The ability to experience pleasure is an important part of subjective well-being and mental health (Kringelbach and Berridge, 2017). Preclinical research on reward processing has dissociated several subcomponents, with partially distinct neural correlates, interacting in a normal hedonic response (Berridge and Robinson, 2003). Hedonic capacity, therefore, not only involves the ability to experience pleasure at the receipt of a reward, i.e., consummatory pleasure, but also comprises reward anticipation, reward learning and decisional components, weighing the estimated efforts of pursuit against the subjective value of a reward (Treadway and Zald, 2011; Der-Avakian and Markou, 2012; Kringelbach et al., 2012; Rømer Thomsen, 2015; Husain and Roiser, 2018). Although the exact conceptualization of these reward components may vary in the literature, the important insight with respect to clinical research is that an impairment of any of these components could potentially lead to a severe reduction or loss of pleasure in one's life. Research on anhedonia, the decrease or loss of hedonic capacity, needs to take this complexity into account.

Anhedonia is a transdiagnostic symptom occurring, as state or trait, in several psychiatric and neurological disorders, including major depressive disorder (MDD; Treadway and Zald, 2011; Cooper et al., 2018), schizophrenia (Gard et al., 2007; Lambert et al., 2018), substance use disorder (Garfield et al., 2014), post-traumatic stress disorder (Nawijn et al., 2015) and Parkinson's disease (Assogna et al., 2011). In addition, anhedonia can be found in the general population (Chan et al., 2012; Dodell-Feder and Germine, 2018; Martino et al., 2018) and has been recognized as a risk factor for several mental disorders (Kwapil, 1998; Gooding et al., 2005; Keedwell et al., 2012; Stringaris et al., 2015; Luby et al., 2018; Guffanti et al., 2019; Ward et al., 2019). Identifying the exact hedonic dysfunction underlying the symptom of anhedonia and potentially distinguishing anhedonic subtypes are important goals of clinical research. For this purpose, measures of anhedonia on different units of analysis, e.g., on neurobiological, behavioral and self-report levels, have to be integrated (Cuthbert and Insel, 2013). This necessitates to be mindful of the exact reward processing components the applied measurement instrument targets. Here, we focused on measures of self-report, specifically the validation of a recent scale designed to capture different facets of state anhedonia: the Dimensional Anhedonia Rating Scale (DARS; Rizvi et al., 2015). The DARS is one of several newly developed anhedonia scales aiming to integrate findings from the neurobiology of reward processing while providing culturally unbiased and contemporary item content (Rizvi et al., 2016).

As a trait, anhedonia has traditionally been assessed with the Revised Chapman Physical Anhedonia Scale (CPAS) and the Revised Chapman Social Anhedonia Scale (CSAS; Chapman et al., 1976, 1980; Bailer et al., 2004). Both were developed as part of a larger set of scales specifically tailored to examine psychosis proneness. Their items comprise different hedonic components, such as anticipation, interest, consummation and effort, but also trait aspects that are not directly related to anhedonia. Since their length (61/40 items) is an impediment for clinical use and the item content has been criticized as culturally biased, new trait measures of anhedonia have been developed (Gard et al., 2006; Gooding and Pflum, 2014; Rizvi et al., 2016). The Temporal Experience of Pleasure Scale (TEPS; Gard et al., 2006) measures predominantly physical pleasure on an anticipatory and a consummatory subscale, whereas the Anticipatory and Consummatory Interpersonal Pleasure Scale (ACIPS; Gooding and Pflum, 2014) focuses on the social domain. For the ACIPS, factor analysis did not distinguish between anticipation and consummation but revealed three interpersonal subdomains instead (Gooding and Pflum, 2014).

As a newly developed state anhedonia scale, the DARS is most importantly compared to the Snaith-Hamilton Pleasure Scale (SHAPS; Snaith et al., 1995) which also measures state anhedonia and has been applied widely, in particular with regard to state anhedonia in depression (Rizvi et al., 2016; Trøstheim et al., 2020). The SHAPS measures hedonic tone during the last few days with 14 hypothetically formulated items, e.g., “I would be able to enjoy my favorite meal.” The item content exclusively targets consummatory pleasure but covers four domains: interests/pastimes, social interaction, sensory experience and food/drink. To avoid cultural, gender and age biases and enable widespread use the items are kept simple and, in part, very general (Snaith et al., 1995). A recent meta-analysis (Trøstheim et al., 2020) based on SHAPS scores of 246 adult samples reported significantly higher, i.e., more anhedonic, scores in patients compared to healthy adults. Patients were diagnosed with current MDD, bipolar disorder, schizophrenia, substance use disorder, Parkinson's disease or chronic pain. MDD patients scored higher than any other patient group, thereby indicating that anhedonia in MDD affects multiple domains of pleasure (Trøstheim et al., 2020).

In contrast to the SHAPS, the DARS (Rizvi et al., 2015) was designed to capture not only consummatory pleasure, but also interest/desire, motivation and effort across the same four domains, thereby offering an important addition to anhedonia research. The DARS comprises 17 items assessing state anhedonia right now. It avoids cultural, gender and age biases as well as too specific item content by asking the participants to provide two or three of their own favorite examples for each of the four domains. Principal component analysis revealed four components distinguishing between the four domains pastimes/hobbies, food/drink, social activities and sensory experiences. Hence, domain-specific subscales have been proposed and high internal consistency has been reported for all scales (Rizvi et al., 2015; Arrua-Duarte et al., 2019). So far, the DARS has demonstrated adequate convergent and divergent validity in community and clinical samples regarding the SHAPS, measures of behavioral inhibition and activation and depression severity (Rizvi et al., 2015; Arrua-Duarte et al., 2019). Sullivan-Toole et al. (2019) found correlations between the DARS and reward-related and effort-related parameter estimates in a novel effort-based decision-making task. In addition, the DARS has shown predictive value above the SHAPS regarding reward functioning and differentiation of treatment-resistant MDD subgroups (Rizvi et al., 2015). A Spanish validation study corroborated the four-factor structure in a diverse clinical sample (Arrua-Duarte et al., 2019).

The primary aim of the present study was to validate the German version of the DARS in a convenience sample of 557 young, mentally healthy adults. This, to the best of our knowledge, is the first report on the factor structure and psychometric properties of the German DARS in its original 17-item version. We expected confirmatory factor analysis (CFA) to corroborate the four-factor structure mapping onto the four domains. Since the differentiation of anhedonic components is an important question in anhedonia research, we additionally evaluated CFA models distinguishing interest, motivation, effort and consummatory pleasure. Based on the literature, we predicted an inferior fit for these models. As measures of convergent validity, we expected moderate correlations to SHAPS, TEPS, ACIPS, CPAS, CSAS, trait positive affect and behavioral activation sensitivity. We expected either no or only very low correlations to the diverging constructs of behavioral inhibition sensitivity and trait negative affect. Since only few items in depression scales specifically target anhedonia, we hypothesized to find low to moderate correlations to depression severity.

As a secondary aim, we assessed the impact of the coronavirus disease 2019 (COVID-19) pandemic on state anhedonia and depression severity. This was not part of the original aim of our study. However, since our data was collected from August 2019 to July 2020, a time during which the COVID-19 pandemic emerged and spread around the world, we added this secondary aim to our analysis and divided our sample into a pre-pandemic and a during-pandemic group. In line with many reports of an increase in depression rates and a deterioration of mental health (Henssler et al., 2020; O'Connor et al., 2020; Pappa et al., 2020; Bueno-Notivol et al., 2021; Cénat et al., 2021), we assumed to find higher depression scores in the during-pandemic group. Further, we examined differences in state anhedonia between both groups, as there is little research on the effects of the COVID-19 pandemic on state hedonic capacity yet (Moccia et al., 2021; Prati and Mancini, 2021). Since anhedonia is a core symptom of depression, it is well-possible that anhedonia levels increase along with depression scores (Moccia et al., 2021). There are reports on fewer positive experiences during lockdown (Klaiber et al., 2020). However, a recent meta-analysis showed no significant effect on positive psychological functioning (Prati and Mancini, 2021). Losing the opportunity to experience many of the usual leisure activities certainly is a consequence of COVID-19 mitigation strategies such as lockdowns (Mutz and Gerke, 2020). But losing the opportunity to do something you usually enjoy does not necessarily imply an impairment of the capacity to enjoy. Given the lack of research on the topic of anhedonia specifically during the COVID-19 pandemic, we conducted two-tailed tests to examine this question.

Adding this secondary aim to our analysis also contributed to the DARS validation, the focus of our study. Assessing state anhedonia with both SHAPS and DARS allowed us to examine potential differences between these measures which might result from their different set-ups. Moreover, dividing our sample into two groups enabled us to perform multigroup CFA on the DARS. We examined whether the COVID-19 pandemic had influenced the factor model of the DARS by testing measurement invariance across both groups and performed latent mean comparisons on all factors of the DARS. This, to the best of our knowledge, is the first report on measurement invariance of the DARS across severe changes to the social environment.

Overall, our analyses resulted in good psychometric properties for the 17-item German DARS. In a sample of 557 young, mentally healthy adults, we confirmed the differentiation of four factors mapping onto the hedonic domains. All six fit indices indicated acceptable or good fit for the correlated four-factor domain model and each item loaded substantially on the assigned factor. Since the DARS measures the same constructs, namely interest/desire, motivation, effort and consummatory pleasure, in different domains, it is reasonable to assume a general factor in addition to the specific domain factors. When a multiple correlated factor model fits the data well, it is not surprising to find a similarly well-fitting bifactor model since the correlation between the factors already indicates shared variance. But computing a bifactor model and the corresponding ω coefficients has specific advantages in assessing multidimensional scales. Bifactor models and omega coefficients can estimate how much variance is explained by each factor (Revelle and Condon, 2019). This is especially relevant when deciding on the use of specific subscales. If the items of a subscale load much higher on the general factor than on their specific group factor, it would be unwarranted to assume that the subscale score reflects a specific subconstruct or domain rather than the overall factor (Rodriguez et al., 2016). In addition, bifactor models may be useful in structural equation modeling. It is easy to integrate the general factor into a structural model and also possible to evaluate whether a group factor can predict an external variable over and above the general factor (Chen et al., 2006). Which model will be most appropriate, e.g., in future studies, therefore will depend on the purpose of the specific study. In our study, the bifactor model provides a more detailed analysis of the sources of variance in the DARS. Although the general factor explained most variance overall, we found ω_h to be substantially lower than ω_t, indicating that relevant portions of variance were explained by the specific domain factors. Except for the social activities domain, more variance was explained by the domain group factors than by the general factor on the subscale level. This is also reflected in the standardized factor loadings of the bifactor model where several items loaded higher onto their respective group domain factor than onto the general factor. Among the domain subscales, food/drink showed the highest domain-specific variance and social activities the lowest. Cronbach's α suggested good reliability for the DARS total scale and acceptable to good reliability for its subscales. Our reliability results were slightly lower than the internal consistency reported by Rizvi et al. (2015) and Arrua-Duarte et al. (2019).

As expected, the DARS correlated moderately to highly with other measures of anhedonia or hedonic capacity, thereby corroborating its convergent validity. Specifically, the high correlation between the DARS total score and the SHAPS, r_s = −0.50, resembles the findings of Rizvi et al. (2015) in a healthy sample and Arrua-Duarte et al. (2019) in a diverse clinical sample. The higher correlation of the DARS total score to the TEPS anticipatory compared to the consummatory subscale could be explained by the DARS's inclusion of interest/desire, motivation and effort as three components besides consummatory pleasure. Possibly, anticipation, as measured by the TEPS, connects stronger to these aspects than consummatory pleasure does. Convergent validity also arises from the patterns of correlations between the DARS subscales and other anhedonia measures, especially from the stronger correlations of the subscale social activities to social compared to physical trait anhedonia scales. Moderate correlations of the DARS total score to related constructs, i.e., PA, BAS reward sensitivity and drive, also confirmed its convergent validity. Only the correlation to BAS fun was weak, a pattern also reported by Rizvi et al. (2015) in a community sample. The lack of correlation between the DARS total scale and punishment sensitivity, as measured by the BIS, as well as the only weak correlation to the NA scale of the PANAS indicate good divergent validity. In accordance with our hypothesis, we found a weak to moderate correlation to the CES-D, weaker than Rizvi et al. (2015) reported. This, however, might have been influenced by the shortened CES-D version we applied to assess depression severity.

The DARS contributes to anhedonia research by providing a valid and reliable measure which integrates different components as well as domains of pleasure. So far, its application in German samples can be recommended in contexts where a comprehensive assessment of state anhedonia is aimed for. Importantly, we only examined its validity in young, mentally healthy adults aged 18 to 30. However, since the DARS asks the participants to provide their own examples, it is unlikely to find strong age biases (Sullivan-Toole et al., 2019). Of note, the factor structure of the 17-item DARS only allows to differentiate between different domains, not between the different reward components. The failure to differentiate between the anhedonic components interest, motivation, effort and consummatory pleasure is in line with the results of the initial development (Rizvi et al., 2015) and the Spanish validation study of the DARS (Arrua-Duarte et al., 2019). A recent study on the newly developed Positive Valence Systems Scale (PVSS; Khazanov et al., 2020), comprising six constructs of the Research Domain Criteria's Positive Valence Systems domain (e.g., reward anticipation, reward valuation, reward satiation) and seven reward types (e.g., food, hobbies, goals), similarly only found factors reflecting the reward types, not the constructs. In general, hedonic domains or reward types seem to have a stronger impact on reward-related self-report than hedonic subprocesses do. The most successful attempts to distinguish components or constructs are therefore found only within one domain, for instance anticipatory and consummatory physical pleasure in case of the TEPS (Gard et al., 2006).

In the DARS, the participants themselves provide two or three of their favorite examples for each domain. An assessment of these examples yielded similar results as reported previously (Rizvi et al., 2015; Arrua-Duarte et al., 2019). As described by Rizvi et al. (2015), examples for hobbies and social activities overlapped in categories such as arts and crafts, lifestyle/culture, leisure/games, fitness/wellness/sports, multimedia/technology, education/training and food/cuisine. However, in the social domain, the social character of the activities was emphasized. In addition, we found a category of examples drawing pleasure from interpersonal interaction, relatedness and intimacy per se.

The DARS differs in important aspects from other established and recently developed state anhedonia scales. In contrast to the SHAPS, it is not exclusively targeted to consummatory pleasure. In contrast to the Motivation and Pleasure Scale–Self-Report (MAP-SR; Llerena et al., 2013), a self-report measure of negative symptoms created based on the Clinical Assessment Interview for Negative Symptoms (Forbes et al., 2010), the DARS is not focused on schizophrenia. Clinically, the DARS is most frequently used in depression research (Rizvi et al., 2015; Dhami et al., 2019, 2020; Bibi et al., 2020; Dunlop et al., 2020). However, due to its comprehensive design it can be applied in a wide range of contexts, in participants from healthy and diverse clinical populations. The Specific Loss of Interest and Pleasure Scale (SLIPS; Winer et al., 2014) differs from all scales described above in assessing recent changes in anhedonia, mainly in the social domain regarding consummatory pleasure and interest (Winer et al., 2014).

We further assessed the effect of the COVID-19 pandemic on depression severity and state anhedonia by dividing our sample into a pre- and a during-pandemic group. In line with our hypothesis and many previous reports on the detrimental impact on mental health (Henssler et al., 2020; O'Connor et al., 2020; Pappa et al., 2020; Bueno-Notivol et al., 2021; Cénat et al., 2021), we found slightly but significantly higher depression scores in the during-pandemic group. Importantly, this effect was not diminished when controlling for sociodemographic differences between groups. We found no group differences in state anhedonia, as measured by DARS and SHAPS sum scores. Moccia et al. (2021) reported that the SHAPS significantly predicted depression severity in a cross-sectional survey study during the early Italian lockdown. However, they did not include pre-pandemic data on the SHAPS. Our results fit a recent meta-analysis of longitudinal studies analyzing the effects of lockdown on mental health (Prati and Mancini, 2021). Prati and Mancini (2021) found a small significant effect on mental health symptoms, including depression, but the effects on positive psychological functioning, i.e., satisfaction with life, positive affect, well-being and quality of life, did not reach significance.

Multigroup CFA also gave no indication of a pandemic effect on the DARS factor model or the latent factor means. Interestingly, the measurement model of the DARS stayed invariant even across detrimental changes to the social environment such as a nationwide lockdown and distancing policies. This speaks to the robustness of the factor model and the high generalizability across different social environments. Possibly, the hypothetical phrasing of the items, such as “I would enjoy these activities,” prevented the severe limitations on pursuing certain leisure activities from affecting the ratings. This also holds true for the SHAPS. Of course, the DARS is special in asking the participants themselves to provide their own favorite examples. The measurement invariance might therefore also be an effect of allowing variance in the examples. Some people might have reported other activities and experiences than they would have under normal circumstances. For the social activities and hobbies domains, we compared the frequencies of the categories arts and crafts, lifestyle/culture, leisure/games, fitness/wellness/sports, multimedia/technology, education/training, food/cuisine, home/garden and intimacy/shared time before and during the pandemic but found no significant differences. Possibly, these categories were too broad to detect relevant distinctions. No differences stood out when inspecting the individual examples in both groups. The during-pandemic group entailed examples which might not have been possible to pursue at the time, such as “going dancing” or “going to a concert.” However, answering this question definitively would necessitate an in-depth qualitative analysis which is beyond the scope of this paper.

Among the strengths of our study are the sample size, which is sufficiently large to reliably perform CFA, as well as the inclusion of several established and recently developed anhedonia scales. This enabled us to examine the convergent validity of the DARS more closely, related to state and trait measures of anhedonia in the social and physical domain. Regarding the pandemic group comparisons, the recency of our pre-pandemic group could be a potential benefit as it offers a more accurate estimate than comparisons to data acquired years before can.

Nevertheless, several limitations must be noted. Our sample is non-representative, predominantly female and highly educated. Due to our inclusion criteria, we only assessed young, mentally healthy adults between 18 and 30 years, thereby limiting the generalizability of our findings. Our study only included self-report measures and only assessed the effect of the COVID-19 pandemic cross-sectionally. We collected no information regarding specific stressors related to the COVID-19 pandemic and can, therefore, make no claim to the cause of the increased level of depressive symptoms in the during-pandemic group. Due to the dynamic nature of the COVID-19 pandemic, constant changes in infection rates and government-ordered mitigation strategies, the circumstances at the time point of participation may have varied strongly within the during-pandemic group.

In conclusion, the 17-item DARS in German translation seems to offer a reliable and valid alternative to the SHAPS when a more comprehensive measure of state anhedonia is needed. The correlated four-factor structure, mapping onto the hedonic domains hobbies, food/drink, social activities and sensory experiences, was corroborated. Bifactor analyses substantiated the multidimensionality of the DARS. Moreover, measurement invariance across pre- and during-pandemic groups was established. The DARS satisfied all hypotheses of convergent and divergent validity and showed acceptable to high internal consistency on all subscales. Future studies in other age groups and in diverse clinical settings should further examine the utility of the German DARS. Our results on the effects of the COVID-19 pandemic are in line with many studies all over the world reporting a deterioration of mental health and an increase in depressive symptoms. Our study also adds to the growing literature of positive functioning during the COVID-19 pandemic. In our sample, we found no indication that state hedonic capacity was affected.

The datasets presented in this article are not readily available because of ethical restrictions. Requests to access the datasets should be directed to Sarah A. Wellan, sarah.wellan@charite.de.

The studies involving human participants were reviewed and approved by the ethics committee of the Charité – Universitätsmedizin Berlin. The patients/participants provided their electronic informed consent to participate in this study.

SW, AD, and HW contributed to conception and design of the study. SW and AD conducted the study and performed the statistical analysis. SW wrote the first draft of the manuscript. All authors contributed to manuscript revision, read, and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.