Functional Relationship between Inhibitory Control, Cognitive Flexibility, Psychomotor Speed and Obesity

In the last decades, it has been proposed that executive functions may be particularly vulnerable to weight-related issues. However, evidence on the matter is mixed, especially when the effects of sociodemographic variables are weighted. Thus, the current study aimed at further examining the relationship between executive functions and obesity. To this aim, we compared treatment-seeking overweight, obese, and morbidly obese patients with normal-weight control participants. We examined general executive functioning (Frontal Assessment Battery–15) and different executive subdomains (e.g., inhibitory control, verbal fluency, and psychomotor speed) in a clinical sample including 208 outpatients with different degrees of BMI (52 overweight, BMI 25–30, M age = 34.38; 76 obese, BMI 30–40, M age = 38.00; 80 morbidly obese, BMI > 40, M age = 36.20). Ninety-six normal-weight subjects served as controls. No difference on executive scores was detected when obese patients were compared with over- or normal-weight subjects. Morbidly obese patients reported lower performance on executive scores than obese, overweight, and normal-weight subjects. Between-group difference emerged also when relevant covariates were taken into account. Our results support the view that morbid obesity is associated with lower executive performance, also considering the critical role exerted by sociodemographic (i.e., sex, age, and education) variables. Our results support the view that executive functioning should be accounted into the management of the obese patient because of non-negligible clinical relevance in diagnostic, therapeutic, and prognostic terms.


Introduction
Obesity represents a strong concern for public health [1] since it is linked to decreased quality of life and worse prognosis for concomitant chronic diseases [2]. Moreover, obesity is a major risk factor for several morbidities [3][4][5][6]. Recently, scientific research has looked beyond the physical consequences of obesity and sought to explore the association between obesity and cognitive processes [7]. Indeed, it appears that obesity predicts adverse neurocognitive outcomes [8]. More specifically, obesity would increase the risk of cognitive defects, besides being a potential predictor of cognitive decline and mild/major neurocognitive disorders [9][10][11][12][13][14][15][16][17][18][19].
The first generation of studies investigating the relationship between obesity and cognition found that obese patients performed worse on tasks assessing global cognitive between obesity and executive performance. In particular, we hypothesize that BMI-ranked groups, i.e., from normal-weight to morbid obesity, show different performances in the executive tasks employed.

Participants
Three hundred and four subjects agreed to participate in this study. Eighty were morbidly obese patients ( 10), and ninety-six were normal-weight subjects that served as controls (fifty-two females, M age = 30.00, SD = 6.76; M education = 12.58, SD = 1.39; M BMI = 23.74, SD = 1.55). Patients were recruited at the Department of Experimental Medicine (outpatient clinic of dietetics, sports medicine, and psychophysical health) of the University of Campania "Luigi Vanvitelli" (Italy). Inclusion criteria were age ≥18 years and formal education ≥5 years. Furthermore, all patients (from overweight to morbidly obese individuals) underwent at least the first outpatient examination, and they had been prescribed a low-calorie diet aimed at weight loss. Exclusion criteria were: BMI < 18.5 (i.e., underweight), previous or current history of intellectual and/or linguistic deficits, previous or current history of psychopathological, psychiatric or neurocognitive disorders (including non-progressive or metabolic dementia), history of alcohol or substance abuse/addiction, or ongoing pharmacological treatment with drugs interfering with cognitive abilities.

Materials and Procedure
For each participant, we collected demographic information, i.e., sex, age and education (years of education were recorded in accordance with the Italian schooling system: 5 years = primary school; 8 years = secondary school; 13 years = high school; 16 years = bachelor's degree; 18 years = master's degree; 20 years = graduate school, PhD), and anthropometric measures (i.e., BMI). Therefore, participants were administered a neuropsychological battery including a screening test of general executive functioning (i.e., the Frontal Assessment Battery-15), and three executive tasks assessing impulsivity/inhibitory control (i.e., Stroop Color-Word Test), cognitive flexibility (i.e., FAS verbal fluency test), and psychomotor speed (i.e., Digit Symbol Substitution Test), respectively. Finally, the Cumulative Illness Rating Scale was used for computing the number of concurrent diseases and related severity.
Frontal Assessment Battery-15 (FAB15) [69]. The FAB15 is a short cognitive screening battery exploring general executive functioning (abstraction, generativity, cognitive flexibility, planning, sensitivity to interference, and inhibitory control). The scoring range is 0-15 and a higher score reflects better performance.
Stroop Color-Word Test (SCWT) [70]. This task assesses the ability to inhibit interference from a dominant response tendency. Moreover, it appears to tap other cognitive domains such as attention, processing speed, cognitive flexibility, and working memory [71,72]. The interference effect on both execution time (Stroop-T) and errors (Stroop-E) served as dependent variables.
FAS Verbal Fluency Test (FAS) [59,73,74]. It is a test assessing phonemic verbal fluency, which requires participants to produce as many words as possible that begin with letters "F", "A" and "S" (omitting personal proper nouns, surnames, and place names). The total number of produced words represents the dependent variable. The FAS fluency is considered a valid measure of selective attention, set-shifting, generativity, and selfmonitoring abilities.
Digit Symbol Substitution Test (DSST) [75,76]. The DSST is a pencil-and-paper test for assessing psychomotor speed; however, it is widely used to explore alternative domains such as processing speed, set-shifting, working memory, and associative and implicit learning. More specifically, participants are presented with a grid of numbers and matching symbols under which there is a test section with numbers and empty boxes. The task consists of filling as many empty boxes as possible with the appropriate symbol. The number of correct number-symbol matches completed in 90 sec is scored and entered as dependent variable.
Cumulative Illness Rating Scale (CIRS) [77]. The CIRS consists of fourteen categories related to different body systems and scores the severity of each condition on a five-point (0-4) Likert scale. Comorbidity and severity indexes entered as dependent variables.

Statistical Analyses
The assumptions of the Generalized Linear Model (GLM) were verified by checking univariate and multivariate normality. For descriptive purposes, we ran parametric (univariate Analysis of Variance, ANOVA) and non-parametric analyses (two-way chi-squared test, χ 2 ), when needed. In addition, correlation analysis (Spearman's correlation, ρ s ) was used to quantify the relationships between the variables under examination. The effects of sex, age, education, and morbidity (e.g., number of illnesses and the respective severity) were taken into account. Indeed, multivariate Analysis of Covariance (MANCOVA) was performed, with the five executive scores entering the model as dependent variables, the BMI subgroups (normal-weight, overweight, obese, and morbidly obese) as fixed factors, and sex, age, education, CIRS-morbidity, and CIRS-severity scores as covariates. Any posthoc analysis was performed according to Bonferroni's correction for multiple comparisons (p bonf ). Eta squared (η 2 ) or partial eta squared (η 2 p ) were used to quantify the effect sizes. Statistical analyses were conducted by using IBM SPSS v. 26 and JASP packages.
In the overweight group, females were younger as compared with males [F (1, 50) = 6.16, p = 0.02, η 2 = 0.11; female, M age = 30.00 years, SD = 6.71 vs. male, M age = 37.12 years, SD = 1.67]. Instead, no difference was found on the education variable (p = 0.99). As concerns performance on executive tasks, males got higher scores than females on the

Multivariate Analysis of Covariance
Results of MANCOVA (Table 3)     Pairwise comparisons based on Bonferroni's correction showed no significant difference between obese patients and normal-and over-weight subjects in any of the executive scores (p bonf range = 0.08-0.99). As concerns patients with morbid obesity, they got lower

Discussion
The current study was designed to compare the performance of morbidly obese, obese, overweight, and normal-weight participants on neuropsychological tasks exploring general executive functioning and specific (i.e., inhibitory control, verbal fluency, and psychomotor speed) executive subdomains. Our study encompassed the simultaneous contribution of sociodemographic variables (sex, age, and formal schooling) and comorbidity/severity indexes as covariates in a GLM.
In a scientific context in which the link between EFs and obesity needs to be further investigated [7,25,[51][52][53][54]81], our findings showed that morbidly obese individuals got poor executive scores regardless of the significant effects of the above-mentioned covariates. Notably, in line with previous studies [50], we found no significant differences between obese, overweight, and normal-weight subjects on the executive domains explored. This evidence suggests that severe obesity, more than obesity, may represent a clinical condition that impacts on cognition. Although a few contributions are available on this topic [34,65,66], their results converge on the idea that morbidly obese patients, net of the presence of eating disorders, would show weaknesses in tasks assessing motor planning, verbal fluency, processing speed, attention/vigilance, and inhibition.
Our results showed that morbidly obese patients got lower FAB15 scores (i.e., lower general executive functioning) than normal-weight, overweight, and obese participants. Therefore, the FAB15 may be sensitive enough to detect executive blunting in morbid obesity and to differentiate morbidly obese from obese patients in terms of executive functioning. Morbidly obese patients scored worse on FAS, DSST and Stroop-T as compared with normal-weight and overweight subjects, although they got similar performance than obese patients.
As concerns the executive subdomains explored, poor performance on the FAS test underlies low self-monitoring/set-shifting capabilities. Reduced psychomotor speed (i.e., DSST) highlights some difficulties in detecting and responding appropriately to rapid changes in the environment. Finally, longer latency time in the Stroop Test suggests inadequate inhibitory control and thus decreased ability to voluntarily suppress interfering information or prepotent habitual responses.
From an anatomofunctional standpoint, the integrity of these domains seems to depend on a large frontal network including (i) anterior and posterior cingulate cortices, inferior frontal gyrus and medial frontal areas for verbal fluency, (ii) dorsolateral, ventromedial, orbital prefrontal and anterior cingulate cortices for inhibitory control, and (iii) the frontoparietal network, especially the middle frontal gyrus and the posterior parietal cortex, for psychomotor speed. Great interest is held by the functional connectivity within the PFC, as well as between PFC and subcortical structures (e.g., basal ganglia, subthalamic nucleus, hippocampal formation) in patients with and without eating disorders [82][83][84][85][86][87][88]. The projections from the PFC towards hypothalamic circuits appear to be involved in the modulation of hunger and satiety signals, whereas the striatal and ventral midbrain circuits are relevant for reward processing [83,89,90]. In morbidly obese patients, likely also due to the more complicated clinical profile (e.g., metabolic syndrome) [91][92][93], abnormal activity within these neural circuits might affect EFs and lead to difficulties in planning regular eating patterns, inability to delay gratification, or inhibit prepotent responses to highly palatable foods [53].
To better understand the relationship between obesity and cognition, there is a need to examine the interaction between obesity and brain physiology. Many studies have found that obesity in middle age correlated with an increased risk of cognitive decline in later life [94]. Indeed, excessive body fat seems to be associated with reduced brain volume in cognitively healthy older adults, but in patients with cognitive deficits, it exerts an additional detrimental effect [95]. More specifically, greater body adiposity and elevated waist-hip ratio were found to be related with generalized structural alterations involving orbital frontal cortex, temporal and parietal cortices, and hippocampus [96][97][98]. In a structural brain mapping cohort study of patients with Mild Cognitive Impairment and Alzheimer's disease (AD), a higher BMI value was found to be correlated with brain volume reduction in the frontal, temporal, and parietal cortices (0.5-1.5%) per one unit increase in BMI [44]. In line with this vein, longitudinal studies have highlighted a significant increase in temporal lobe atrophy between 13 and 16% for each one-unit increment in BMI [45]. Still, brain scanning evidence has suggested that obese individuals showed lower gray matter density in the middle frontal gyrus, primary somatosensory cortex, and putamen than normal-weight controls [99,100].
Although the mechanisms underlying the relationship between obesity and cognitive functions need to be further investigated, physiology research has identified a reliable mechanism of action in the activity of the neuro-immune-endocrine systems. Adipose tissue fulfills neuro-immune-endocrine functions, participating in the homeostasis [101]. In obesity, adipose tissue produces cytokines such as interleukin (specifically IL1ß and IL6), interferon γ (IFNγ), Tumor Necrosis Factor α (TNFα), and Monocyte Chemoattractant Protein 1 (MCP1), which promote chronic inflammation [102][103][104][105][106][107]. Hypertrophy of adipocytes induces a state of chronic inflammation that causes changes in the brain by inducing neuroinflammation. For instance, some studies have shown that inflammatory processes, especially in the hypothalamus, often follow high-fat feeding [15,108,109]. Moreover, chronic low-grade inflammation alters the blood-brain barrier (BBB) due to endothelial dysfunction, generating neuroinflammation and increasing oxidative stress, often resulting in cognitive impairment [110,111]. Since the BBB interfaces the periphery with the central nervous system, it could represent the bridge between peripheral inflammation and obesity-related differences in cognitive processes. Studies on obese animal models and human obese patients have shown a relationship between diet and cognitive functioning, especially in working memory and learning [110][111][112].
Furthermore, systemic inflammation could predict cognitive decline and dementia [113][114][115][116]. Adipose tissue may also affect β-amyloid metabolism [117,118]. Indeed, some neuropathological features of AD (e.g., amyloid plaques, neurofibrillary tangles) appear to be most observed in obese compared with normal-weight older adults [119]. In addition, a higher concentration of amyloid plaques has been detected in the hippocampal formation of cognitive intact morbidly obese patients when compared with non-obese subjects [120,121]. A high-fat diet may increase either body weight or cognitive impairment [122]. Longitudinal data have shown that higher caloric intake was related to an increased risk of AD at 6.3 years follow-up [123]. In addition, high-fatty acids diets, or diets rich in simple sugars, may undermine brain physiology, harming the integrity of the bloodbrain barrier (BBB) [124], protecting the central nervous system from bloodborne toxins. AD and vascular dementia are linked to BBB dysfunction [125] and longitudinal studies sustained a significant relationship between midlife obesity and lower BBB functioning [45].
This study supports the hypothesis that patients with morbid obesity show inefficiencies in executive functioning, including when the contributions of sociodemographic factors and comorbidity indexes are taken into account. The main limitation of the present study is that we used BMI as the sole indicator of obesity, although previous research has shown that it is not a good indicator of obesity, besides that its relationship to EFs is notoriously mixed. BMI is unable to discriminate between muscle and adipose tissues and cannot assess regional adiposity [126]. Conversely, waist circumference or waist-to-hip ratio (WHR) could be superior to BMI as measures of obesity since they showed higher correlations with obesity-related risk factors [127][128][129]. Furthermore, robust evidence is available on the predictive role of WHR for executive performance in obese patients [49]. An additional limitation, in terms of external validity, is that we tested treatment-seeking overweight, obese, and morbidly obese patients; therefore, our results are not generalizable to patients not requiring a dietary intervention. Finally, one should consider that, although some cognitive tasks assessing EFs are based on a verbal response set (e.g., verbal fluency, Stroop test), most executive tests require hand motor responses to be properly performed (e.g., the DSST) [130,131]. Thus, it could be interesting to test the mediation effects exerted by hand motor function in the relationship between EFs and obesity.

Conclusions
Our results support the relationship between severe/morbid obesity and poor EFs, also considering the contribution of sociodemographic variables and the severity of con-current diseases. Specifically, results from the present study suggest that: (1) patients with morbid obesity demonstrate decreased general executive functioning compared with normal-weight, overweight, and obese participants; (2) morbidly obese patients score worse than normal-weight and overweight subjects on cognitive tasks exploring selective executive subdomains, namely, cognitive flexibility, psychomotor speed, and inhibition processes; (3) EFs need to be accounted into the management of obese patients due to the non-negligible clinical relevance in diagnostic, therapeutic, and prognostic terms; (4) the direction of the relationship between obesity and executive defects needs to be further investigated. The study of this matter remains a strong challenge for future research.

Institutional Review Board Statement:
The study was conducted in accordance with the Declaration of Helsinki and approved by the Ethics Committee of the University of Campania "Luigi Vanvitelli".
Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.

Conflicts of Interest:
The authors declare no conflict of interest.