrent positive relationship between social anhedonia and social anxiety symptoms at baseline (cov = 22.44, SE = 3.88, p < 0.001), 6-week follow-up (cov = 5.30, SE = 1.43, p < 0.001), and 6-month follow-up (cov = 6.17, SE = 1.59, p < 0.001). Finally, there were no significant cross-lagged relationships between social anhedonia symptoms and social anxiety symptoms (all βs < 0.40, all ps > 0.067).
Figure 3. Results of a cross-lagged panel model representing the autoregressive, concurrent, and cross-lagged relationships between social anhedonia and social anxiety. Note: Cross-lagged effects are bolded and highlighted in yellow. Social anhedonia and social anxiety are concurrently related to each other at each timepoint; however, they do not significantly predict each other over time.
Lastly, in the model examining the relationship between social anhedonia and ADHD (Figure 4), there were significant autoregressive effects found between ADHD symptoms across all timepoints (all βs > 0.77, all ps < 0.001). There was also a concurrent positive relationship between social anhedonia and ADHD symptoms at baseline (cov = 8.83, SE = 2.81, p = 0.002), 6-week follow-up (cov = 3.64, SE = 1.04, p < 0.001), and 6-month follow-up (cov = 4.31, SE = 1.18, p < 0.001). Finally, there were no significant cross-lagged relationships between social anhedonia symptoms and ADHD symptoms (all βs < 0.14, all ps > 0.100).
Figure 4. Results of a cross-lagged panel model representing the autoregressive, concurrent, and cross-lagged relationships between social anhedonia and attention-deficit/hyperactivity disorder (ADHD). Note: Cross-lagged effects are bolded and highlighted in yellow. Social anhedonia and ADHD are concurrently related to each other at each timepoint; however, they do not significantly predict each other over time.
Discussion
To the best of our knowledge, this was the first study to examine longitudinal trajectories of social anhedonia and their clinical correlates in autistic children. Consistent with our first hypothesis, we found that social anhedonia symptoms were concurrently associated with internalizing symptoms at baseline in autistic children even after controlling for autism severity. Notably, symptoms of social anhedonia were similar across age, sex, race/ethnicity, and family income, underscoring its clinical utility in autistic children. We also found a relatively stable trajectory of social anhedonia symptoms in autistic children that remained consistent across demographic characteristics. Finally, we found that symptoms of social anhedonia predicted increases in depression; this was not the case for symptoms of social anxiety or ADHD. Overall, these results converge with existing research highlighting the clinical importance of social anhedonia in autistic children.
Social anhedonia in autistic children
Congruent with previous research, autistic children varied in their degree of interest in social interaction, with a little over one-third of the sample (38%) meeting the criteria for social anhedonia on parent reports. This is in line with the literature indicating elevated symptoms of social anhedonia in this population (Chevallier et al., Reference Chevallier, Grèzes, Molesworth, Berthoz and Happé2012; Gadow & Garman, Reference Gadow and Garman2020). However, these levels are below prior reports, which found that about half of autistic children met the criteria for social anhedonia (Gerber et al., Reference Gerber, Griffin, Keifer, Lerner and McPartland2025). Given that symptoms of anhedonia in children generally peak around adolescence (Dodell-Feder & Germine, Reference Dodell-Feder and Germine2018; Gupta, Eckstrand, & Forbes, Reference Gupta, Eckstrand and Forbes2024), it is likely that differences in sample age contributed to this finding. Specifically, the sample analyzed by Gerber et al. (Reference Gerber, Griffin, Keifer, Lerner and McPartland2025) included children aged 8–18, while the current study examined a sample of children aged 6–11. This difference in sample age may also explain why the study by Gerber et al. (Reference Gerber, Griffin, Keifer, Lerner and McPartland2025) found an association between social anhedonia and age when the current study did not. Nonetheless, we still found significant levels of social anhedonia in this study, suggesting that it is important to assess in younger autistic children.
Although our findings suggest a longitudinal relationship between social anhedonia and depression, depression rates are generally higher in female autistic adolescents compared with their male counterparts (Oswald et al., Reference Oswald, Winter-Messiers, Gibson, Schmidt, Herr and Solomon2016; Schwartzman, Williams, & Corbett, Reference Schwartzman, Williams and Corbett2022). Thus, it is noteworthy that we did not find comparable sex differences in social anhedonia. However, these findings are consistent with prior work (Gadow & Garman, Reference Gadow and Garman2020; Gerber et al., Reference Gerber, Griffin, Keifer, Lerner and McPartland2025), suggesting that there may be elements of multifinality, such that symptoms of social anhedonia may exhibit differential relationships with co-occurring psychiatric symptoms in autistic males and females. Alternatively, it is possible that the diagnostic criteria for depression may be less attuned to identifying depression in autistic males compared with autistic females, potentially leading to underdiagnosis in males despite similar clinical profiles (Zheng et al., Reference Zheng, Adams, Taylor, Pezzimenti and Bishop2021). Taken together, these findings highlight the need for further research to explore sex differences in co-occurring psychiatric conditions and their relationship to social anhedonia in autistic youth.
Further, our results demonstrate a cross-sectional relationship between symptoms of social anhedonia and internalizing symptoms above and beyond autism symptom severity. Specifically, we found that symptoms of social anxiety and depression were concurrently associated with greater social anhedonia, while generalized and separation anxiety were not. Overall, these findings replicate prior research (Gerber et al., Reference Gerber, Griffin, Keifer, Lerner and McPartland2025) in an independent sample and extend them to younger autistic children. Further, the results indicated that symptoms of ADHD were linked to lower social anhedonia, consistent with our prior findings (Gerber et al., Reference Gerber, Griffin, Keifer, Lerner and McPartland2025). Previous work has shown that youth with ADHD exhibit heightened responsiveness to social rewards (Kohls, Herpertz-Dahlmann, & Konrad, Reference Kohls, Herpertz-Dahlmann and Konrad2009), while co-occurring ADHD in autistic youth has been linked to a relatively heightened responsiveness to social rewards (Baumeister et al., Reference Baumeister, Moessnang, Bast, Hohmann, Aggensteiner, Kaiser, Tillmann, Goyard, Charman, Ambrosino, Baron-Cohen, Beckmann, Bölte, Bourgeron, Rausch, Crawley, Dell’Acqua, Dumas, Durston and Brandeis2023). Thus, in the context of the broader literature, our findings suggest the possibility of a clinically relevant autistic phenotype characterized by hyper- rather than hypo-responsiveness to social rewards (e.g. Yi, Wang, Song, & Han, Reference Yi, Wang, Song and Han2022).
Trajectories of social anhedonia
Results indicated that social anhedonia was consistent across 6 months in autistic youth. This suggests that it may be a relatively stable characteristic in autistic children, consistent with social anhedonia symptoms for those on the schizophrenia spectrum rather than the state-like changes seen in individuals with depression (Blanchard, Horan, & Brown, Reference Blanchard, Horan and Brown2001). However, it is also possible that 6 months may not provide enough time to effectively assess changes in social anhedonia in autistic youth. Thus, future research should examine longitudinal stability across longer periods of time. Further, additional research is needed to better understand how stable social anhedonia is across the lifespan in autism and whether specific factors contribute to variability in this characteristic over time. In the present study, there was a significant group-level decrease in social anhedonia between baseline and 6-week follow-up. It is possible that improvements in social anhedonia symptoms at 6 weeks could be a result of the sustained contact involved in participating in a longitudinal study. This suggests that external factors can influence the presence of social anhedonia in autism. Nonetheless, trajectories of social anhedonia were not different for children based on age, sex, and IQ. Thus, symptoms of social anhedonia may represent an important transdiagnostic clinical characteristic and potential intervention target in autistic children.
Further, the current study presents evidence regarding the longitudinal associations between symptoms of social anhedonia and internalizing symptoms in autistic children. Contrary to our hypotheses, our investigation did not reveal a significant predictive relationship between symptoms of social anhedonia and social anxiety over the 6-month study period. However, despite the relative stability of social anhedonia in autistic youth, findings demonstrated a bidirectional relationship between symptoms of social anhedonia and depression over time. This suggests a reciprocal influence between social anhedonia and depression, where these constructs iteratively impact each other over time. These results build upon prior research, underscoring the pivotal role of social anhedonia symptoms as a risk factor for depression in autistic children. Our findings are consistent with developmental models of internalizing symptoms in autistic children, which propose that early negative social experiences lead to social withdrawal and downstream internalizing symptoms (Wood & Gadow, Reference Wood and Gadow2010; Yarger & Redcay, Reference Yarger and Redcay2020). Overall, this highlights the critical importance of assessing symptoms of social anhedonia in autistic children and lays the groundwork for targeted interventions for autistic children.
Limitations
This study is subject to several limitations that warrant consideration. First, while the CASI-5 social anhedonia subscale has demonstrated strong internal reliability in prior use (Gadow & Garman, Reference Gadow and Garman2020; Gerber et al., Reference Gerber, Griffin, Keifer, Lerner and McPartland2025), it nonetheless has limited data in terms of psychometric validation. This limitation exacerbates challenges with overlapping items on the depression and autism subscales, restricting conclusions regarding the specificity of these results. Further, reliance on parent report introduces potential difficulties in discriminating among social anhedonia, depression, and social anxiety, given the behavioral overlap between these constructs. For example, without an understanding of the internal context, parents may conflate symptoms of social anhedonia (lack of desire for social contact) with symptoms of social anxiety (difficulty engaging in social contact). This may be particularly relevant for the youth with limited language who completed an ADOS module 1 or 2, where it may be less clear what constitutes social anhedonia. Additionally, although we examined trajectories of social anhedonia over a 6-month period, this time frame may not be sufficient to capture measurable changes. Further, it is possible that some participants received interventions during the course of this longitudinal study, which may have influenced our findings. Finally, this study did not include a comparison group of neurotypical youth, limiting the ability to assess the specificity of these effects. Future research should address this gap by examining whether these findings are unique to autistic children.
Conclusion
In summary, this work replicates prior work indicating elevated prevalence of social anhedonia in autistic children and the concurrent relationships between these symptoms and co-occurring psychopathology. Our findings also shed light on the trajectories of social anhedonia in autistic children, revealing a relatively stable trajectory over the course of 6 months. Finally, the results demonstrate a bidirectional relationship between social anhedonia and depression symptoms over time, suggesting that symptoms of social anhedonia may help identify autistic children at risk for the development of depression. Future research should integrate objective indicators of social reward processing (e.g. neural or visual attentional biomarkers) with questionnaire and parent-report measures to improve understanding of social anhedonia in autistic youth. Overall, these results underscore the clinical significance of identifying autistic children elevated in social anhedonia and emphasize the need for a comprehensive assessment of these symptoms in autistic children.
Acknowledgements
The authors wish to thank all the participating families and staff for their time and effort on this project.
Funding statement
This research was supported by an NIMH grant to J.C.M. (U19 MH108206). In addition, the effort by A.H.G. was supported by an NIMH T32 fellowship (MH18268).
Competing interests
G.D. serves on the Scientific Advisory Boards of Akili Interactive, Nonverbal Learning Disability Project, and Tris Pharma, has received speaker fees from WebMD, and receives royalties from Apple, Guilford Press, and Springer Nature Press. A.R.L. has served as a consultant for Jaguar Gene Therapies and Lab 1636, LLC. J.C.M. consults with Customer Value Partners, Bridgebio, Determined Health, and BlackThorn Therapeutics, has received research funding from Janssen Research and Development, serves on the Scientific Advisory Boards of Pastorus and Modern Clinics, and receives royalties from Guilford Press, Lambert, Oxford, and Springer. All other authors have declared that they have no competing or potential conflicts of interest.
Open access
