Reward Deficits in Co‐Occurring Social Anxiety and Depression: The Roles of Anhedonia, Boredom, Loneliness, and COVID‐19 Social Restriction

ABSTRACT

Prior research points toward reward deficits as crucial to the link between social anxiety disorder and major depressive disorder. However, studies have not explored specific types of reward deficits or how social‐environmental factors may impact them. The current study examined whether the indirect effects of three reward deficits (anhedonia, boredom, and loneliness) accounted for the connection between social anxiety and depression. Questionnaires were collected in a sample of college students (N = 523) during the first year of the COVID‐19 pandemic, allowing examination of environmental stressors (i.e., remote learning and current friend network size) on reward deficits. Indirect effect models showed that anhedonia, boredom, and loneliness each uniquely partly accounted for the association between social anxiety and depression (accounting for 20%, 38%, and 14% of the association, respectively), with boredom accounting for the largest proportion. Further, analyses suggested that anhedonia was associated with the social anxiety‐depression link only among remote learners, while loneliness appeared to be only associated among students who had smaller friend networks. Findings highlight the importance of reward deficits in co‐occurring social anxiety‐depression symptoms. We present novel findings showing the strong role of boredom in the connection between social anxiety and depression symptoms, suggesting a new transdiagnostic treatment target. Social‐environmental stressors, such as remote learning and contact with fewer friends, may exacerbate the connection between reward deficits and social anxiety‐depression co‐occurrence and serve as areas for intervention.

Social anxiety disorder (SAD) is one of the most common mental health conditions in the United States, with a lifetime prevalence of 12.1% (Stein et al. 2017). SAD is highly comorbid with major depressive disorder (MDD; Koyuncu et al. 2019), a condition that also has a high lifetime prevalence of 23.9% (Tam et al. 2020). SAD typically develops earlier than MDD and is a significant risk factor for subsequent MDD (e.g., Beesdo et al. 2007; Belzer and Schneier 2004; Ohayon and Schatzberg 2010). Individuals with comorbid SAD/MDD experience more severe symptoms and functional impairment (e.g., Ledley et al. 2005; Norton et al. 2008; Stein et al. 2001) as well as a poorer response to treatment (Crawley et al. 2008; Ledley et al. 2005).

To explain SAD/MDD comorbidity, sensitivity shift theory (SST; Richey et al. 2019) focuses on the reward motivational deficits characteristic of both SAD and MDD. SST posits that temperamental, biological, and environmental factors converge to, across development, shift individuals with heightened social sensitivity toward a social defeat syndrome (Richey et al. 2019). Social defeat syndrome includes deficits in positive affect and social approach behaviors (i.e., pursuit of reward), which lead to a reduction of positively reinforcing social experiences. The outsized costs compared to gains of social interactions end in chronic social defeat and lay the foundation for social anxiety symptoms, while increased social withdrawal and loss of social enjoyment lead to depression. While theories of social anxiety often focus on the role of increased distress and high social avoidance (Heimberg et al. 2014; Clark and Wells 1995), SST highlights the loss of positive affect and low social approach. Thus, reward deficits, which may include anhedonia, boredom, and loneliness, are theoretically important but understudied among co‐occurring social anxiety and depression. We tested this in college students during a COVID‐19 pandemic partial shutdown, which provided naturalistic social stressors (e.g., isolation, online learning) with impacts on reward motivational factors (Gigli et al. 2024; Wieman et al. 2024).

Anhedonia—low pursuit of or pleasure in typically enjoyable activities—is not included in most theories of SAD. However, research on low positive affect, one of the key features of anhedonia (Craske et al. 2016), suggests that anhedonia is more closely related to social anxiety than other forms of anxiety (Brown et al. 1998; Kashdan 2004; Kashdan 2007). In MDD, anhedonia is a core clinical symptom that occurs in most people with the disorder (Cao et al. 2019; Franken et al. 2007). As such, anhedonia may help explain the relation between SAD and MDD. Consistent with predictions of SST, network analyses have identified anhedonia (de la Torre‐Luque and Essau 2019; Heeren et al. 2018) and low positive affect (Langer et al. 2019) as important bridge symptoms between SAD and MDD. Anhedonia prospectively mediates the relation between fear of positive evaluation (an aspect of social anxiety) and subsequent depression (Jordan et al. 2018) and between anxiety and subsequent depression (Winer et al. 2017). In a cross‐sectional study, anhedonia partially contributed to a serial indirect effect between social anxiety and depression (Hermann et al. 2023). These findings highlight the potential importance of anhedonia in explaining the co‐occurrence between social anxiety and depression.

Boredom is another possible reward‐related factor in SAD/MDD comorbidity. SST does not discuss boredom, but boredom's overlap with anhedonia suggests it may drive some of the effects of anhedonia. Indeed, prior work has shown that boredom is related to, but distinct from, anhedonia in that boredom involves a lack of interest in current activities and a desire for more engaging activities (Barbalet 1999; Goldberg et al. 2011). In comparison to the general affective state of anhedonia, boredom arises from situational (i.e., lack of external stimulation) and/or cognitive factors (i.e., inattention, low cognitive resources; Goldberg et al. 2011). Boredom has moderate‐to‐large correlations with depression and anxiety (e.g., Fahlman et al. 2013; Hager et al. 2022; Krotava and Todman 2014; Yao et al. 2023; Zarei 2020), and it predicts depression longitudinally (Spaeth et al. 2015). Boredom has rarely been studied in relation to social anxiety, but two studies reported moderate‐to‐large correlations (Xiao et al. 2021; Yao et al. 2023).

Loneliness is another potential reward deficit involved in social anxiety‐depression comorbidity. SST describes how chronic social stress and related social anxiety lead to deficits in social reward contingencies, and loneliness may emerge as an outcome of this process. Indeed, loneliness is generally defined as a discrepancy between one's perceived and desired level of social connection (Zavaleta and Samuel 2014), resulting in reduced social reward reinforcement (Young 1982). Several core features of social anxiety may facilitate this discrepancy: social avoidance, maladaptive coping behaviors (e.g., avoiding eye contact), and the tendency to compare one's perception of an interaction to the ideal interaction (Heimberg et al. 2014; Jones et al. 1990). Consistent with this, social anxiety is strongly and positively associated with loneliness in clinical and nonclinical samples (e.g., Inderbitzen‐Pisaruk et al. 1992; O'Day et al. 2019; Rozen and Aderka 2023). Loneliness is also a well‐established correlate of depression (e.g., Erzen and Çikrikci 2018; Hager et al. 2022; Ren et al. 2022) and predicts future depression (e.g., Cacioppo et al. 2006; Richardson et al. 2017; Wei et al. 2005), potentially through reward motivational deficits stemming from beliefs that social prospects cannot be improved (Vanhalst et al. 2012). Cross‐sectional studies suggest that loneliness plays a significant role in depression's link to social anxiety (Baytemir and Yildiz 2017) and general anxiety (Ebesutani et al. 2015; Palgi et al. 2020). Longitudinal evidence suggests complex associations: one adult study found that social anxiety predicted future loneliness, which predicted future depression and social anxiety (Lim et al. 2016), whereas adolescent samples show variability in the temporal ordering of these effects (Danneel et al. 2019) and that loneliness, social anxiety, and depression develop somewhat independently (Danneel et al. 2020). Together, this literature positions loneliness as a critical correlate—and sometimes intermediary—in the social anxiety‐depression relationship, while highlighting the need to examine these complex associations in context of other potentially influential factors.

In the current study, we compared the indirect effects of three reward‐related factors that are plausibly related to the co‐occurrence of social anxiety and depression: anhedonia, boredom, and loneliness. We grounded our hypotheses in SST and examined these associations cross‐sectionally. As such, this study was well‐suited to identify variables that are 1) based on the theoretical etiology of comorbid SAD/MDD and 2) relevant for the targeted treatment comorbid social anxiety‐depression. We selected anhedonia, boredom, and loneliness as possible explanatory factors due to their relevance to reward and their independent associations with social anxiety and depression. We hypothesized that these constructs would uniquely partially account for the association between social anxiety and depression in a sample of college students. We also compared the strength of the indirect effect of each variable.

Analyses further examined whether reward deficits associated with social anxiety‐depression co‐occurrence varied by environmental factors introduced during COVID‐19 pandemic social restrictions. Social environments greatly impact reward processing (Aoki et al. 2021; Harkness et al. 2021; Hill et al. 2017), and many pandemic‐related changes may have intensified associations between reward deficits and mental health. The pandemic introduced a broad range of stressors—including uncertainty, financial loss, physical inactivity, illness, and prolonged isolation—that have been linked to emotional distress (Ebrahimi et al. 2021; Ganesan et al. 2021). In line with SST's focus on social adversity and given our sample population, we examined two social‐contextual pandemic factors particularly relevant to college students: social network size and online learning. The impact of similar environmental factors was highlighted in longitudinal work showing that pandemic social restrictions amplified loneliness and depression but reduced social anxiety (Lim et al. 2022). Further, the research literature on the social effects of online learning is particularly scarce (Cramarenco et al. 2023). We hypothesized that associations through anhedonia, boredom, and loneliness would be stronger among students who were enrolled exclusively in online courses and who reported smaller friend social networks.

1. Methods

1.1. Participants

Participants were 523 undergraduate students (mean age = 20.49 years, SD = 4.59) at a large southeastern university recruited through the psychology department research pool. A majority of participants were cisgender women (75.5%) and identified as either White (41.3%) or Black (40.9%; see Table 1 for full demographics). Of the 611 total participants that completed the survey, 80 (13.1%) were excluded for incorrect responses to at least one of four attention check items, and eight participants had missing data on a portion (n = 4) or all (n = 4) of at least one questionnaire. Listwise deletion was used for these 8 participants based on: 1) the small percentage of participants with missing data (1.5%), 2) the small percentage of missing values (0.4%), 3) a non‐significant Little's MCAR test (χ ²(32) = 32.31, p = 0.86), and 4) t‐tests not showing a difference in questionnaire scores as a function of missingness on other questionnaires (ps > 0.45). This yielded the final sample of 523 participants.

Table 1.

Sample characteristics.

	Mean (SD)	Median	Frequency (%)
Age (range = 18–58 years)	20.49 (4.59)	19
Gender
Cisgender man			98 (18.7%)
Cisgender woman			395 (75.5%)
Transgender man			4 (0.8%)
Transgender woman			1 (0.2%)
Non‐binary			3 (0.6%)
Genderfluid			1 (0.2%)
Other			11 (2.1%)
No label			7 (1.3%)
Prefer not to answer			3 (0.6%)
Ethnicity
Latinx			51 (9.8%)
Not latinx			472 (90.2%)
Race
White or Caucasian			216 (41.3%)
Black or African American			214 (40.9%)
Asian or Asian American			25 (4.8%)
Arab, Middle Eastern, or Arab			4 (0.8%)
American
Pacific Islander			3 (0.6%)
American Indian, Native			1 (0.2%)
American, or Alaskan Native
Multiracial			52 (9.9%)
Other			8 (1.5%)

The clinical cutoffs for possible MDD on the Patient Health Questionnaire‐9 (≥ 10) and SAD on the Mini‐Social Phobia Inventory (≥ 6) were met by 38.4% and 50.1% of the sample, respectively. The cutoff score on the Lubben Social Isolation Scale – 6 (< 6) indicated social isolation from friends in 22.4% of the sample and from family in 15.5% of the sample. Data were collected between November 1st and December 9th, 2020. During this time, a spike in Covid‐19 cases and deaths had led to statewide mandates restricting indoor and outdoor gatherings, expanding mask mandates, and prohibiting the sale of alcohol after 10:00 pm in order to limit time in communal indoor spaces. University courses were delivered through a mix of in‐person, online, and hybrid formats at this time. See Table 2 for study variable means and frequencies.

Table 2.

Emotion and environmental variables by group.

	Full sample (n = 523)		In‐person (n = 241)		Online only (n = 282)
	Mean (SD)	Frequency (%)	Mean (SD)	Frequency (%)	Mean (SD)	Frequency (%)	Mann– Whitney U p
Emotion
Social Anxiety	5.83 (3.48)		5.62 (3.31)		6.01 (3.62)		0.257
Depression	8.74 (6.57)		8.24 (6.06)		9.17 (6.95)		0.232
Anhedonia	8.11 (8.08)		7.25 (7.15)		8.84 (8.75)		0.108
Boredom	45.16 (12.44)		45.39 (12.06)		44.96 (12.77)		0.701
Loneliness	22.23 (6.61)		21.76 (6.04)		22.63 (6.77)		0.195
Environment
At least one of the following course types:
In‐person		130 (25.0%)		130 (53.9%)		0 (0.0%)
Online—live course delivery		380 (72.7%)		179 (74.3%)		201 (71.3%)
Online—recorded/asynchronous delivery		385 (73.6%)		164 (68.0%)		221 (78.4%)
Hybrid		161 (30.8%)		161 (66.8%)		0 (0.0%)
Current housing:
University housing or residence hall		132 (25.2%)		98 (40.7%)		34 (12.1%)
Off campus housing		136 (26.0%)		43 (17.8%)		93 (33.0%)
Parent or guardian's residence		243 (46.5%)		96 (39.8%)		147 (52.1%)
Other		12 (2.3%)		4 (1.7%)		8 (2.8%)
Friend Network Size	8.12 (3.51)		8.55 (3.45)		7.76 (3.52)		0.016*
Family Network Size	8.35 (2.84)		8.41 (2.90)		8.30 (2.81)		0.627
Extent displeased by activities being cancelleda	4.79 (1.66)		5.00 (0.10)		4.60 (0.10)		0.018*
Time interacting with friends in‐personb	1.81 (1.41)		2.04 (1.38)		1.61 (1.41)		< 0.001**
Time interacting with friends remotelyb	2.79 (1.21)		2.86 (1.20)		2.72 (1.21)		0.177
Time interacting with relatives in‐personb	2.06 (1.32)		1.94 (1.34)		2.16 (1.29)		0.059
Time interacting with relatives remotelyb	2.31 (1.21)		2.44 (1.22)		2.19 (1.20)		0.013*

Note: Depression = Patient Health Questionnaire‐9. Social anxiety = Mini‐Social Phobia Inventory. Anhedonia = Specific Loss of Interest and Pleasure Scale. Boredom = Multidimensional State Boredom Scale—Disengagement subscale. Loneliness = UCLA Loneliness Scale‐10. Friend Network Size = Lubben Social Network Scale‐6—Friends. Lubben Social Network Scale‐6—Family.

^a (1‐extremely pleased to 7‐extremely displeased).

^b (0‐none to 4‐a great deal).

^*Significant at p < 0.05.

^**Significant at p < 0.001.

1.2. Procedure

All participants provided informed consent prior to the start of the study. Participants completed all measures online and received research course credit upon completion. Procedures were approved by the university's Institutional Review Board.

1.3. Measures

1.3.1. Depression

Depression symptom severity was assessed with the 9‐item Patient Health Questionnaire‐9 (PHQ‐9; Spitzer et al. 1999). Participants indicated how often they were bothered by symptoms of depression in the past 2 weeks on a 4‐point Likert scale from 0 (not at all) to 3 (nearly every day). Items were summed to yield a total score in which higher scores indicated greater symptom severity. Internal consistency was excellent in the current sample (Cronbach's α = 0.91).

1.3.2. Social Anxiety

The severity of social anxiety symptoms was evaluated using the 3‐item Mini‐Social Phobia Inventory (Mini‐SPIN; Connor et al. 2001). Participants rated how much they are impacted by fear and embarrassment over the past week on a 5‐point Likert scale from 0 (not at all) to 4 (extremely). The total score was calculated by summing the items with higher scores indicating greater symptoms of social anxiety. Internal consistency was good in the present study (α = 0.82).

1.3.3. Anhedonia

The 23‐item Specific Loss of Interest and Pleasure Scale (SLIPS; Winer et al. 2014) assessed recent changes in anhedonia. Participants rated symptoms on a 0–3 scale with increasing numbers indicating greater anhedonia. Responses of 3 indicated trait level anhedonia (e.g., ‘I have never enjoyed going out with anyone’) and were recoded to 0 to focus on recent changes in anhedonia rather than trait anhedonia, as recommended in prior studies (Winer et al. 2014, 2017). Higher total summed scores indicated a greater recent increase in anhedonia (Winer et al. 2017). Internal consistency was excellent in the present study (α = 0.91).

1.3.4. Boredom

Boredom was evaluated using the 10‐item Disengagement subscale of the Multidimensional State Boredom Scale (MSBS‐D; Fahlman et al. 2013). The Disengagement subscale is the “core component of boredom” and represents the unique aspects of boredom that do not overlap with anxiety and depression (p. 82; Fahlman et al. 2013). Compared to other subscales, MSBS‐D also has the highest factor loading (λ = 0.95) on the MSBS boredom latent variable and the highest correlation with boredom proneness (r = 0.61; Fahlman et al. 2013). Participants rated their feelings about their lives over the past 2 weeks on a 7‐point Likert scale from 1 (strongly disagree) to 7 (strongly agree) such that a higher total score indicated greater boredom. Internal consistency was good in the present study (α = 0.89).

1.3.5. Loneliness

The 10‐item UCLA Loneliness Scale (ULS‐10; Russell 1996) assessed loneliness by asking participants to report how often they felt isolated on a 4‐point Likert scale from 1 (never) to 4 (always). After five items were reverse coded, all items were summed to get a total score with higher scores indicating greater feelings of loneliness. Internal consistency was good in the present study (α = 0.89).

1.3.5.1. Course Instruction Type

Participants provided the number of university courses they were currently taking in the following categories: 1) Entirely in‐person, 2) Entirely online with live course delivery, 3) Entirely online with recorded delivery, and 4) Hybrid (classes that are partly in‐person and partly online). Participants taking any courses in categories 1 and 4 were classified into the in‐person group (INP, n = 241), and participants taking only courses in categories 2 and 3 were classified into the online only group (ONL, n = 282). Note that those in the INP group could also be taking one or more online class. See Table 2 for group means on social variables.

1.3.5.2. Social Network Size

The 6‐item short form of the Lubben Social Network Scale (LSNS‐6; Lubben et al. 2006) assessed the size of participant's social network. The scale included two 3‐item subscales (Family Network and Friend Network), with the Friend Network subscale being used in the current analyses. The three items requested (in an open response field) the number of friends participants 1) “see or hear from at least once a month,” 2) “feel at ease with that you can talk about private matters,” and 3) “feel close to such that you could call on them for help.” As such, these items asked about friends in general as well as close friendships. The number of friends was recoded to align with the 5‐point Likert scale used by the LSNS‐6: 0 = none, 1 = one, 2 = two, 3 = three or four, 4 = five to eight, and 5 = nine or more friends. The scores for each item were summed, such that higher scores indicated larger social networks.

1.4. Analytic Strategy

PROCESS (version 4.1; Hayes 2022) for R (version 4.1.2; R Core Team 2021) was used to estimate all the regression coefficients and indirect effects with percentile bootstrap confidence intervals (CIs) and 10,000 bootstrap samples. Percentile bootstrapping is recommended over other bootstrapping methods (e.g., bias‐corrected) in mediation analyses due to lower Type I error (Hayes and Scharkow 2013; Tibbe and Montoya 2022). Effects were interpreted as significant when the 95% CI did not include zero. The indirect effects model tested the parallel (i.e., unique) indirect effects of social anxiety on depression through anhedonia, boredom, and loneliness (PROCESS model 4). As an alternative model to the indirect effects model, we examined moderation models (see Material S1). To test the relative size of the specific indirect effects, we followed up with exploratory tests of contrasts between the three indirect effects, again using percentile bootstrapping. Moderated indirect effect models included the moderator at all three paths in the models (PROCESS model 59). Influential cases (i.e., potential multivariate outliers) for each path were identified by the spread and cutoff values of studentized deleted residuals (SDR), standardized difference in fit (DFFITS), and standardized difference in betas (DFBETAS; Cousineau and Chartier 2010). To identify sensitivity to influential cases, any difference in analyses without influential cases were reported in footnotes and generally showed little impact on model interpretation. The HC3 estimator of standard errors was used due to violations of the homoscedasticity assumption as indicated by visual inspection of residual plots (Davidson and MacKinnon 1993). Pearson correlations between all study variables are in Table 3. We considered the primary analyses to be the specific indirect effects in the parallel indirect effect model as well as tests of whether these indirect effects varied across the two moderators in the moderated indirect effect models. To evaluate a conservative robustness check of these primary analyses, we also reported 99% bootstrap confidence intervals (the most conservative available), with full results provided in the Supplemental Materials. Pairwise comparisons among indirect effects were examined as exploratory.

Table 3.

Correlations among variables.

	Depression	Social anxiety	Anhedonia	Boredom
Depression	–
Social Anxiety	0.39*	–
Anhedonia	0.54*	0.31*	–
Boredom	0.60*	0.42*	0.47*	–
Loneliness	0.52*	0.35*	0.58*	0.53*

Note: Sample size is N = 523. M = mean. SD = standard deviation. Depression = Patient Health Questionnaire‐9. Social anxiety = Mini‐Social Phobia Inventory. Anhedonia = Specific Loss of Interest and Pleasure Scale. Boredom = Multidimensional State Boredom Scale—Disengagement subscale. Loneliness = UCLA Loneliness Scale‐10

^* p < 0.001.

2. Results

2.1. Indirect Effects Model

The results described below show unstandardized betas (B) and Figure 1 displays the standardized betas (β). The total effect of social anxiety on depression was significant, B = 0.73, 95% CI: [0.58, 0.88], R ² = 0.15. Social anxiety was significantly associated with anhedonia, B = 0.72, 95% CI: [0.52, 0.93], boredom, B = 1.50, 95% CI: [1.20, 1.80], and loneliness, B = 0.66, 95% CI: [0.50, 0.81]. Depression was significantly associated with anhedonia, B = 0.21, 95% CI: [0.13, 0.28], boredom, B = 0.18, 95% CI: [0.13, 0.23], and loneliness, B = 0.15, 95% CI: [0.06, 0.24] (Figure 1).

Figure 1.

Indirect Effects of Social Anxiety on Depression. Note. Coefficients are standardized betas. All paths and effects are significant at p < 0.01. The coefficient from Social Anxiety to Depression is the total effect with the direct effect in parentheses.

The direct effect of social anxiety on depression was significant, B = 0.21, 95% CI: [0.05, 0.37], and there were significant indirect effects through anhedonia, B = 0.15, 95% CI: [0.08, 0.22], boredom, B = 0.28, 95% CI: [0.19, 0.37], and loneliness, B = 0.10, 95% CI: [0.04, 0.16] (Figure 1). In terms of proportional effect size (indirect effect/total effect), anhedonia, boredom, and loneliness respectively explained 20% (0.15/0.73), 38% (0.28/0.73) and 14% (0.10/0.73) of the relation between social anxiety and depression. The indirect effect of boredom was significantly larger than the indirect effects of anhedonia, B _difference = 0.13, 95% CI [0.01, 0.24], and loneliness, B _difference = 0.18, 95% CI: [0.06, 0.30], but the indirect effects of anhedonia and loneliness were not significantly different from each other, B _difference = 0.05, 95% CI: [−0.05, 0.15]¹. Effects survived correction for multiple comparison (see Material S1).

2.2. Moderated Indirect Effects Models

2.2.1. Course Instruction Type: Online Only Versus In‐Person

The direct effect of social anxiety on depression was only marginally significant, B = 0.18, 95% CI: [−0.03, 0.39]. The indirect effect through anhedonia was significantly larger for online only classes, B _ONL = 0.22, 95% CI: [0.11, 0.35], than for in‐person classes, B _INP = 0.06, 95% CI: [0.002, 0.13], B _difference = 0.16, 95% CI: [0.03, 0.30]. Indeed, the indirect effect of anhedonia was 3.6 times greater in the ONL group and explained 29% (0.22/0.75 [indirect effect/total effect]) of the association between social anxiety and depression in this group (compared to 8% [0.06/0.75] in the INP group). Using the corrected 99% CI made the group difference non‐significant but showed that the indirect effect of anhedonia was only in online‐only students, not in‐person students (see Material S1). There were no group differences in the indirect effects through boredom, B _ONL = 0.25, 95% CI: [0.14, 0.38], B _INP = 0.33, 95% CI: [0.18, 0.49], B _difference = −0.07, 95% CI: [−0.27, 0.11], or loneliness, B _ONL = 0.10, 95% CI: [0.02, 0.20], B _INP = 0.09, 95% CI: [0.01, 0.19], B _difference = 0.01, 95% CI: [−0.12, 0.13]².

2.2.2. Friend Social Network Size

The conditional (i.e., moderated) indirect effects described here contrast the effects at ±1 SD from the mean of the moderator (i.e., friend social network size). The direct effect of social anxiety on depression was not significant (B = 0.16, 95% CI: [−0.24, 0.57]). The indirect effect through anhedonia did not significantly depend on friend network size but there was a non‐significant trend showing a greater role of anhedonia for people who reported a larger compared to smaller friend network size (B _difference = 0.08, 95% CI: [−0.05, 0.22]). The indirect effect through boredom did not depend on friend network size (B _difference = −0.05, 95% CI: [−0.22, 0.12]). The indirect effect of loneliness was significantly moderated by friend network size such that the effect was greater and significant when friend network size was smaller (B = 0.14, 95% CI: [0.06, 0.24]) but not significant when friend network size was larger (B = 0.03, 95% CI: [−0.05, 0.10]; B _difference = 0.12, 95% CI: [0.0003, 0.24])³. This difference was not significant when using the corrected 99% CI, but the indirect effect in smaller friend network size remained significant and the effect in larger friend network size remained non‐significant (see Material S1).

3. Discussion

The high rate and severity of comorbid social anxiety and depression highlight the importance of identifying potential targets for intervention and prevention. The current study—informed by SST (Richey et al. 2019)—compared reward‐related factors to help understand their contribution to social anxiety‐depression co‐occurrence in a sample of college students. Data were collected during the first full semester of the COVID‐19 pandemic, which created a natural stressor for social access and reward. Results showed that anhedonia, boredom, and loneliness each accounted for unique variance in the relation between social anxiety and depression, consistent with cross‐sectional indirect effects. Anhedonia, boredom, and loneliness represent different aspects of reward (i.e., poor reward processing, lack of exposure to stimulation, and low social reinforcement, respectively), which facilitated a nuanced examination of reward dysfunction in co‐occurring social anxiety‐depression. The indirect effect of boredom was particularly strong, while the strength of anhedonia and loneliness depended on environmental factors.

SST posits that social anhedonia may serve as a key mechanism underlying the transition from SAD to MDD (Richey et al. 2019). Our cross‐sectional data cannot directly speak to this transition, but our findings suggest that anhedonia is a potential connection between co‐occurring social anxiety and depression. SST characterizes social anhedonia, in part, as low social motivation (Richey et al. 2019), which can help explain why anhedonia is a link between social anxiety and depression. People high in social anxiety report lower social approach motivation (Blay et al. 2021), exhibit fewer approach behaviors (e.g., social initiation; Alden and Taylor 2010), and expect social failure (e.g., low social rank; Trower and Gilbert 1989). In line with behavioral research and theories of depression (e.g., Carvalho et al. 2011; Lewinsohn and Graf 1973), SAD may increase depression because low social approach behaviors reduce the probability of pursuing and experiencing reward. Further, perception of social failures may be heightened by the relatively low levels of positive affect (i.e., greater anhedonia) that socially anxious individuals feel during social experiences (Alden and Taylor 2010; Goodman et al. 2021). When these individuals experience low reward in the face of their ideal social standards, they experience self‐criticism and perceptions of social failure (Alden and Taylor 2010). Such negative self‐directed thoughts are common triggers of depression and likely factors in social anxiety‐depression comorbidity (Ingram et al. 2001). Complementing low social reward and reduced approach behaviors, individuals with SAD experience extreme social distress and strive to avoid it by not initiating social interactions or disengaging from them via safety behaviors (e.g., not speaking; Heimberg et al. 2010; Piccirillo et al. 2016). Regulating negative emotions through avoidance can markedly reduce opportunities for positive events, positive emotions, and goal attainment (Kashdan and Weeks 2010), such that avoidance likely works synergistically with anhedonia to further promote depression. Although SST singles out social anhedonia, a meta‐analysis showed that social and physical anhedonia are similarly elevated in MDD (Gandhi et al. 2022). Our data suggest that general anhedonia can partially account for variance in the social anxiety‐depression association, but future research may explore anhedonia subdomains.

Anhedonia accounted for more of the association between social anxiety and depression in online‐only students compared to in‐person students. Indeed, the indirect effect of anhedonia was only significant in online‐only students and not significant in students taking in‐person courses. Although the group difference did not survive multiple comparison correction, the pattern points to a potentially meaningful effect that appears specific to online‐only students. This finding aligns with SST's prediction that, among socially anxious individuals, infrequent social interactions will reduce social motivation and reward sensitivity (Richey et al. 2019). Students who take classes in isolation from peers have fewer opportunities for social interactions and other engaging activities typical of college life. Isolation from peers may allow socially anxious individuals to more easily avoid social activities, especially for anxious individuals, who have difficulty choosing rewards that conflict with threat‐avoidance (Pittig et al. 2021). This social avoidance may be associated with depression due to the reduced likelihood of opportunities for reward reinforcement (Carvalho et al. 2011). In contrast, when someone has a rewarding experience, they are more likely to feel a reduction in fear over time (i.e., fear extinction learning; Papalini et al. 2019; Pittig 2019). As such, opportunities for more varied and rewarding experiences through in‐person and hybrid classes may have improved social fear extinction. This represents a feedback loop wherein reduced avoidance increases reward and learning, which further reduces avoidance and increases opportunity for reward and learning (Taylor et al. 2022). Thus, it is possible that taking only online classes strengthened the risk for depression in people with social anxiety through decreased reward approach behaviors and increased avoidance. Given the unique pandemic context of our data, we consider these findings preliminary, and research should examine how post‐pandemic online learning may impact anhedonia and social anxiety‐depression comorbidity.

The indirect effect of boredom on the social anxiety‐depression association was significantly larger than that of anhedonia and loneliness, suggesting that boredom may be particularly important in accounting for these co‐occurring symptomatologies. Consistent with the few previous studies examining the relationship between social anxiety and boredom (Xiao et al. 2021; Yao et al. 2023), our findings demonstrated a strong association between them and extend the literature by showing that this relationship persists even after accounting for other reward‐related factors. This is the first study to suggest and test the role of boredom in accounting for the link between social anxiety and depression, which expands upon the social motivational deficits outlined in the SST. The indirect effect of boredom suggests participants had desire for engagement that was limited by cognitive and environmental barriers. One such barrier may be attentional control deficits, which have been associated with boredom (Perone et al. 2021; Yakobi et al. 2021), lower positive affect (Morrison and Heimberg 2013), and social anxiety and depression (Derakhshan 2020; Judah et al. 2013). Another cognitive variable, repetitive negative thinking, may also help account for the association between boredom and depression, as found previously (Hager et al. 2022). Indeed, boredom may increase depression in people with social anxiety by triggering the tendency they already have toward repetitive negative thinking (Wong et al. 2019).

As with anhedonia, social withdrawal in people with social anxiety may have contributed to boredom and its link to depression. Although the two aspects of social access that we examined (friend network size and in‐person classes) did not impact the indirect effect of boredom, boredom‐related withdrawal is common in non‐social domains (e.g., occupational; Bruursema et al. 2011). Future research on types of boredom (e.g., ‘indifferent boredom,’ ‘searching boredom’; Goetz et al. 2014) may provide insight into the circumstances under which boredom strengthens the ties between social anxiety and depression.

Finally, we found that loneliness also contributed to the co‐occurrence of social anxiety and depression, building upon the substantial literatures showing greater loneliness in people with social anxiety (Inderbitzen‐Pisaruk et al. 1992; O'Day et al. 2019) or depression (e.g., Cacioppo et al. 2006; Ren et al. 2022). Loneliness has been shown to increase physiological and mental stress (Nausheen et al. 2007; Nowland et al. 2018; Zawadzki and Gavrilova 2021) and decrease neural reward responsiveness (Cacioppo et al. 2009). Thus, loneliness may contribute to the chronic social stress and reward dysfunction that SST implicates in social deficits and depression (Richey et al. 2019). Loneliness may also increase risk for depression by contributing to low self‐efficacy (Liu et al. 2024), which is a cognitive risk factor of depression (Maddux and Meier 1995; Muris 2002; Steca et al. 2014). Previous work has also found that loneliness is associated with low self‐efficacy in people with social anxiety (Kashdan and Roberts 2004). Depression may be more likely among socially anxious individuals whose loneliness triggers their perceived lack of efficacy in social skills.

Our analysis of loneliness further suggested that its explanation of the social anxiety‐depression link was only among people with smaller friend networks, whereas this association was not significant for those with larger networks. After removing 27 influential cases, although the standardized indirect effect remained roughly twice as large when network‐group was smaller versus larger, the overall moderation became non‐significant. This sensitivity indicates the moderation should be interpreted cautiously and is considered preliminary. Our measure of friend network size (Friend Network subscale of the LSNS‐6) incorporated information on number of friends and number of close friends. This close friend distinction is important given that the negative association between loneliness and social network size is only consistent when considering close relationships (e.g., Reinhard et al. 2022; Rico‐Uribe et al. 2016; Rigon et al. 2019; Stokes 1985). Individuals with small networks of close friends likely have fewer social resources to mitigate the impact of loneliness on strengthening depression in people with social anxiety. This may be especially true during times of social stress, as research has shown that loneliness was more closely related to having fewer close relationships during pandemic isolation compared to a pre‐pandemic period (Kovacs et al. 2021). Given the modest effect seen here and the limitations of capturing social network with a single short questionnaire (Gupte and Eliassi‐Rad 2012; Roberts and Dunbar 2011), future research should more fully explore whether and how a larger friend network may disrupt the role of loneliness in social anxiety‐depression co‐occurrence.

4. Clinical Implications and Future Directions

Our results have several potential clinical implications for reducing social anxiety‐depression co‐occurrence and refining existing interventions. First, boredom emerged as the strongest indirect link between social anxiety and depression, suggesting that boredom may be a distinct and clinically relevant aspect of reward deficits associated with these symptoms that is not fully captured by anhedonia alone. Future research is needed to examine whether clinical assessment of boredom offers incremental utility beyond measuring anhedonia alone. This distinction could be important if future research demonstrates that boredom is more responsive to a tailored administration of behavioral activation (BA), a well‐established intervention that uses activity engagement to overcome loss of reward. Indeed, modified BA has been shown to ease boredom in specific populations (Magidson et al. 2020; Taube‐Schiff et al. 2016), but its adaptations for boredom in people with social anxiety or depression are yet to be explored. Consistent with our finding that boredom showed a unique association with social anxiety‐depression co‐occurrence, boredom may warrant further investigation as a distinct treatment target. Research is inconclusive regarding individual predictors of response to BA versus cognitive therapy (Driessen et al. 2022); however, future research may examine whether individuals with elevated boredom may benefit from special emphases, such as on stimulating and effortful BA. Further, prior research suggests that boredom may be associated with novelty seeking, including more easily shifting toward novel unhealthy patterns or increased risk‐taking behaviors (Bench and Lench 2019; Danckert and Elpidorou 2023), highlighting another potential avenue for future clinical research.

Second, the potential moderation by online learning suggests that intervention approaches may need to consider virtual contexts—a growing reality of education and mental health care. Virtual academic platforms that foster meaningful, diverse social interactions have been linked to improved academic motivation and success (Lotz et al. 2015; Weaver and Albion 2005) and may also be relevant to the anhedonia, social anxiety, and depression concerns highlighted in our moderation analyses. Our findings further suggest that the size of the one's close friend network modulates the association between loneliness and social anxiety‐depression co‐occurrence. This connects social anxiety‐depression research to prior literature showing that disconnected individuals may benefit from building close, supportive relationships rather than simply increasing the number of acquaintances (Fiorillo and Sabatini 2011; Schwartz and Litwin 2017). These findings suggest that social network characteristics may be relevant to understanding social anxiety and depression co‐occurrence, highlighting a potential avenue for future treatment and prevention research, particularly in virtual or socially restricted environments.

Third, prior research suggests that reward‐related processes may be relevant for social anxiety interventions, such as exposure therapy. Exposure therapy has traditionally emphasized conquering high anxiety scenarios, but low reward in everyday social encounters may also warrant consideration. Consistent with the present findings, loneliness may represent an important component of this low reward domain that warrants further investigation. Indeed, people with social anxiety report reduced intimacy in close relationships (see Ginat‐Frolich et al. 2024) and can have specific fears (e.g., perceived criticism, conflict, self‐disclosure) related to these important relationships that reduce enjoyment (Cuming and Rapee 2010; Langer et al. 2017; Porter et al. 2019). Future research may test the hypothesis that exposures designed to increase intimacy and challenge these fears may target loneliness and depressive symptoms. In exposure therapy, clinicians guide clients toward fully experiencing the anxiety of a scenario, but prior research has also suggested that interventions that encourage fully engaging in positive experiences may confer benefits as well. For example, research suggests that savoring interventions and comedic improvisational theater may help people indulge in positive social emotions while reducing anxiety (Chen et al. 2026; Phillips Sheesley et al. 2016). Researchers have also advocated for integrating BA with exposure therapy to address issues of motivation and anhedonia in people with social anxiety and depression (Judah et al. 2020). Indeed, increasing physical activity has been shown to improve loneliness and increase social interaction (Ahn et al. 2023; Shvedko et al. 2020). Overall, the reward deficits highlighted in the current study suggest a potentially fruitful direction for future intervention research examining whether approaches that emphasize positive emotion and reward may complement traditional efforts to reduce fear and avoidance in social anxiety treatment (Kashdan and Weeks 2010).

Fourth, mindfulness‐based interventions may represent another approach to address the reward‐related concerns highlighted by our findings. Such interventions aim to increase non‐judgmental awareness of the present moment and have been shown to reduce anhedonia (Abouzaid and Abdelhamid 2024; Cernasov et al. 2024), including social anhedonia (Carlton et al. 2021). Prior work has proposed mindfulness to help socially anxious individuals attend to the rewarding aspects of social interactions and see how their own actions contribute to these outcomes (Richey et al. 2019). Given our finding that boredom was the strongest indirect link between social anxiety and depression, it is particularly noteworthy that mindfulness has been shown to reduce boredom (Lee and Zelman 2019; Waterschoot et al. 2021), as well as improve loneliness (O'Day et al. 2021), perhaps by fostering objective observation and acceptance rather than rumination on negative thoughts. This prior research suggests that mindfulness‐based approaches may be relevant to boredom and loneliness by encouraging non‐judgmental awareness and acceptance of these experiences, potentially creating space for engagement in meaningful activities. This prior research suggests that mindfulness may enhance motivation and social reward sensitivity, with future research needed to examine these mechanisms in comorbid social anxiety and depression.

5. Limitations

This study should be interpreted within the context of its limitations. Because the study is cross‐sectional, claims regarding the directionality of the results cannot be made. As we tested multiple conceptually similar models, Type I error necessarily increased. Using corrected, 99% CIs supported the robustness of our indirect effects; the effects of environmental moderators no longer quite reached significance when looking at group differences, reinforcing these findings as preliminary (see Material S1). The directional associations modeled in the current study align with those in the SST framework and are supported by longitudinal research, but the temporal relationships between variables in this study are likely complex. Indeed, social anxiety may have a bidirectional association with loneliness (Lim et al. 2016) and low reward sensitivity (Potsch and Rief 2023), which is an aspect of anhedonia. Research has also demonstrated evidence for a bi‐directional association between depression and loneliness (Cacioppo et al. 2006; Ren et al. 2022), but the direction from loneliness to depression is more consistent in college students (Richardson et al. 2017; Vanhalst et al. 2012). Additionally, data were collected from college students relatively early in the COVID‐19 pandemic when enhanced safety measures were in effect. Although the social limitations during the pandemic were of interest in the current study, they likely impacted results and limited generalizability. In fact, other stressors seen during the pandemic—including recreational, financial, psychical and mental health—may modulate the links between social anxiety and depression, and future work should continue to unpack the role of environment on social anxiety‐depression co‐occurrence. Despite the limitations, the current study leveraged a unique opportunity to examine reward‐related factors in the context of natural social stressors, including remote education.

6. Conclusion

We examined reward‐related factors that may help account for variance in the co‐occurrence of social anxiety and depression symptoms in a cross‐sectional dataset. Anhedonia, boredom, and loneliness—representing different aspects of reward deficits—each accounted for a portion of the association between social anxiety and depression. Boredom had a stronger effect than anhedonia or loneliness, with evidence that the latter two had increased effects in the context of specific social stressors during the COVID‐19 pandemic (online classes and smaller friend networks). Results highlight the need for targeted interventions for reward deficits, especially to address the often‐overlooked experience of boredom. Such interventions could disrupt the association between social anxiety and depression and help individuals manage their interconnected symptoms.

Consent

All participants provided informed consent, and all procedures were approved by the Old Dominion University Institutional Review Board.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File

Data Availability Statement

The data that support the findings of this study are openly available in Reward Deficits in Comorbid Social Anxiety and Depression at https://doi.org/10.17605/OSF.IO/MNFE6.