The stability of multidimensional subclinical apathy during a pandemic and its relations to psycho-behavioral factors

Apathy is a clinical symptom prevalent in many neuropsychiatric pathologies. Subclinical apathy is found in 35% of the general population. Despite high prevalence and negative consequences, underlying mechanisms are poorly understood, perhaps because the concept of apathy is one-dimensional. The current investigation aims to address the incidence of multidimensional apathetic trait in three distinct forms in a student population, to specify its determinants and to evaluate its stability during a global pandemic. Two online surveys, conducted 1 year apart on two separate cohorts of university students, with qualitative measures and validated scales. The final analysis included, respectively, 2789 and 1678 students. The three forms of apathetic trait were present, with the same debilitating consequences as apathetic symptom but independent determinants. Executive apathy was predicted by depressive symptoms, emotional apathy by motivational deficit and initiative apathy comprised a mixed executive-emotional form and a pure deficit of action initiation. The three forms of subclinical apathy remained similar in the context of increased depressive symptoms due to a global pandemic. This study confirmed the presence and independence of three forms of subclinical apathy in healthy students, which remained similar even in the light of increased depressive scores. These results shed light on cognitive and neuronal mechanisms underlying multidimensional apathy, allowing new, targeted treatments.

Specifically, 28.68% of respondents presented a single form of apathetic trait, 11.68% combined two forms, and 2.12% had all three forms. Figure 2 illustrates the percentage of combinations for the three forms of apathetic trait.
Analyses of correlations between the three forms of apathetic trait showed that executive and emotional apathy were no correlated (r = − 0.065; p > 0.05), whereas initiative apathy was positively correlated with both executive (r = 0.405; p < 0.001) and emotional apathy (r = 0.205; p < 0.001).
Identification of variables correlated with each form of apathy. Executive apathy. Sociodemographic ANOVAs on executive apathy scores revealed that no sociodemographic factor showed a significant, practical effect.
Emotional apathy. Sociodemographic ANOVAs on emotional apathy scores revealed that gender was the only sociodemographic variable to have a large, significant effect (F(2,2786) = 218; p < 0.001; η 2 = 0.135). Bonferroni tests revealed that men had higher scores than women (t(2786) = 20.85; p < 0.001; d = 0.39), while there was no difference between men and trans individuals or between women and trans individuals.
Initiative apathy. Sociodemographic ANOVAs on initiative apathy scores revealed that gender was the only sociodemographic variable to have a significant, small effect (F(2,2786) = 13.7; p < 0.001; η 2 = 0.010). Bonferroni tests revealed that men had higher scores than women (t(2786) = 4.89; p < 0.001; d = 0.10), while there was no difference between men and trans individuals or between women and trans individuals.
Psychopathology Correlation analyses showed that initiative apathy scores were positively correlated with depressive symptoms (
Specifically, 31.34% of respondents presented a single form of apathetic trait (14.36% executive, 11.56% emotional and 5.42% initiative apathy), 13.41% combined two forms, and 2.09% had all three forms. Figure 4 illustrates a comparison of the percentage of combinations for the three forms of apathetic trait in first-years students, in both studies.
Among the respondents, 12.22% presented a single form of apathetic symptom (4.41% executive, 5.24% emotional and 2.56% initiative apathy), 1.66% had mixed symptoms (0.18% executive and emotional forms, www.nature.com/scientificreports/ 1.07% executive and initiative forms, and 0.42% emotional and initiative forms), and none (0.00%) presented all three symptoms together. Analyses of correlations between the three forms of subclinical apathy showed that executive and emotional apathy were still not correlated in the second study (r = − 0.037; p > 0.05), whereas initiative apathy was still positively correlated with both executive (r = 0.459; p < 0.001) and emotional apathy (r = 0.181; p < 0.001).
Interests of two main predictors: gender and depression. ANOVAs on executive and initiative apathy scores revealed that gender had no significant effect. ANOVAs on emotional apathy scores revealed that gender had a medium, significant effect (F(1,1676) = 161; p < 0.001; η 2 = 0.088). Bonferroni tests revealed that men had higher scores than women (1676) = 12.7; p < 0.001; d = 0.31).
Correlation analyses showed that depressive symptom score on the BDI were positively correlated with executive apathy scores (r = 0.286; Stability of the three forms of apathy. For first-year students, Table 2 presents a comparison of the scores of the three forms of subclinical apathy and depressive symptoms, in 2020 and 2021. Only BDI-II score showed a large significant difference in the new cohort tested 1 year later, with an increase of 5.37 points (BDI 2021-2020 = 14.52-9.15).

Discussion
The results revealed the presence and stability of the three forms of apathetic trait in students over a year and identified form-specific determinants, confirming their independence. In addition, by revealing for the first time a specific link between one form of apathy (the executive form) and depressive symptoms, they shed new light   Subclinical apathy was present in three distinct forms in a student population and associated with the same debilitating consequences as apathetic symptom. The present investigation, based on two studies involving, respectively, 2789 and 1678 students, confirmed the presence of subclinical apathy in three distinct forms in a student population. About 40% of respondents presented an apathetic trait, in agreement with the 35% found in the only other study in a similar population 11 . About 10-14% of respondents presented an apathetic symptom, which was far above the 2% found in other studies with young people 9 . However, in these studies, no distinction was made between the three forms of apathy and the scales used were different: 2% may correspond to participants with the most severe form of apathy (i.e., initiative apathy). An exploration of factors related to education and general functioning clearly revealed, in agreement with some other studies, that subclinical apathy, of whichever form, was associated with debilitating consequences in academic and daily life 13 . Initiative apathy was the most debilitating form in daily life, and emotional apathy specifically in social life. As previously shown, all negative repercussions were similar, although perhaps differing in intensity, to those of the apathetic symptom 19,30 . The present results are thus completely consistent with the usual definition of apathy and support keeping the overarching term "apathy" for all three forms, despite their not having the same underlying mechanisms.
Regarding the population of students, our results show that the dominant form of apathy differed according to field of study, suggesting that course choices may be related to the difficulties associated with the specific form of apathy affecting a student. Students with executive or initiative apathy were more often studying art and literature whereas students with emotional apathy more often studied technical subjects and sciences. This question, as yet unexplored, could be an interesting object of investigation for educational scientists.

From the independence of the three forms of apathy to a better characterization of each.
Our results provide strong evidence for the independence of the three forms of apathy. This is because they highlight that the majority of participants with a subclinical apathy presented a single form of apathy. They also show the existence of specific correlates or predictors for each of the three forms. Of the respondents with apathetic trait, 10-15% had only executive apathy, 10% only emotional apathy and 5-8% only initiative apathy. We expected a lower prevalence of the third form in a young healthy population because it is the most severe form, associated with greater disability 31 .
Most factors specifically linked to a form of apathy concerned psychopathology. Regarding executive apathy, our results suggest that it may be a secondary form. It was the only form specifically linked to psychiatric or somatic disorders and the only one to be predicted by depressive symptoms and fatigue. Executive apathy may www.nature.com/scientificreports/ therefore be secondary to another disorder, potentially modulated by executive deficits. In the literature, executive apathy has been linked to deficits in executive functions (Alzheimer: Perri et al. 25 ; Schizophrenia: Raffard et al. 26 ), executive deficits which also correlate with depression 32,33 and fatigue 34,35 . These results cast doubt on the nature of executive apathy, but also shed light on the confusion between apathetic and depressive symptoms, suggesting that the overlap may be specific to the executive form. Emotional apathy, in contrast, was the only form to be predicted by the two motivational components: consummatory and anticipatory pleasure. However, only consummatory pleasure was specific to emotional apathy; anticipatory pleasure also predicted the initiative apathy. Emotional apathy was therefore the only form linked to a deficit of hedonic pleasure in the moment. To our knowledge, the only link previously highlighted between apathy and motivation was non-specific: a link between deficit of anticipatory pleasure and one-dimensional apathy 36 . Our results revealed the importance of considering all the components of motivation to improve understanding of apathy in its multidimensional form.
Finally, and interestingly, our results may suggest the presence of two subtypes of initiative apathy: a mixed form with both executive and emotional deficits, and a specific deficit. Initiative apathy was correlated with the two other forms of apathy over 2 years. Moreover, in 60-70% of cases, initiative apathy was present in association with another form. This mixed form hypothesis is also supported by neuroanatomical and clinical studies in patients with neurodegenerative disease, especially Parkinson's and Alzheimer's 25,37,38 . These studies showed a combination of cognitive and emotional deficits and the presence of lesions characteristic of the two other forms. The second subtype of initiative apathy could be a pure deficit in action initiation, the last step before action is taken. Although the percentage was low, the fact remains that 5-8% of our sample had initiative apathy without any other associated form. Such pure initiative apathy has previously been highlighted in amyotrophic lateral sclerosis, confirming the existence of this pure form as a symptom 39,40 related to specific cerebral regions (anterior cingulate cortex and supplementary motor area) in patients 24,41 and in healthy people 31,42 .
The stability of subclinical apathy and the effects of the COVID-19 pandemic. The three dimensions of subclinical apathy were stable under different circumstances. Indeed, before and during a global pandemic, there were no major variations of means, standard variations, correlates, predictors and cut-off for apathetic trait and symptom. Only initiative apathy as a pure deficit of action initiation showed a slight decrease in the second study. However, social distancing and lockdown could explain this reduction, by changed social norms and reduced everyday activities for everyone.
Furthermore, the stability of apathy contrasts with the sharp increase in depression. One year later, depression was still a specific predictor of executive apathy. However, in the second study, the rise of depression during the COVID-19 pandemic allowed us to improve our knowledge of the link between executive apathy and depression. Depressive symptoms doubled over 1 year in students, probably because of lockdowns, curfews and psychological difficulties induced by the COVID-19 pandemic 43,44 , but executive apathy remained stable. Thus, even if depressive symptoms are a strong predictor of executive apathy, it appears that executive apathy remains a separate symptom with other explanatory factors.

What role do gender and self-esteem play in apathy?
To our knowledge, no other studies explored the link between gender and a specific form of apathy. However, others have revealed a positive correlation between male gender and one-dimensional apathy 30,45 . Our results showed more precisely that men presented more emotional apathy than women and trans individuals whereas transgender experienced more executive apathy than cisgender (male and female). The male gender was still a predictive factor of emotional apathy 1 year later. Although no real explanation for this association is yet apparent, controlling for gender seems important for studying apathy.
The identification of self-esteem as a predictor for executive and initiative apathy may offer new perspectives for non-pharmacological therapeutic interventions focused on enhancing self-esteem when these two forms of apathy are present. Two previous studies, involving patients with traumatic brain injury 41 and psychotic disorders 36 , showed that low self-esteem, by inducing demotivating beliefs, may prevent people from undertaking new or complex tasks, thereby inducing apathy.

Conclusion
This study confirmed the existence and stability of multidimensional subclinical apathy in young healthy students under different circumstances (global pandemic vs not) and showed that apathetic trait was associated with the same debilitating consequences in all spheres of daily life as apathetic symptom. By identifying specific characteristics for three forms of apathy, the present results may aid exploration of the mechanisms underlying multidimensional apathy (especially self-esteem, anticipation, and motivation) in healthy people and in various pathologies to propose new, targeted therapies.

Methods
The current investigation, based on two studies, aims to address the incidence of multidimensional apathetic trait in three distinct forms (executive, emotional and initiative) in a student population, to specify its determinants and to evaluate its stability during a global pandemic. Two online surveys, conducted 1 year apart, tested 2789 and 1678 different university students, with qualitative measures and validated scales.
Participants. Participants were recruited in a French University. All registered students received an email giving them the opportunity to participate in an online survey about 'Personality in university students' , created on LimeSurvey. To participate, the only criteria were to be a student, aged between 18 and 28 years and to www.nature.com/scientificreports/ be a native French-speaker. All participants gave electronic informed consent, and the study was approved by the University review board (UNISTRA/CER/2020-13) and the French data protection authority (CNIL). They weren't paid for their participation. In total, for study 1, 3144 students (26.5% male) completed the survey. After eliminating 355 respondents who did not match the criteria, 2789 students constituted the study sample. For study 2, 1678 new first-year students (38.6% male) constituted the study sample (see Table 3 for sociodemographic characteristics of the studies).
Online survey. For Study 1, the survey was divided in four parts: sociodemographic characteristics, education, general functioning, and psychopathology (see Supplementary data 2 online). All these factors were assessed with qualitative measures. Four validated scales were added to complete the assessment of psychopathology. The names and themes of the four questionnaires were not specified online. is an 18-item scale assessing anticipatory (10 items) and consummatory (8 items) pleasure. Each item is score from 1 to 6. Anticipatory pleasure refers to pleasure due to active search for or anticipation of a pleasant activity (score range: 10 to 60). Consummatory pleasure refers to pleasure directly due to accomplishing a pleasant activity (score range: 8 to 48). A low score for each subscale indicates pleasure disorder 46 .
For study 2, the survey was divided into two parts: sociodemographic characteristics and psychopathology. Two validated questionnaires were used to assess the psychopathology: the DAS 47 and the BDI-II (Beck et al. 52 ).
See Supplementary data 3 online for correlations between all the questionnaires for both studies.
Statistical analysis. Each participant had three DAS subscores: one for each form of apathy.
To determine the percentage of participants with a subclinical apathy, a cut-off was calculated for each subscore, based on the mean and the standard deviation of studies samples (Study 1, 2789 students, Study 2, 1678 first-year students). The suggested cut-offs were + 1 standard deviation from the mean for subclinical apathy/ apathetic trait and + 2 standard deviation from the mean for clinical apathy / apathetic symptom 11,39,48 .
Both studies had large sample size, so according to the Central Limit Theorem, data could be approximated by a normal distribution and parametric tests were used 49 .
Pearson correlation tests were implemented between DAS scores. To identify variables correlating with each form of apathy, Pearson correlation tests were implemented with continuous variables; ANOVAs and Bonferroni post-hoc tests were used with categorical variables.
Exploratory approach was used for regressions analyses since the selection of predictor variables was not based on a priori hypotheses. Predictors of each form of apathy were specified by multiple linear regression analyses, using one DAS score as reference and all the other variables from the online survey as possible regressors of interest (see Supplementary data 4 online). The purpose was to determine which variables in the survey Table 3. Sociodemographic characteristics of the sample. www.nature.com/scientificreports/ most influenced each form of apathy, thanks to the Akaike Information Criterion (AIC). The smaller the AIC value, the better the model fit, using the fewest possible variables. R-square values (the percentage variation of the reference explained by the regressor of interest) were calculated. In study 2, to compare questionnaire scores across time, Student's t test was used when there was an assumption of equal variances with Levene's test (p > 0.05); Welch's t test was used when there was a violation of the assumption of equal variances with Levene's test (p < 0.05).
The statistical significance level was set at 0.05. As both studies have a large sample size, effect size measures were always added, to prevent big data bias and to analyze only practical effects 50,51 . For correlation squared (R 2 and η 2 ) used in regressions and ANOVA, practical effect was small around 0.01, medium around 0.06 and large from 0.14. For correlation coefficient (r) used in correlations, practical effect was small around 0.10, medium around 0.30 and large from 0.50. For standardized mean difference (d) used in t-test, practical effect was small around 0.20, medium around 0.50 and large from 0.80. Only practical significant results are described in the Results section below.
Ethical standards. The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national and institutional committees on human experimentation (Research ethics committee of the University of Strasbourg-UNISTRA/CER/2020-13) and with the Helsinki Declaration of 1975, as revised in 2008.

Data availability
The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.