Schizophrenia is among the world’s most debilitating illnesses [Reference Murray and Lopez1]. Despite the striking nature of psychotic symptoms in schizophrenia, neurocognitive dysfunction is recognized as a core feature of the condition. These deﬁcits are present from the first psychotic episode or earlier, impairing functional recovery and resulting in substantial socio-occupational disability [Reference Bagney, Rodriguez-Jimenez, Martinez-Gras, Sanchez-Morla, Santos and Jimenez-Arriero2, Reference Green, Kern, Braff and Mintz3]. The most pronounced cognitive impairments among patients with schizophrenia are associated with memory, executive function, and processing speed [Reference Dickinson, Ragland, Gold and Gur4, Reference Nuechterlein, Barch, Gold, Goldberg, Green and Heaton5]. Furthermore, antipsychotic medications have little impact on cognition [Reference Goldberg, Goldman, Burdick, Malhotra, Lencz and Patel6], and other pharmacological approaches for the treatment of these deficits have demonstrated limited success [Reference Keefe, Buchanan, Marder, Schooler, Dugar and Zivkov7].
Physical performance is deﬁned as the ability to satisfactorily perform muscular work, and is influenced by both physical capacity and motor abilities [Reference Bouchard, Shephard, Stephens, Sutton and McPherson8]. Various components of physical performance are significantly correlated with everyday functioning, including strength, ﬂexibility, balance, mobility, motor coordination, muscle mechanics, and cardiorespiratory ﬁtness. Among these, cardiorespiratory fitness [Reference Strassnig, Brar and Ganguli9] and functional mobility [Reference Viertiö, Sainio, Koskinen, Perälä, Saarni and Sihvonen10] are notably impaired in patients with schizophrenia.
Profoundly low cardiorespiratory ﬁtness (CRF; i.e., the ability of the circulatory and respiratory systems to supply oxygen to working muscles during sustained physical activity [Reference Colcombe, Kramer, Erickson, Scalf, McAuley and Cohen11]) is present from the first psychotic episode in patients with schizophrenia [Reference Vancampfort, Rosenbaum, Probst, Soundy, Mitchell and De Hert12]. Low CRF is a strong and independent predictor of cardiovascular disease (CVD; [Reference Laursen, Nordentoft and Mortensen13]), which is the largest single cause of death among patients with schizophrenia [Reference Hennekens, Hennekens, Hollar and Casey14]. In addition, low CRF has been associated with cognitive dysfunction in the general population [Reference Voss, Chaddock, Kim, VanPatter, Pontifex and Raine15, Reference Etnier, Nowell, Landers and Sibley16], and several studies have also reported this relationship in patients with schizophrenia [17–Reference Firth, Stubbs, Rosenbaum, Vancampfort, Malchow and Schuch21].
Functional mobility is defined as the capacity to move through the environment (e.g., standing, bending, walking) in order to perform activities of daily living (ADLs) [Reference Forhan and Gill22]. Functional mobility has emerged as a useful clinical marker for predicting overall health decline [Reference Viccaro, Perera and Studenski23], disability in ADLs [Reference Huang and Reichardt24], and health-related quality of life [Reference Forhan and Gill22]. Furthermore, poorer mobility is independently associated with poorer executive function [Reference McGough, Kelly, Logsdon, McCurry, Cochrane and Engel25], memory [Reference Kimura, Kazui, Kubo, Yoshida, Ishida and Miyoshi26], and global cognition [Reference Mun-San Kwan, Lin, Chen, Close and Lord27]. Thus, there is a growing interest in the development of a brief and easy-to- task for the assessment of functional mobility.
Rising from the floor, which can be assessed using the supine-to-standing (STS) test, is a basic functional mobility task that requires appropriate levels of muscle strength, joint coordination, balance, and flexibility. Previous studies have indicated that STS results exhibit a significant correlation with measures of typical functional mobility (e.g., gait speed and timed up-and-go (TUG) test results) and are a significant predictor of failing health and function in older adults [Reference Bergland and Laake28, Reference Klima, Anderson, Samrah, Patel, Chui and Newton29]. Because the STS test is less than 1 min in duration and requires no specialized equipment or training, it is relatively simpler than other tests for functional mobility [Reference Klima, Anderson, Samrah, Patel, Chui and Newton29, Reference Tiedemann, Shimada, Sherrington, Murray and Lord30]. In this respect, STS will be useful in measuring functional mobility in schizophrenia patients. However, to our knowledge, no studies have investigated STS test results in patients with schizophrenia, although a few studies [31–Reference Chwastiak, Rosenheck, McEvoy, Keefe, Swartz and Lieberman33] have explored the relationship between cognitive function and functional mobility in patients with schizophrenia.
As stated above, while several studies have examined the relationship between CRF or functional mobility and a single domain of cognition in patients with schizophrenia, no studies to date have investigated how individual physical performance domains (e.g., CRF, functional mobility) are associated with different cognitive domains in this patient population. Furthermore, there is a need to consider the association between cognitive function and disorder-related clinical symptoms in patients with schizophrenia. Both depression and psychotic symptoms appear to negatively affect memory and psychomotor speed [Reference Holmen, Egeland, Andersen, Bigseth and Engh18, Reference Douglas and Porter34]. In addition, health indicators such as total cholesterol levels and body mass index (BMI) may impact neurocognition in patients with schizophrenia [Reference Krakowski and Czobor35, Reference Rashid, Lim, Lam, Chong, Keefe and Lee36].
In the present study, we aimed to investigate the relationship between physical performance (i.e., CRF and functional mobility) and cognitive function (i.e., executive function and working memory) in individuals with schizophrenia. We hypothesized that poorer physical performance would be associated with lower cognitive function after adjusting for factors that are known to affect both physical performance and cognitive function in patients with schizophrenia.
Patients aged 18–60 with a primary diagnosis of schizophrenia or schizoaffective disorder were recruited via posters placed in local hospitals with psychiatric units and community mental health centers, and via advertisements in local newspapers. Some patients were also referred by other local health professionals.
Participants were eligible for the study if they had a diagnosis of schizophrenia or schizoaffective disorder according to the Diagnostic and Statistical Manual for Mental Disorders, 4th Edition, Text Revision (DSM-IV-TR), received antipsychotic treatment, and were at least 18 years old. Face-to-face diagnostic interviews were conducted by a psychiatric/mental health nurse and a board-certified psychiatrist who reviewed the history, symptoms, and psychosocial function of each patient using all available sources of information, in accordance with DSM-IV-TR criteria. All participants were required to be on a stable dose and dosing regimen of antipsychotics for at least 4 weeks prior to assessment. Participants were excluded if they had participated in any exercise program in the 3 months o before the start of the study or displayed evidence of significant cardiovascular, neuromuscular, endocrine, or other somatic/substance use disorders that would prevent safe participation.
A total of 65 participants were screened, five of whom failed to meet initial eligibility criteria.
The present cross-sectional study included 60 patients diagnosed with schizophrenia. All procedures were approved by the Investigational Review Board of Severance Mental Health Hospital. All participants provided written informed consent. Participants were required to perform all assessments between 10 a.m. and 4 p.m. on the same day, in the following order: psychiatric measures, cognitive tasks, and physical performance tasks. The time interval between the administration of assessments ranged from 15 min to 1 h.
2.3 Demographic information
Demographic characteristics (age, sex, years of education) and baseline clinical data (age at onset, duration of illness, number of hospital admissions, antipsychotic dosing (PDD:DDD ratio), total cholesterol, and BMI were collected.
2.3.1 Antipsychotic dosing
To calculate doses of antipsychotic drugs, the prescribed daily dose (PDD) in milligrams was divided by the defined daily dose (DDD) to yield a PDD:DDD ratio. DDD is defined as the assumed average maintenance dose per day for a drug used for its main indication in adults.
2.4 Clinical symptoms
Clinical symptoms were determined via interviews and self-report questionnaires.
2.4.1 Brief psychiatric rating scale (BPRS)
The BPRS is an 18-item, semi-structured introductory interview that is completed by the interviewer [Reference Overall and Gorham37]. Each item is rated on a scale from 0 to 6. Scores between 15 and 30 indicate minor symptoms, while those above 30 indicate major symptoms. Previous studies have demonstrated the reliability of the BPRS in Korean patients [Reference Kim, Lee and Jeon38].
2.4.2 Beck depression inventory (BDI)
The BDI was used to assess depressive symptoms [Reference Beck, Steer and Carbin39]. The BDI is a 21-item self-report inventory that asks participants to choose the one statement that best describes their feelings during the past 2 weeks. The Korean version of the BDI has been identified as both valid and reliable [Reference Lee40].
2.4.3 State-trait anxiety inventory (STAI)
The STAI is a self-rated instrument that contains two 20-item subscales for the measurement of anxiety [Reference Spielberger41]. One scale measures state anxiety, while the other measures trait anxiety. A standardized Korean version of the STAI has been developed by Han and colleagues [Reference Han, Lee and Tack42].
2.5 Neurocognitive function
2.5.1 Stroop word-color test
The Stroop Word-Color Test was used to assess components of executive function representing a person’s ability to deal with conﬂicting stimuli [Reference Stroop43]. This test involves pairs of conﬂicting stimuli that are presented simultaneously (i.e., the name of one color printed in another color). In the congruous condition, participants are required to read the names of colors printed in black ink (W) and to name diﬀ ;erent color patches (C). Conversely, in the color-word (CW) condition, color-words are printed in an inconsistent color ink. Thus, in this incongruent condition, participants are required to name the color of the ink instead of reading the word.
2.5.2 Sternberg working memory task (SWM)
The classic SWM task [Reference Sternberg44] was administered using Inquisit 188.8.131.52 (Millisecond Software, LLC Seattle, WA). Each trial of the task consisted of a set of two to five white digits presented in a sequence (1200 ms each). A yellow probe digit appeared 2500 ms after the last digit (maintenance period), at which point participants were required to press an appropriate button indicating whether it had been present in the previously displayed sequence. Participants were provided with visual feedback regarding the accuracy of their responses. Task sessions were divided into equally distributed “in” (probe present in the memory sequence) and “out” (probe not present in the memory sequence) trials (120 trials in total, preceded by 15 training trials).
2.6 Physical performance
2.6.1 Step test
Measuring post-exercise recovery heart rate (HR) after a step test is a common method for determining cardiorespiratory (CR) fitness. The present study used the 3-minute YMCA step test (American College of Sports Medicine, 2006), which involves stepping at a rate of 24 steps/minute for 3 min. Following exercise, HR was counted for 1 min with the participant in the seated position. Low post-exercise recovery HR is considered indicative of good CRF.
2.6.2 STS test
The STS test is used to assess a person's ability to transition from a supine position to a standing position. Previous studies have utilized the STS test to evaluate functional mobility [Reference Schenkman, Ellis, Christiansen, Barón, Tickle-Degnen and Hall45] and physical performance [Reference Klima, Anderson, Samrah, Patel, Chui and Newton29] in older adults and patients with neurodegenerative conditions. Beginning from a mat on the floor, participants were instructed to rise from the supine position at a comfortable rate, using any motions they desired. The administrator measures the amount of time it takes to rise from the supine position. The STS test has been associated with excellent test-retest reliability (intraclass correlation coefficient = 0.9) [Reference Alexander, Galecki, Nyquist, Hofmeyer, Grunawalt and Grenier46].
2.7 Statistical analysis
We first analyzed the demographic and clinical characteristics of participants using descriptive statistics. Multiple regression analyses were then used to adjust for relevant covariates and identify predictive factors associated with neurocognitive function. After performing univariate analyses to identify the associated factors, multivariate analyses were conducted using Stroop Word-Color test and SWM task results as the dependent variables. Analyses were performed using SPSS version 18.0 (SPSS, Chicago, NJ, USA). The level of statistical significance was set at p < 0.05.
3.1 Descriptive statistics
A total of 60 participants were included in the analysis. Demographic information, clinical characteristics, physical performance, and cognitive function results are presented in Table 1.
3.2 Correlation analysis
Table 2 presents the associations between neurocognitive function and the following factors: demographic information, clinical characteristics, and physical performance. Stroop CW response latency (RL) was positively correlated with age (p = 0.005), age at onset (p = 0.047), and step test results (index) (p = 0.009). In addition, Stroop CW correct rate (CR) was negatively correlated with age (p = 0.004), duration of illness (p < 0.001), number of hospital admissions (p = 0.024), and STS test results (p < 0.001). SWM RL was positively correlated with age (p = 0.006), duration of illness (p = 0.03), BDI (p = 0.001) results, and step test results (index) (p = 0.026). Our findings also indicated that SWM CR was inversely correlated with age (p = 0.027) and STS test results (p = 0.007).
The results of additional analyses based on Global Assessment of Functioning (GAF) scores are shown in STable 1,2, and 3. In the high-functioning group, we observed significant positive correlations between the step text index and response latency in the Stroop CW and SWM tasks. STS test time was also correlated with SWM CR in this group. In the low-functioning group, STS test results were strongly correlated with Stoop task CR.
DDD, defined daily dose; PDD, prescribed daily dose; BMI, Body mass index; BPRS, Brief Psychiatric Rating Scale; BDI, Beck Depression Inventory; SAI, State Anxiety Inventory; TAI, Trait Anxiety Inventory; CW, Color-word condition; RL, response latency; CR, correct rate; WM, working memory task; STS, supine-to-standing.
CW, color-word condition; RL, response latency; CR, correct rate; WM, working memory task; DDD, defined daily dose; PDD, prescribed daily dose; BMI, body mass index; BPRS, Brief Psychiatric Rating Scale; BDI, Beck Depression Inventory; SAI, State Anxiety Inventory; TAI, Trait Anxiety Inventory; STS, supine-to-standing.
Pearson’s correlation coefficients:
** Correlation is significant at the 0.01 level (2-tailed).
* Correlation is significant at the 0.05 level (2-tailed).
3.3 Multiple regression analysis
Multivariable linear regression was applied to determine which factors were associated with Stroop CW test and SWM test results (Table 3). Signiﬁcantly associated variables by univariate analyses above were included as independent variables. Our analysis revealed that age at onset, total cholesterol, and step test results (index) were significantly associated with Stroop CW RL, while the duration of illness and STS time (s) were significantly associated with Stroop CW CR. Step test results (index) were the only significant predictor of RL in the SWM task, while STS time (s) was the only significant predictor of SWM CR. Step test results were positively associated with both Stroop CW RL and SWM RL, while STS test results were negatively associated with both Stroop CW CR and SWM CR. However, clinical symptoms were not associated with any variables in Stroop or SWM task.
In the present study, we aimed to investigate the associations between physical performance and neurocognition among patients with schizophrenia. Our findings indicated that poorer functional mobility as determined by the STS test was associated with lower accuracy on executive function and working memory tasks after adjusting for other clinical correlates. Furthermore, poorer CRF as determined by the step test was associated with slower performance on these two cognitive tasks. In addition, total cholesterol levels and the duration of illness were significantly associated with latency and accuracy on the executive function task, respectively.
Recent studies involving the general population and older adults have demonstrated that functional mobility is associated with cognitive function, including executive function and memory [Reference McGough, Kelly, Logsdon, McCurry, Cochrane and Engel25, Reference Kimura, Kazui, Kubo, Yoshida, Ishida and Miyoshi26, Reference Demnitz, Zsoldos, Mahmood, Mackay, Kivimäki and Singh-Manoux47, Reference Donoghue, Horgan, Savva, Cronin, O’regan and Kenny48]. Although there is also evidence to support this relationship in patients with schizophrenia, previous studies are limited in that they included particular sub-groups of patients only (e.g., older adults, early schizophrenia) [Reference Leutwyler, Hubbard, Jeste, Miller and Vinogradov31, Reference Lallart, Jouvent, Herrmann, Beauchet and Allali32] or utilized subjective measures only [Reference Chwastiak, Rosenheck, McEvoy, Keefe, Swartz and Lieberman33]. However, our study demonstrated that, when related variables including age and clinical symptoms are considered, the objective STS is the most explanatory factor for cognitive domains such as executive function and memory. Although such findings suggest a close relationship between mobility and cognition in patients with schizophrenia, follow-up studies are required to verify our results.
Previous research has indicated that the STS is useful for evaluating functional mobility and ADL performance in older adults and patients with Parkinson’s disease [Reference Schenkman, Ellis, Christiansen, Barón, Tickle-Degnen and Hall45, Reference Alexander, Galecki, Nyquist, Hofmeyer, Grunawalt and Grenier46], as well as those undergoing neurological rehabilitation [Reference Schenkman, Cutson, Kuchibhatla, Chandler, Pieper and Ray49]. Similarly, the STS may be useful in patients with schizophrenia due to the impact of sedentary behavior and low levels of physical activity in this population [Reference Vancampfort, Firth, Schuch, Rosenbaum, Mugisha and Hallgren50]. The STS is advantageous in that it is more efficient than other functional mobility tasks, requiring less time and space than the TUG test or measurements of gait speed. In addition, the STS test is easy to understand and requires no specific training for evaluators, making it ideal for patients with schizophrenia. Taken together, the accumulated evidence suggests that the STS represents a quick, easy method for assessing functional mobility in patients with schizophrenia in both clinical and practical settings.
Note: Bold values are statistically significant.
CW, color-word condition; RL, response Latency; CR, correct rate; WM, working memory task; BDI, Beck Depression Inventory; STS, supine-to-standing.
Given the complexity of functional mobility, it is not surprising that multiple cognitive functions are associated with performance in this domain. Among the multiple brain regions associated with mobility, the cuneus not only controls limb movement, but also processes and memorizes environmental information [Reference Beloozerova and Sirota51]. The prefrontal cortex is responsible for executive function and is known to be active during the planning and initiation of a series of actions [Reference Frith, Friston, Liddle and Frackowiak52]. Previous studies have reported abnormalities in neural synchrony in the cuneus of patients with schizophrenia [Reference Hoptman, Zuo, Butler, Javitt, D’Angelo and Mauro53], along with hypofrontality [Reference Davidson and Heinrichs54]. Thus, it is possible that dysfunction of this region exerts a negative impact on the series of processes involved in rising from the floor. However, the cause-effect relationship between mobility and neurocognition remains to be elucidated.
In accordance with our findings, previous studies have demonstrated an association between CRF and executive function/memory in patients with schizophrenia [17–Reference Firth, Stubbs, Rosenbaum, Vancampfort, Malchow and Schuch21]. Recently evidence also indicates that adult neurogenesis may be reduced in patients with schizophrenia, which in turn may contribute to impaired cortical-to-hippocampal connectivity [Reference Reif, Fritzen, Finger, Strobel, Lauer and Schmitt55]. However, some investigators have found that improving CRF through aerobic exercise increases levels of neural growth factors (e.g., brain-derived neurotrophic factor (BDNF)) in patients with schizophrenia, helping to support the survival of existing neurons and encouraging the growth and differentiation of new neurons and synapses [Reference Kimhy, Vakhrusheva, Bartels, Armstrong, Ballon and Khan19, 18–Reference Firth, Stubbs, Rosenbaum, Vancampfort, Malchow and Schuch21, Reference Vakhrusheva, Marino, Stroup and Kimhy56]. Aerobic exercise promotes neurogenesis and synaptic plasticity, exerting positive effects on neurocognition in schizophrenia. The resulting improvements in CRF lead to increases in the number of synapses in the frontal and parietal gray matter [Reference Colcombe, Kramer, Erickson, Scalf, McAuley and Cohen11], which play a role in resolving the response competition elicited by incongruence response cues (e.g., Stroop task). In addition, a study by Pajonk and colleagues [Reference Pajonk, Wobrock, Gruber, Scherk, Berner and Kaizl20] reported that improving CRF increases levels of metabolic and synaptic plasticity-related proteins in the hippocampus, which may contribute to improved cognition, particularly with regard to short-term memory.
Interestingly, higher total cholesterol levels were associated with better performance on cognitive tasks after controlling for other clinical symptoms. Similarly, Elias and colleagues [Reference Elias, Elias, D’agostino, Sullivan and Wolf57] observed that total cholesterol levels are positively correlated with cognitive function in healthy adults. Some researchers have speculated that serum cholesterol concentrations play a role in synthesizing neurotransmitters and forming neuronal membranes [Reference Hibbeln, Umhau, George, Shoaf, Linnoila and Salem58]. Plasma cholesterol also enhances cognitive function by modifying brain membrane fatty acid composition and altering physiological properties to activate membrane neurotransmitter receptors [Reference Burger, Gimpl and Fahrenholz59]. Krakowski and Czobor [Reference Krakowski and Czobor35] revealed that increases in cholesterol levels were associated with improved cognition in patients with schizophrenia receiving antipsychotics. When taken with these previous results, our findings suggest that physical health factors, including cholesterol levels and physical performance, are closely related to cognitive function in patients with schizophrenia.
In the present study, longer duration of illness was associated with poorer performance on the Stroop task. It is well established that longer durations of illness among patients with schizophrenia are associated with greater impairments in higher-order cognitive functions such as attention and executive function. For example, using a battery of standardized neuropsychological tests, Smet and colleagues [Reference Smet, Tandon, Goldman, Decker, Lelli and Berent60] reported that patients with a longer duration of illness exhibited significantly greater impairments in executive function as determined using the Trail-Making Test and Wisconsin Card Sort Test. In addition, in a study by Albus and colleagues [Reference Albus, Hubmann, Ehrenberg, Forcht, Mohr and Sobizack61], patients with chronic schizophrenia exhibited poorer performance on measures of visuomotor processing, attention, and executive function (abstraction/flexibility) when compared to patients with first-episode schizophrenia.
The present study demonstrated robust associations between physical performance and neurocognition. While previous studies focused on a single component (e.g., CRF), we also considered functional mobility, which is closely related to the functional outcomes of schizophrenia as well as CRF. In addition, our findings indicated that total cholesterol level is significantly associated with cognitive function. Moreover, unlike previous studies, we also controlled for the effects of other disorder-related clinical correlates (e.g., psychotic symptoms, depression). Our findings suggest that addressing physical health factors may improve cognitive function in patients with schizophrenia. Because direct treatment of cognitive dysfunction remains difficult [Reference Holmen, Egeland, Andersen, Bigseth and Engh18], the clinical value of our findings cannot be understated, as we have identified modifiable factors that may impact neurocognition. While antipsychotic medications can improve psychotic symptoms, their impact on cognitive dysfunction is limited [Reference Hasan, Falkai, Wobrock, Lieberman, Glenthoj and Gattaz62]. Furthermore, non-pharmacological interventions such as cognitive remediation therapy exert only minor effects, which are eventually lost over time [Reference Wykes, Huddy, Cellard, McGurk and Czobor63]. In the future, exercise intervention may help to improve cognitive symptoms in patients with schizophrenia, thereby leading to improvements in function and prognosis [Reference Sommer and Kahn64].
Given several limitations of the present study, our findings should be interpreted with caution. First, although the STS test has been used to evaluate functional mobility in other clinical populations, the test has not yet been standardized for the assessment of patients with schizophrenia. Subsequent studies should aim to replicate our findings using this tool. Second, the structured diagnostic tools were not used, but diagnostic verification was double-checked by two psychiatric practitioners. Third, although the time of day at which neurocognitive tests are administered may influence performance among patients with schizophrenia [Reference D’Reaux, Neumann and Rhymer65], cognitive tasks were not administered at the exact same time in the present study. Fourth, the generalizability of our findings may be limited due to a small sample size and the inclusion of patients with chronic schizophrenia with a rather long duration of illness. Finally, as the study was cross-sectional, no causal correlations could be established. Interaction effects may exist, as better cognitive abilities may provide better opportunities for a healthy lifestyle. Further studies are also required to investigate the mechanisms underlying schizophrenia symptoms.
The results of the present study demonstrate the relationship between physical performance and neurocognition in patients with schizophrenia. Functional mobility and CRF were correlated with accuracy and RT during cognitive tasks including the Stoop and Sternberg task, respectively. In addition, total cholesterol levels were positively associated with Stroop task accuracy, after controlling for clinical symptoms (e.g., psychotic symptoms). These findings suggest that physical performance is closely associated with neurocognition among patients with schizophrenia. Given that these factors are modifiable, exercise intervention may help to improve cognitive symptoms in patients with schizophrenia, thereby leading to improvements in function and prognosis. Subsequent studies should aim to clarify the mechanisms underlying the association between physical health and neurocognition in individuals with schizophrenia.
This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (NRF-2018R1C1B6005103) and Hallym University Research Fund.
Roh designed the study and structured this manuscript. Ryu wrote the protocol of exercise programs. Jung, Kim CH and Kim DH managed the literature searches and analyses (including the statistical analysis). Lee HB and Lee SK contributed to the interpretation of the results. Shin and Kim J wrote the first draft of the manuscript. Shin and Kim J managed and contributed equally the entire this study process. All authors contributed to and have approved the final manuscript.
Declaration of Competing Interest
The authors declare that they have no competing interests.
Appendix A Supplementary data
Supplementary material related to this article can be found, in the online version, at doi: https://doi.org/10.1016/j.eurpsy.2019.06.010.