Cardiorespiratory fitness mediates the relationship between depressive symptomatology and cognition in older but not younger adults

Aging is commonly associated with emotional, physical, and cognitive changes, with the latter, particularly affecting executive functioning. Further, such changes may interact. For instance, depressive symptomatology is a known risk factor for developing cognitive deficits, especially at older ages. In contrast, an active lifestyle, reflected in high cardiorespiratory fitness (CRF) levels, has proven to protect against adverse effects on cognition across the adult lifespan. Hence, this study aimed to investigate the relationships between depressive symptomatology, CRF, and cognition during critical developmental stages, namely in young adults (YA), when cognitive abilities are at their peak, and in older adults (OA), when they may start to decline. Eighty-one OA with ages between 60 and 89 years ( M = 70.46; SD = 7.18) and 77 YA with ages between 18 and 34 years ( M = 22.54; SD = 3.72) went through (i) a sociodemographic interview, (ii) an emotional assessment, (iii) a battery of cognitive tests, and (iv) a physical evaluation assessing CRF levels, visceral fat and body-mass index. Results showed that OA exhibited lower general cognitive performance, inhibitory control, cognitive flexibility, memory, and CRF. Depressive symptoms and anxiety were not different among groups, with CRF mediating the relationship between depressive symptoms and cognition in the OA group. The present study provides valuable insights into the interplay between emotional, physical, and cognitive well-being. Additionally, it calls attention to how lifestyle factors can play a protective role against the adverse effects that depressive symptoms have on cognition, particularly at older ages.


Introduction
Aging, as part of biological development, is associated with emotional, physical, functional and cognitive changes.Despite the interindividual variability observed (Wu et al., 2021), late adulthood can be a period of vulnerability to cognitive deficits, particularly regarding executive functions (EF) (Kirova et al., 2015).Evidence shows that older adults (OA) perform worse in different EF domains than younger adults (YA) (Kirova et al., 2015;Bedard et al., 2002).Further, previous studies have demonstrated that age-related cognitive decline manifests an augmented likelihood of conversion into dementia when concomitant depressive symptoms are present (Diniz et al., 2013;Modrego and Ferrández, 2004).Interestingly, depression, alongside dementia, is the most common mental disorder in the OA population, which presents a higher percentage when compared to any other age group (World Health Organization, 2023).Even YA shows impairments in complex tasks requiring EF when suffering from depression (Castaneda et al., 2008).Hence, the presence of depressive symptomatology is an important risk factor to account for when investigating cognitive decline across the lifespan.
Nevertheless, other factors, such as physical activity, seem to mitigate the rate of cognitive decline and the effects of depressive symptomatology on cognition at different ages.Thus, physical activity and cardiorespiratory fitness (CRF) have been shown to act as protective factors that can alter the trajectory of age-related cognitive decline (Falck et al., 2019;Northey et al., 2018).CRF is a component and objective measure of physical activity that reflects the ability of the cardiovascular and respiratory systems to supply oxygen during prolonged physical activity (Lee et al., 2010).Specifically, higher CRF has been associated with overall better cognitive function (Freudenberger et al., 2016), including memory (Hayes et al., 2016;Vesperman et al., 2022) and EF (Freudenberger et al., 2016;Hayes et al., 2016;Mekari et al., 2019;Pentikäinen et al., 2019).This is particularly important in advanced stages of life, once CRF tends to decline with aging (Jackson et al., 2009).Furthermore, lower CRF levels in middle-age have also been associated with a risk of developing dementia in later life (DeFina et al., 2013).
Considering this evidence, a cardiorespiratory hypothesis has been proposed, stating that improved cognitive functioning may be partly explained by the positive physiological processes that result from physical activity (Agbangla et al., 2019).In accordance, previous research has shown that physical activity and CRF are associated with structural and functional brain changes, such as increased gray matter volume (Raichlen et al., 2020), promotion of neuroplasticity in the medial temporal lobes (Raichlen et al., 2020;Erickson et al., 2011), and improved hippocampus and prefrontal cortex connectivity in YA (Kronman et al., 2020).Specifically, Kronman et al. (2020) showed that people with higher CRF had more effective connections between the hippocampus and prefrontal areas -regions directly related to EF (Mekari et al., 2019;Monteiro-Junior et al., 2016) -but also to emotional circuits -that are commonly involved in depression (Firth et al., 2020).
Despite the substantial influence of genetic predisposition in the etiology of mental health disorders such as depression (Hyde et al., 2016), emerging empirical data has highlighted the role of environmental and lifestyle determinants.For example, several meta-analyses have shown that low physical activity is associated with a higher prevalence of depression (Schuch et al., 2018;Teychenne et al., 2010).Accordingly, studies show that individuals with depression engage in 50 % less moderate-to-vigorous physical activity (Vallance et al., 2011), presenting lower CRF than the general population (Boettger et al., 2009).Similarly, a systematic review conducted by Schuch et al. (2016) revealed that compared to individuals with high CRF, those with low CRF experienced a 76% higher incidence of depression, whereas those with medium CRF had a 23 % elevated likelihood of suffering from depression.Nonetheless, a meta-analysis found a modest correlation (r = − 0.16) between the severity of depressive symptoms and CRF in healthy and depressed individuals (Papasavvas et al., 2016).
Based on the spectrum of emotional, cognitive and somatic consequences associated with depressive symptoms, several hypotheses have emerged to elucidate the potential mechanisms underlying the role of CRF.First, the emergence of depressive symptoms may lead to reduced motivation to engage in physical activity, resulting in greater physical inactivity, which, in turn, leads to decreased CRF levels (Hollenberg et al., 2003;Nikolakaros et al., 2020).Second, depressive symptoms are associated with multiple disturbances, such as sleep disorders (DSM-5) (American Psychiatric Association, 2013) that can disrupt the role of sleep in the restoration and consolidation of physical and cognitive health, contributing to greater physical inactivity and obesity (Watenpaugh, 2009), as well as worse cognitive performance (Rasch and Born, 2013).Finally, the psychosocial factors associated with depressive symptoms, such as social isolation, anxiety and negative affect (DSM-5) (American Psychiatric Association, 2013), can disrupt healthy lifestyle behaviors and, thus, have detrimental effects on CRF and cognition.
Considering these postulates and the underlying analogous neural correlates of depression, cognition, and CRF, our main aim was to further explore these relationships at two different stages of the lifespan.Additionally, we intended to evaluate if depressive symptoms and CRF were predictors of general cognition, short-term memory, and EF at those stages, i.e., when these abilities are at their peak -YAand when they may start to decline -OA.Additionally, we tested if an active lifestyle proxy, such as the CRF, could mediate the relationship between depressive symptoms and cognition.Specifically, and considering the evidence on the relationship between depression, CRF and cognitive status, we hypothesized that OA would present higher depressive symptomatology, lower CRF, and lower cognitive abilities (worse performance on tests measuring overall cognition, short-term memory, and EF) in comparison with YA; and that the adverse effects that an increase in depressive symptomatology has on cognition would be driven by the decrease of CRF at both age-groups.

Participants
OA over 60 years of age were recruited through contacts with associations and daycare centers.In addition, YA between 18 and 35 years of age were recruited at the university campus and received course credits for their participation in the study.Study protocols were in accordance with the principles outlined in the Declaration of Helsinki and received approval from the local ethics committee (CE.CVS 095/2018).Exclusion criteria were presenting history of stroke, transient ischemic attack, head injury, epilepsy, Parkinson's disease, Alzheimer's disease, or other neurological or psychiatric diseases.Further, volunteers were excluded if they were taking anxiolytic or antidepressant medication or scored below the cut-off for probable dementia established in the Portuguese version of the Mini-Mental State Examination (MMSE) (Folstein et al., 1975) by Santana and colleagues (2016).In addition, included participants scored for total independence or mild dependency in the Instrumental Activities of Daily Living Scale (IADL) (Lawton et al., 1969;Reis et al., 2012).A priori sample size calculations on G*Power software (https://www.psychologie.hhu.de,accessed on 12 November 2019) estimated for the regression analyses a minimum sample size of participants per age group, considering a medium effect size (f 2 = 0.15), an alpha (α) of 0.05, and a statistical power of 0.80.
Eighty-one participants (71.6 % female) with ages between 60 and 89 years (M = 70.46;SD = 7.18) were included in the OA group, while 77 participants (64.9 % female) with ages between 18 and 34 years (M = 22.54; SD = 3.72) were included in the YA group.(Folstein et al., 1975).General cognition was evaluated with the MMSE.A maximum of points can be achieved, and cutoff scores are adapted according to the participants' years of formal education as established in the Portuguese version of the MMSE (illiterate: ≤ 15; from 1 to 11 years of formal education: ≤ 22; >11 years of formal education: ≤ 27 (Santana et al., 2016);).(CANTAB®, 2019).CANTAB is a validated cognitive research software that comprises 18 standardized tests to assess different domains of cognitive function.It was used to evaluate memory and EF using the Spatial Span and the Multitasking tests, respectively.

Cambridge neuropsychological test automated battery -CANTAB
The Spatial Span (SSP) test is based on the Corsi block tapping test and measures short-term memory.In this test, a set of gray boxes lights up a color in a specific sequence.The participant's task is to remember the sequence and then touch the boxes on the screen in the same order.The sequence length increases throughout the test, and the participant has a maximum of 3 attempts at each sequence.The longest sequence of boxes successfully recalled by the participants was recorded for further analyses.
The Multitasking (MTT) test measures cognitive flexibility and inhibitory control through the ability of the participant to use multiple sources of potentially conflicting information to guide behavior.In each trial, an arrow appears on the middle, left, or right side of the screen, and the participant is asked to click on the right/left side button in accordance with the direction or the side that the arrow points to or appears on the screen.During training, the participant learns to respond according to the arrow's direction/side of the screen, separately.During the assessment stage, each trial is preceded by a cue indicating whether the participant should respond according to direction or side (the rule is changed randomly).In some trials, the direction of the arrow and the side in which the arrow appears are incongruent.The outputs used for further analyses were the Incongruency Cost, an indicator of inhibitory abilities, and the Multitasking Cost, a measure of cognitive flexibility.The former is calculated by subtracting the mean response latency (in ms) of congruent trials from that of the incongruent trials.A higher Incongruency Cost indicates that the participant takes longer to process conflicting information.The latter output is calculated by subtracting the mean response latency during single task block(s) from that during multitasking block(s).A positive score indicates difficulties in managing multiple sources of information (i.e., less flexibility).(Portuguese validation, Campos and Gonçalves, 2011).This self-report questionnaire comprises 21 sets of 4 statements.For each set, the participant must choose the statement that better describes how they have felt for the last two weeks.The score ranges from 0 to 63 points, with higher scores indicating more depressive symptoms.The internal consistency of the Portuguese version is high (Cronbach's alpha = 0.90), and it shows high convergent validity with other depressive symptomatology scales (Campos and Gonçalves, 2011).Cruz and Mota, 1997;Spielberger et al., 1983).STAI-Y measures state and trait anxiety separately, with 20 questions for each kind.For each question, the participant has to rank on a 4-point Likert scale how much that sentence describes them at the moment (state scale) or generally in their life (trait scale).Scores range from 0 to 80 points for each scale.In the current study, we have only used the trait subscale, where a higher score indicates more severe anxiety as a personality trait.The Portuguese version of STAI-Y has demonstrated high internal consistency (Cronbach's alpha = 0.85) even when the trait subscale is considered alone (Cronbach's alpha = 0.88.

Physical measures
2.2.3.1.International physical activity questionnaire -IPAQ (Craig et al., 2003).This self-report questionnaire evaluates physical activity across five domains, including activity related to work, physical activity as a means of transport, domestic and gardening activities, leisure time activity, and sedentary time.The questions under each domain provide the score for walking, moderate-intensity activity, vigorous-intensity activity, and overall activity level.Responses to the questionnaire were used to categorize participants according to the five levels of physical activity defined by Jurca et al. (2005) for the calculation of CRF without performing exercise testing.

Anthropometric assessment.
The participants' height was measured by a stadiometer.Weight and visceral fat level were measured with a bio-impedance scale (TANITA MC-780MA Segmental).Resting heart rate (HR) values were collected with a wrist pulsometer (POLAR M200).Finally, height and weight were used to calculate the body-mass index for each participant using the standard formula of weight (Kg) divided by the square of height (in meters).

Cardiorespiratory fitness -CRF.
CRF was estimated from the following equation (Jurca et al., 2005): CRF = (sex*2.77)-(Age*0.10)-(BMI*0.17)-(Resting Heart Rate*0.03)+ Self-reported Physical Activity derived from IPAQ +18.07.Sex was coded as 0 for females and 1 for males.Jurca's et al. (2005) equation shows similar estimates of CRF as graded exercise tests (GXT) -considered the gold standardwith the advantages of being a simple, low-cost and low-risk measure, especially beneficial for clinical and research settings with OA.
Table 1 summarizes OA and YA's mean and standard deviation for emotional (BDI-II and STAI-Y), cognitive (MMSE score and CANTAB subtests results), and physical (CRF, visceral fat and BMI) variables.

Procedure
Data collection was carried out in two different moments.First, participants in the study received a detailed explanation of the procedures, provided written informed consent, and completed a standard interview to collect demographic information and assess their physical and cognitive health status, ensuring compliance with the exclusion and inclusion criteria.Then, in a second session, participants underwent (1) a cognitive evaluation using CANTAB, (2) an emotional assessment using BDI-II and STAI-Y Trait, and (3) a physical evaluation and IPAQ assessment.The physical evaluation encompassed the measurement of resting heart rate with a pulsometer (POLAR M200), the utilization of bio-impedance scale (TANITA MC-780MA Segmental) and a stadiometer to later determine the CRF level.

Statistical analysis
Statistical analyses were conducted using IBM Statistical Package for Social Sciences (SPSS; Version 27).Scores for each variable (i.e., age, sex, BDI-II, STAI-Y Trait, MMSE, MTT Incongruency Cost, MTT Multitasking Cost, SSP Forward Span Length, CRF, Visceral fat, BMI) that deviated at least 2.2 times the interquartile range from its mean were considered univariate outliers (Hoaglin and Iglewicz, 1987) and winsorized before conducting the statistical tests (Tukey, 1962).The alpha level was set at p ≤ .05.
Firstly, we conducted a series of independent samples t-tests to evaluate the possible differences between groups on CRF, MTT Incongruency Cost and MTT Multitasking Cost.For the variables whose distributions deviate from normality (MMSE, BDI-II and STAI-T scores, SSP Forward Span Length, Visceral fat and BMI levels), Mann-Whitney tests for independent samples were conducted to test the possible differences between groups.Secondly, for each age group separately, linear regression analyses were used to evaluate the predicting value of depressive symptomatology (BDI-II) on the cognitive and physical variables, as well as the predicting value of the physical variables (used as independent predictors, given the high correlation among them and the potential effects of multicollinearity) on the cognitive ones.Finally, we conducted mediation analyses using model 4 of PROCESS macro for SPSS v4.3 (Hayes, 2017) with the emotional, physical, and cognitive variables.Notes: SD = standard deviation.
C. Barros et al.
Other regression analyses having the STAI-T (anxiety measure), Visceral fat and BMI as predictors of cognition and physical variables are shown in Supplementary Table 1.

Mediation models
Since no regression models were significant for the YA group (see Table 2), mediation models were only carried out for the OA group.We used the cognitive measurements significantly predicted by depressive symptoms and CRF as output variables separately.Consequently, mediation analyses were performed to assess the mediation role of CRF in the relationship between depressive symptomatology and 1) general cognition, 2) cognitive flexibility, and 3) memory.

Discussion
The main objective of this paper was to explore the interplay between depressive symptomatology, cognition, and CRF at two different stages of the lifespan: YA and OA.Our results showed that OA presented lower cognitive performance in all cognitive tests, lower CRF, and higher visceral fat and BMI scores, according to our hypothesis.However, no differences were found between both age-groups for depressive symptoms and anxiety.Furthermore, as expected, CRF mediated the relationship between depressive symptoms and cognition (general cognition and memory).However, contrary to our hypothesis, this mediation was only significant in OA, suggesting that the adverse effects that depressive symptomatology has on cognition seem to be driven by decreased CRF at older ages.
Regarding age-group differences, we found that OA, compared to YA, presented lower general cognitive performance, inhibitory control, cognitive flexibility, memory and CRF.These differences are in accordance with previous literature showing that aging can be associated with cognitive decline in multiple domains (Kirova et al., 2015;Bedard et al., 2002) and with lower CRF (Jackson et al., 2009).However, we found no group differences in depressive symptomatology.Even though OA is C. Barros et al. widely described as displaying higher levels of depression (World Health Organization, 2023), our sample comprised only individuals with subclinical symptoms (with current depression diagnosis or medication being exclusion criteria), which could explain the lack of differences.Moreover, although depressive symptomatology predicted active lifestyle indices (e.g.CRF) for both age groups, it predicted cognitive performance only on the OA.These results suggest that while the impact of depressive symptoms on active lifestyles, as reflected by CRF, can be observed from a young age, their impact on cognition manifests solely in later stages of life.This goes in line with longitudinal studies showing that, in older populations, the level and duration of depressive symptoms are associated with lower CRF levels measured two or even four years later (Hollenberg et al., 2003).
As previous literature has shown, CRF can be a protective factor against age-related cognitive decline (Falck et al., 2019;Northey et al., 2018) and further protect from conversion to dementia (DeFina et al., 2013).Accordingly, our results showed that CRF was a predictor of cognitive functioning in all the domainsoverall cognition, short-term memory, and EFin OA but not YA.This pattern aligns with Hayes et al. (Hayes et al., 2016), who also found that CRF was associated with cognition and EF only in OA.The lack of an association between CRF and cognition in YA may be attributed to the fact that cognitive abilities usually reach their peak performance during this life period.Consequently, the impact of CRF on cognition may be masked by other factors within this age-group.Overall, the present results are in accordance with and extend the age-dependence hypothesis (Hötting and Röder, 2013), which sustains that CRF impacts cognitive and brain function during childhood, with its influence fading during young adulthood.Our results showed that, in older adulthood, when cognitive decline may develop, CRF exerts once again a beneficial impact on cognitiongeneral  cognition, short-term memory, and EF.Furthermore, studies show that for OA, key factors influencing their quality of life revolve around health, social connections, independence in daily activities, and engagement in an active lifestyle (Agustí et al., 2023).Additionally, when considering the sense of satisfaction with health status for OA, the absence of physical illnesses and psychological problems becomes essential, as well as maintaining adequate levels of physical activity (Rizo, 2017).Thus, an active lifestyle seems to emerge as pivotal determinant in enhancing the overall well-being and life satisfaction among OA, alongside improving cognitive abilities.
Regarding the mediation models, results revealed the mediating role of CRF in the interplay between depressive symptomatology and general cognition as well as memory, but only among OA.In essence, decreased CRF was shown to drive the negative effect that depressive symptoms have on general cognition and memory performance.Previous literature has evidenced the adverse effects that depressive symptomatology has on cognition, especially in OA (Diniz et al., 2013;Modrego and Ferrández, 2004), while also showing the protective role that greater CRF can have on cognition (Freudenberger et al., 2016) and memory (Hayes et al., 2016;Vesperman et al., 2022).Thus, in the presence of depressive symptomatology, individuals tend to exhibit reduced levels of physical activity (Vallance et al., 2011), and this decline may be associated with a subsequent decrease in CRF (Boettger et al., 2009).Results from this study corroborate this hypothesis, showing that this chain of events may have significant negative implications for cognitive and memory performance.
Several mechanisms have been suggested that explain the complex relationship between depressive symptomatology and CRF and their impact on cognitive abilities.For instance, depressive symptoms encompass a lack of motivation, social isolation, anxiety, negative affect, and sleep disorders (American Psychiatric Association, 2013) that have been shown to contribute to greater physical inactivity (Nikolakaros et al., 2020;Watenpaugh, 2009).Our results show that, in fact, depressive symptoms have a detrimental impact on physical well-being, predicting worse CRF in both OA and YA.However, despite previous studies that have shown that depressive symptoms also have a negative impact on cognitive performance (Rasch and Born, 2013), our statistical models suggest that the negative impact of depressive symptoms on cognitive performance is totally mediated by CRF.That is, a decrease in CRF capacity explains the negative effect that higher depressive symptoms have on cognition.These results further extend previous literature by showing that an individual's lifestyle and physical well-being better explain the influence of depressive symptoms on cognition.
Other hypotheses suggest that these results may be explained by the overlap in the neural bases of the effects of a healthy and active lifestyle and those of depressive symptomatology, which affect brain regions essential for cognitive performance (Raichlen et al., 2020;Kronman et al., 2020;Firth et al., 2020).This is consistent with the cardiorespiratory hypothesis (Agbangla et al., 2019), which establishes that improved cognitive functioning is related to the positive neurobiological outcomes of physical activity.Nevertheless, no mediation model was found regarding EF, although there was an interaction between depressive symptomatology and CRF predicting EF performance.In this way, a possible moderator role of CRF in the relationship between depressive symptoms and EF was observed, particularly concerning cognitive flexibility.In other words, the negative effect of depressive symptomatology on cognitive flexibility can be exacerbated in individuals with low CRF.Thus, it seems that a different mechanism involving an interaction between depressive symptomatology and CRF may be responsible for the impact of depressive symptoms on higher cognitive abilities such as EF.Future studies are needed to clarify the moderator role that depressive symptomatology and an active lifestyle play in shaping EF in later life.
The present work showed that the effect of depressive symptomatology on cognition does not appear at younger ages.It is important to note that our sample comprised subclinical depressive symptoms, meaning that all participants were healthy.Additionally, during young adulthood, cognition, as well as brain function and structure, are at their peak.Thus, it is possible that subclinical symptoms are not enough to impact cognition compared to clinical depressive symptoms -which previous literature has shown to affect cognition at younger ages (Castaneda et al., 2008).In contrast, according to our results, in old age, when cognition may start to suffer alterations, even subclinical symptoms seem enough to trigger and exacerbate cognitive differences.However, the cross-sectional design of our study poses an undeniable limitation.Additionally, our focus solely on subclinical depressive symptoms overlooks the potential influence of more severe depression symptoms, on YA, or other emotional well-being constructs on both agegroups, which should be considered as well.Moreover, CRF serves just one aspect of an active lifestyle, providing only a limited perspective on this multifaceted concept, which is also a constraint of our study.Therefore, further studies are needed to evaluate the effect of clinical depression on cognition, considering longitudinal approaches and accounting for active lifestyle-related variables to establish more robust conclusions.
Nevertheless, the present paper emphasizes the pivotal role of CRF as a potential active lifestyle index mediating the impact of depressive symptomatology on cognitive function and memory abilities, particularly in OA.Thus, early detection and intervention for depressive symptoms and promoting CRF through physical activity may help delay or diminish age-related cognitive decline.However, given the differences between age-groups, interventions addressing cognitive health and emotional well-being should be tailored to each age-group specific needs and characteristics and designed to target specific cognitive domains for the most successful outcome.Thus, understanding the intricate relationships between emotional, cognitive, and physical wellbeing in different stages of adulthood calls for an in-depth approach to health promotion.These practical implications can guide the development of targeted interventions and health promotion strategies to support the population's emotional, cognitive, and physical well-being.

Conclusions
In conclusion, our results demonstrated that OA presented lower performance on tests measuring overall cognition, short-term memory, and EF, as well as lower CRF than YA.Additionally, subclinical depressive symptomatology predicted CRF in OA and YA while also predicting cognitive abilities only in OA.Likewise, CRF predicted cognitive performance only in the OA group.Finally, the adverse effects that an increase in depressive symptomatology has on cognition and memory may be driven by the decrease of CRF in late life.Hence, the present study stands out in its comprehensive exploration of the relationships between emotional, cognitive, and physical well-being across two different and crucial stages of adulthoodwhen cognition is at its peak and when it may start to decline.Further, it highlights the pivotal role of CRF as an index of active lifestyles that protects from the impact of depressive symptomatology on cognitive function and memory abilities in OA, while also being the first study to indicate its mediating role between depressive symptoms and cognition.Thus, our results suggest avenues for improving the quality of life of OA by emphasizing the importance of maintaining an active lifestyle and managing depressive symptoms.This may help maintain a healthy cognitive trajectory into older age, fostering longer years of independence and autonomy.Interventions aimed at enhancing CRF and addressing even subclinical depressive symptomatology could potentially mitigate age-related cognitive decline and contribute to better cognitive outcomes in late life.Future studies could build upon our study by incorporating additional variables related to emotional wellbeing and active lifestyle while employing machine learning methodologies to have a deeper understanding of those variables' impact on cognitive processing across different ages.
Supplementary data to this article can be found online at https://doi.

Declaration of competing interest
Authors declare no conflict of interest Fig. 1.Scatter plots showing the relationship between cognitive performance scores and BDI-II (left column) and CRF scores (central columns) for OA.The relationship between BDI-II scores and physical variables (visceral fat and CRF) is shown on the right column for OA (top) and YA (inside the black square).

Fig. 3 .
Fig. 3. Schematic descriptions of total and mediated effects of depressive symptomatology (BDI-II) on memory.(Top) Schematic description of the total effect of depressive symptomatology (BDI-II) on memory (SSP Forward Span Length) -path c-.(Bottom) Schematic description of the mediated effect of depressive symptomatology (BDI-II) on memory through CRF -path β 1 and β 2 , and c'-.

Table 2
Regression models with the BDI-II score and CRF level as predictors and cognitive and physical variables as outcomes for the OA and YA groups, separately.