Physical activity is unrelated to cognitive performance in pre-bariatric surgery patients Svenja Langenberg, Mareike Schulze, Merle Bartsch, Kerstin GrunerLabitzke, Christian Pek, Hinrich K¨ohler, Ross D. Crosby, Michael Marschollek, Martina de Zwaan, Astrid M¨uller PII: DOI: Reference:

S0022-3999(15)00079-3 doi: 10.1016/j.jpsychores.2015.03.008 PSR 8993

To appear in:

Journal of Psychosomatic Research

Received date: Revised date: Accepted date:

8 January 2015 9 March 2015 11 March 2015

Please cite this article as: Langenberg Svenja, Schulze Mareike, Bartsch Merle, GrunerLabitzke Kerstin, Pek Christian, K¨ohler Hinrich, Crosby Ross D., Marschollek Michael, de Zwaan Martina, M¨ uller Astrid, Physical activity is unrelated to cognitive performance in pre-bariatric surgery patients, Journal of Psychosomatic Research (2015), doi: 10.1016/j.jpsychores.2015.03.008

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

Physical activity and cognition

PT

Short title: Physical activity and cognition

SC RI

Physical activity is unrelated to cognitive performance in pre-bariatric surgery patients

Svenja Langenberg1, Mareike Schulze2, Merle Bartsch1, Kerstin Gruner-Labitzke3, Christian Pek2,

1

MA

NU

Hinrich Köhler3, Ross D. Crosby4, Michael Marschollek2, Martina de Zwaan1, Astrid Müller1*

Department of Psychosomatic Medicine and Psychotherapy, Hannover Medical School,

Peter L. Reichertz Institute for Medical Informatics, University of Braunschweig - Institute

PT

2

ED

Hannover, Germany

of Technology and Hannover Medical School, Hannover, Germany Department of Surgery, Herzogin Elisabeth Hospital, Braunschweig, Germany

4

Neuropsychiatric Research Institute and Department of Psychiatry and Behavioral Science,

AC

CE

3

University of North Dakota School of Medicine and Health Sciences, Fargo, ND, USA

*Correspondence address: Astrid Müller, Department of Psychosomatic Medicine and Psychotherapy, Hannover Medical School, Carl-Neuberg-Str. 1, D-30625 Hannover, Germany. Tel.: +49 511 532 9179, Fax: +49 511 532 3190, E-mail: [email protected]

1

ACCEPTED MANUSCRIPT

Physical activity and cognition

ABSTRACT Objective: To investigate the relationship between physical activity (PA) and cognitive

PT

performance in extreme obesity. Methods: Seventy-one bariatric surgery candidates (77.5% women) with a mean body mass index (BMI) of 46.9 kg/m2 (SD = 6.0) and a mean age of

SC RI

41.4 (SD = 11.9) years completed SenseWear Pro2 activity monitoring for seven days. Cognitive functioning was assessed by a computerized test battery including tasks of executive function (Iowa Gambling Task), visuospatial short-term memory (Corsi Block

NU

Tapping Test) and verbal short-term memory (Auditory-Verbal Learning Test).

MA

Questionnaires assessing eating disturbances and depressive symptoms were administered. Somatic comorbidities were assessed by medical chart review. Results: The level of PA was low with mean steps per day within wear time being 7140 (SD = 3422). Most patients were

ED

categorized as sedentary (31.0%) or low active (26.8%). No significant association between

PT

PA estimates and cognitive performance was found. Lower PA was modestly correlated with higher BMI but not with age, somatic comorbidity or depressive symptoms. Moderated

CE

regression analyses suggested a significant interaction effect between depression and PA in

AC

predicting performance on the Corsi Block Tapping Test. Patients with (29.6%) and without (70.4%) regular binge eating did not differ with respect to PA or cognitive function. Conclusion: The findings indicate no association between daily PA and cognitive performance in morbidly obese patients. Future studies should explore the relationship between the variables with regard to dose-response-questions, a broader BMI range and with respect to potential changes after substantial weight loss due to bariatric surgery.

Keywords: accelerometry, bariatric surgery, cognitive function, obesity, physical activity

2

ACCEPTED MANUSCRIPT

Physical activity and cognition

INTRODUCTION

PT

Past research provides evidence for an association between obesity and impaired cognitive performance, particularly in the domains of memory, learning and executive

SC RI

functioning (1-4). These deficits are particularly common in individuals with morbid obesity (BMI ≥ 40 kg/m2) seeking bariatric surgery (5,6) and have been attributed to the effects of somatic conditions that are common in this group including hypertension, type 2 diabetes,

NU

sleep apnea, pain disorder and dyslipidemia (e.g., 7-14). Also, depression that is known to be

MA

prevalent in morbidly obese patients is thought to negatively affect cognitive function (14,15). Considering findings from non-obese individuals, low physical activity (PA) is likely to contribute to cognitive impairment (16). Bed rest studies clearly demonstrate the negative

ED

effect of prolonged physical inactivity on executive functioning in healthy males (17).

PT

Conversely, numerous studies indicate a positive effect of aerobic exercise on cognition (1822); however, the mechanisms of exercise-related benefits remain unclear. Aerobic fitness has

CE

been linked to better cognitive function via improvement in cerebral blood flow and oxygen

AC

delivery (23,24). Neuroimaging findings suggest a positive association between aerobic exercise and cerebral white matter integrity (25) or between exercise and better cardiovascular fitness and higher volume of specific brain areas important for learning and memory (26). Prior research has reported a mostly sedentary lifestyle and low PA in individuals with obesity (27-31) which is possibly connected with alterations in cognitive domains potentially contributing to an unhealthy or even disturbed eating behavior (16). As suggested recently by Galioto et al. (2013) (32), executive dysfunction could negatively affect adherence to postoperative guidelines and lifestyle changes. Specifically, high disinhibition, organizational and planning problems as well as memory deficits in bariatric surgery candidates may be reflected in common behavioral factors that hinder effective weight loss, e.g. high susceptibility to alimental stimuli, poor control of eating behavior (33,34), missing adherence 3

ACCEPTED MANUSCRIPT

Physical activity and cognition

to follow-up appointments or difficulties in implementing a healthy diet and physical activity into daily life. In fact, inferior performance on aspects of cognitive function was associated

PT

with lower rates of reported adherence in bariatric surgery patients (35). Hence, the supposed relationship between low PA and cognitive dysfunction in obesity has implications for

SC RI

designing effective weight loss, weight loss maintenance, and prevention strategies. Following this line, improvement in PA that is connected to an enhancement of cognitive performance (36) could facilitate changes in lifestyle behaviors (16). With respect to bariatric

NU

surgery, improvements in PA and cognitive function may positively influence patients’

MA

adherence to postoperative guidelines and increase postoperative weight loss outcome (32). To date, research clarifying the potential association of PA and cognitive performance in obesity is still in its infancy. The majority studies examining PA in obese samples have

ED

used surveys due to ease of administration and relatively low burden for the participants.

PT

However, the congruence between self-reported and objectively determined estimates of PA is rather poor (37). Objective measurement devices such as accelerometers are thought to

CE

more accurately assess PA compared to self-ratings (38). They can successfully measure

AC

movement intensity as counts or steps per time interval and can also assess active energy expenditure in natural settings (39,40). Moreover, accelerometers provide large amounts of data that can be used to identify distinct behavior types of PA based on duration, intensity and regularity of activities (41,42). Recently, Galioto et al. (2014) (29) explored the extent of PA and its relationship to attention, executive function, memory and verbal fluency in severely obese individuals. In this study, 31 bariatric surgery candidates were presented with a standardized cognitive test battery and wore an activity monitor, fastened above the ankle, for seven days prior to surgery that combined acceleration and timing information to measure steps per minute. Almost half of the sample (45.1%) was categorized as sedentary or low physically active according to standardized steps per day indices suggested by Tudor-Locke and Bassett (2004) (43). 4

ACCEPTED MANUSCRIPT

Physical activity and cognition

Contrary to expectations, after adjusting the results for potential covariates (e.g., BMI, somatic comorbidities) no significant association was found between objectively measured

PT

PA and cognitive function. However, the relatively small sample size might have severely limited statistical power in this study.

SC RI

Given the lack of studies that have addressed the proposed link between PA and cognitive function in obesity, the aim of the present study was to determine the relationship between these variables in a sample of patients with extreme obesity by using accelerometers

NU

and computer-assisted cognitive tasks. Considering prior studies in non-obese samples, we

MA

hypothesized that greater levels of PA would be related to better cognitive performance.

PT

ED

METHODS

Study Design

CE

Data for the present study were obtained between March 2013 and February 2014.

AC

Participants were bariatric surgery candidates who were recruited within the routine preoperative psychiatric evaluation at Hannover Medical School. Inclusion criteria were obesity grade 2 or 3 (BMI ≥ 35.0 kg/m2) and age between 18 and 65 years. Exclusion criteria were a history of prior bariatric surgery, intellectual disability, developmental or learning disorders, sensory impairment, psychosis, neurological disorders (e.g. multiple sclerosis, stroke), dementia, history of severe head trauma, current substance abuse, and insufficient German language skills. The assessments for the present investigation were scheduled after the routine preoperative evaluation and were conducted by independent assessors who were not included in the psychiatric evaluation.

5

ACCEPTED MANUSCRIPT

Physical activity and cognition

All participants gave written informed consent for participation according to procedures approved by the Institutional Ethics Committee and received a compensation of 50

PT

€.

SC RI

Assessment

The participants wore a physical activity monitor for 24 hours on at least seven consecutive days, completed computerized cognitive tests and filled out self-ratings. The

NU

assessments were performed in a standardized way by two trained assessors (S.L. and C.P.)

MA

who were continuously supervised by the last author.

Physical activity

ED

An objective assessment of daily physical activity (PA) was conducted by means of

PT

the SenseWear Pro2 armband (SWA; BodyMedia, Inc., Pittsburgh, PA), a multi-sensor device worn on the right upper arm over the triceps muscle. Based on configuration with personal

CE

demographic data (subject ID, age, height, weight, gender, handedness and smoking status), it

AC

continuously records motion data from a biaxial accelerometer and physiologic variables (galvanic skin response, skin temperature, near body temperature and heat flux) per aggregated one-minute time interval. By means of proprietary software algorithms provided by BodyMedia, each minute of data is classified into an activity class (e.g., sitting, walking, stair stepping), allowing for an application of context-specific linear regression models to calculate energy expenditure (44,45). Previous studies have reported an accurate and reliable estimation of energy expenditure by the SWA in a lab setting when compared to the reference standards doubly labeled water or indirect calorimetry (45,46). The SWA has been used to objectively assess PA in obese cohorts including bariatric surgery candidates within standardized activities (28,47-49). 6

ACCEPTED MANUSCRIPT

Physical activity and cognition

In the present study, patients were asked to wear the SWA for 24 hours on at least seven consecutive days. In order to achieve a reliable estimation of habitual PA, validity

PT

criteria were defined corresponding to best research and practice recommendations (50,51). Specifically, data were considered to be valid if the SWA was worn for a minimum of 12

SC RI

hours/day on at least 3 days. Data from 10 participants were determined to be invalid based upon these criteria. Minimal wear time requirements were not met by five participants and five had to be excluded due to technical failure of the SWA (see Figure 1).

NU

Data were transformed into text files by SenseWear Research software version 4.1

MA

(Body Media, Inc., Pittsburgh, PA). Raw data (e.g., total number of steps) were then processed into aggregated data (e.g., mean steps per minute within wear time) and exported to IBM SPSS Statistics software version 22 for further statistical analysis.

ED

Outcome variables used for analysis were mean steps per minute, percentage of PA,

PT

mean active energy expenditure per minute (all three of them in relation to minutes wear time) and mean step frequency. Steps per day were calculated by dividing cumulated steps

CE

within wear time by the number of wear time days. In addition, a composite physical activity

AC

score based on frequency, intensity and duration of respective PA periods was used in order to evaluate PA in the present sample (52). Frequency and duration were determined by PA intervals detected while intensity of PA was computed based on the related accelerometric values.

Neuropsychological Performance Computer-assisted behavioral tasks were administered on the first day of PA assessment to measure executive functioning, memory and learning abilities. A computerized version of the Iowa Gambling Task (IGT) (53) was used to assess executive functioning, particularly real-life decision-making under ambiguity. On a touch screen, four decks of cards were presented to the participants, each labeled as deck A, B, C 7

ACCEPTED MANUSCRIPT

Physical activity and cognition

and D. They all contained monetary gains (reward) and losses (penalty) differing with regard to their frequency and amount. While deck A and B yielded high and immediate monetary

low losses resulting in an overall profit (advantageous).

PT

profit but high losses (disadvantageous), deck C and D yielded modest monetary profit and

SC RI

At the beginning, subjects were provided with a virtual amount of 2000 € and were instructed to gain as much money or lose as little money as possible until the task was terminated by the computer. At the end of the task (at trial 100), the IGT net score was

NU

calculated by subtracting the sum of cards chosen from the disadvantageous decks from the

MA

sum of cards chosen from the advantageous decks [(C+D)-(A+B)]. A lower net score indicates a poorer ability to choose advantageously and served as main outcome variable. A computerized version of the Corsi Block Tapping Test (54) was administered in

ED

order to measure the capacity of visuospatial short-term memory. The span task consisted of

PT

nine points arranged irregularly in a pattern correspondent to the original version established by Corsi (1972) (55). The points flashed up in randomized sequences (56) with a minimal

CE

starting span of four points. Immediately after each presented sequence, the subjects were

AC

instructed to reproduce the sequence correctly. One level was characterized by two sequences of equal length, increasing gradually with one point. The task was discontinued as soon as the stop criterion of two inaccurate trials of same length was reached. The total score [(block span)*(number of correct trials)] was used as the primary outcome measure, taking into account both the block span and the performance on each trial within one level and thus resulting in a sensitive and statistically reliable performance estimation (57). The assessment of verbal learning and short-term memory was realized with a computerized version of the Auditory-Verbal Learning Test (58,59). A list of 15 discrete words per trial was read automatically with all words separated by 3-s-intervals. Each of the five consecutive learning trials was followed by a free recall period of 60 seconds in which the subject was instructed to repeat as many of the 15 words as possible. In order to avoid 8

ACCEPTED MANUSCRIPT

Physical activity and cognition

potential distraction mechanisms, no feedback about the correctness and number of repeated words was given (59). Total number of words correctly repeated was counted for each of the

PT

five trials. A total score was calculated using the sum of trials I-V and served as the primary

SC RI

outcome variable (60).

Questionnaires

To control for depressive symptoms that may have confounded PA and/or cognitive

NU

performance, we administered the German version of the 9-item Patient Health Questionnaire-Depression Scale (PHQ-9) (61) that showed a good reliability in the present

MA

sample (Cronbach's α = 0.82). A PHQ-9 total score ≥ 10 indicates the presence of major depressive disorder (62).

ED

The German version of the Eating Disorder Examination-Questionnaire (EDE-Q) (63)

PT

was used to characterize the sample with regard to specific eating disorder psychopathology. The questionnaire consists of the subscales ‘restraint’, ‘eating concern’, ‘weight concern’ and

CE

‘shape concern’ and includes additional items to identify objective binge eating (OBE; i.e.

AC

eating an objectively large amount of food with a sense of loss of control). Regular binge eating was defined as 4 or more OBEs during the past 28 days. A mean total EDE-Q score ≥ 2.3 served as an indicator for the presence of any eating disturbance (64). Cronbach's α of the mean total score was 0.85. Weight, height and sociodemographic data were self-reported. Somatic comorbidities were assessed by medical chart review, particularly hypertension, diabetes, dyslipidemia, sleep apnea, chronic pain, and the number of comorbid somatic disorders.

Data analysis Analyses were performed using IBM Statistical Package for Social Sciences (SPSS, version 22.0). Descriptive statistics (e.g., means, SD, frequency distributions) were generated 9

ACCEPTED MANUSCRIPT

Physical activity and cognition

for all variables. To calculate potential group differences between excluded and included participants and between patients with and without regular binge eating, we performed

PT

nonparametric tests (Mann-Whitney’s U test) given that most PA estimates and most cognitive variables were not normally distributed. Categorical variables were compared

SC RI

across group using χ2-tests. One-tailed Spearman’s rank correlations (due to the directional hypotheses) were used in order to examine the relationship between PA and cognitive variables. To analyze possible moderator effects (e.g., interaction between age, BMI,

MA

moderated regression analyses were used.

NU

depression, or somatic comorbidity and PA in predicting cognitive performance), hierarchical

Statistical significance was defined as p < 0.05. Alpha levels were Bonferroni

PT

ED

corrected for multiple testing as appropriate.

CE

RESULTS

AC

Descriptive characteristics

The initial study sample consisted of 81 patients. Of those, 10 patients were dropped from the study because of invalid accelerometry data, resulting in a final sample of 71 participants (77.5% women) with a mean age of 41.4 years (SD = 11.9, range 19-63) (see Figure 1). Excluded and included individuals did not differ with regard to age, gender or BMI (results not reported here). The participants had an average BMI of 46.9 kg/m2 (SD = 6.0, range 36.3-60.5) with the majority (93.0%) suffering from obesity grade 3 (BMI ≥ 40.0 kg/m2) and only 7% from obesity grade 2 (BMI: 35-39.9 kg/m2). With regard to education, about one quarter of the sample (25.4%) had 12 years of school education or more. Most participants (59.2%) were employed and 7.0% were in school or apprenticeship. The remaining participants were 10

ACCEPTED MANUSCRIPT

Physical activity and cognition

currently unemployed (18.3%) or retired (15.5%). In terms of marital status, 49.3% were

SC RI

[Figure 1]

PT

married, 31.0% were single, 14.1% were divorced, and 5.6% were widowed.

Somatic comorbidities, eating disorders and depressive symptoms

The point prevalence of any somatic disorder in the sample was 60.6% with a mean

NU

number of comorbid somatic disorders of 1.1 (SD = 1.0, range 0-4). Among all assessed

MA

somatic disorders, the highest prevalence was found for hypertension (54.9%) followed by diabetes (16.9%), sleep apnea (15.5%), chronic pain (9.9%), and dyslipidemia (5.6%). According to the EDE-Q cutoff, 54 patients (76.1%) were identified as having eating

ED

disturbances. Twenty-one patients (29.6%) reported regular binge eating. Furthermore, 34

PT

patients (47.9%) were diagnosed with major depressive disorder based on the PHQ-9 cut-off.

CE

Physical activity and cognitive performance

AC

Table 1 shows the results of PA and cognitive performance in the present sample. The patients wore the SWA on average 6.6 days (SD = 1.4, range 3-11) with a mean wear time of 21.4 hours per day (SD = 2.4, range 12.6-23.9). The wear time included 3 to 9 weekdays (M = 4.9, SD = 1.0). Mean steps per day within wear time were 7140 (SD = 3422, median = 6556). According to the standard step per day indices proposed by Tudor-Locke and Bassett (2004) (43), 31.0% of participants were sedentary (< 5000 steps/day), 26.8% were low active (5000 to 7499 steps/day), 23.9% were somewhat active (7500 to 9999 steps/day), 12.7% were active (10000 to 12499 steps/day), and 5.6% were highly active (≥ 12500 steps/day).

[Table 1] 11

ACCEPTED MANUSCRIPT

Physical activity and cognition

Table 2 summarizes the bivariate correlations between PA estimates and measures of

PT

cognitive performance (note that the sample size was N = 68 due to listwise deletion because of missing data in some of the tests). The listwise deletion method was used to handle missing

SC RI

data resulting in a final sample of N = 68 for correlation analyses. At the first glance, we discovered significant negative correlations (p < 0.05) between two out of five PA estimates and the IGT net score: the composite PA factor and mean steps per minute within wear time.

NU

However, after adjusting for multiple testing (p < 0.05/5) these correlations were no longer

MA

significant. No meaningful association emerged between any of the PA estimates and measures of working memory/learning (CBT/wordlist). In addition, the relationships between parameters of PA and age, BMI, number of

ED

comorbid somatic disorders, and depressive symptoms were investigated. As can be seen in

PT

Table 2, lower PA estimates were modestly correlated with higher BMI. No significant relationships were found between PA and age, somatic comorbidities, and depression. Group

CE

comparisons of patients with and without regular binge eating did not reveal significant

AC

differences with respect to PA estimates or cognitive variables (results not reported here).

[Table 2]

The correlation of cognitive measures with age, BMI, number of comorbid somatic disorders, and depressive symptoms revealed some week correlations. Age was inversely correlated with working memory and learning performance (CBT: rs = -0.29, p < 0.01; wordlist: rs = -0.33, p < 0.005). BMI was related to working memory (CBT: rs = 0.24, p < 0.05). Furthermore, weak negative correlations revealed between the number of somatic disorders and working memory performance (CBT: rs = -0.21, p < 0.05) and between depressive symptoms and learning (wordlist: rs = -0.27, p < 0.05). 12

ACCEPTED MANUSCRIPT

Physical activity and cognition

Given the significant correlations between BMI and PA measures as well as the aforementioned correlations between age, BMI, somatic comorbidity and depression with

PT

some of the cognitive tests, we conducted subsequent partial correlations between PA and cognitive performance, controlling for age, BMI, number of somatic disorders, and depressive

SC RI

symptoms. However, the results of the partial correlations (not reported here) did not show any significant relationship between PA and cognitive function.

In order to test potential moderator effects of age, BMI, somatic comorbidity, or

NU

depression, separate hierarchical moderated regression analyses were calculated for each of

MA

the three cognitive tests. However, the only significant interaction effect was found between depression and PA in predicting working memory performance. In the corresponding regression analysis, task performance on the CBT was used as the main dependent variable.

ED

The composite PA score (cPA) and the total PHQ-9 score (depressive symptoms) were

PT

included as potential predictor variables (note that all variables were centralized in accordance with Cohen, Cohen, West, & Aiken, 2003 (65)). In the first step, cPA was not a significant

CE

predictor for CBT performance, R2 = 0.002, F(1, 68) = 0.13, p = 0.72. When adding

AC

depression as a second predictor, the changes in R2 was not significant, resulting in an overall explanation of CBT variance of only 2%; changes in R2 = 0.018, changes in F(1, 67) = 1.22, p = 0.27. In a third step, the interaction effect “depression x cPA” was calculated. Entering the interaction term in the hierarchical regression analysis resulted in a significant increase in CBT’s variance explanation; changes in R2 = 0.124, changes in F(1, 66) = 9.58, p = 0.003. Depression, PA, and the interaction between those variables explained 14.4% of variance in the CBT. The interaction effect is illustrated in Figure 2.

[Figure 2]

13

ACCEPTED MANUSCRIPT

Physical activity and cognition

DISCUSSION The present study aimed at providing further insight into the relationship between PA

PT

and cognitive function in obesity. Contrary to our hypothesis, on the level of bivariate correlations the relationship between PA and cognitive functioning in pre-bariatric surgery

SC RI

patients did not reach significance. Also, subsequent partial correlations adjusting for age, BMI, somatic comorbidity, and depressive symptoms did not change the outcome. Similarly, the findings of hierarchical moderated regression analyses did not indicate significant

NU

interaction effects between these variables and PA in predicting cognitive performance. The

MA

only exclusion was found with regard to depressive symptoms that may be thought to impact the relationship between PA and working memory. In the present sample of pre-bariatric surgery patients those individuals with high depression scores performed worse on the CBT,

ED

regardless of whether they showed low or high PA (see Fig. 2).

PT

The results are in line with those from Galioto et al. (2014) (29) who also did not find a significant association between PA and cognitive function in extremely obese patients. Both

CE

studies investigated pre-bariatric surgery patients and used objective measures to determine

AC

PA and computer-assisted behavioral tasks to assess cognitive function. Although the studies differed in some details - e.g. the location of the PA monitor (ankle vs. upper arm), some of the PA indices, the utilized cognitive tests, and the substantially higher sample size in the present investigation – they produced a similar outcome. More than half of our sample was classified as sedentary or low physically active according to standard step per day indices (43) which is in accordance to the rates reported by Galioto et al. (2014) (29) and consistent with earlier reports on low PA in obese samples (27,28,30,31). As already noted by Galioto et al. (2014) (29), the generally low level of PA in pre-bariatric surgery patients may explain the missing association between PA and cognitive function. Given the dose-response-relationship that has been observed in prior studies with non-bariatric samples (66-68), there is a possibility that many of our participants did not reach 14

ACCEPTED MANUSCRIPT

Physical activity and cognition

the threshold of PA intensity that is required to positively affect cognitive performance. It is noteworthy that the link between PA and cognitive function may represent a rather non-linear

PT

relationship with an association between cognitive functioning and moderate but not light or vigorous exercise (18,69). Taken this consideration into account, our findings mainly suggest

SC RI

that light levels of daily PA appear to have no relationship with cognitive functioning in bariatric surgery candidates. Future research should address the question of whether moderate levels of exercise are related to cognitive performance in morbidly obese individuals.

NU

Another consideration pertains to the hypothesis that social isolation might lessen the

MA

beneficial effect of exercise on cognitive function (70,71). Half of the present study sample reported not living in a relationship, and more than one third was currently not employed or was retired. Some of our patients may have experienced negative effects of their high BMI on

ED

employment opportunities (72,73) or they may have experienced stigmatization (74,75) that

PT

may have been accompanied by a certain level of social isolation potentially suppressing the effect of PA on cognition. However, this assumption needs further clarification and must be

CE

taken as speculative.

AC

Besides the low PA it appears that the present study group was fairly typical for a prebariatric surgery sample in terms of BMI, somatic comorbidities, the level of eating disturbances including regular binge eating, and the extent of depressive symptoms. These variables may have contributed to the lacking association between PA and cognitive function. As mentioned above, patients with high levels of depression indeed performed worse on the CBT irrespective of the level of PA. However, no effects were found for other variables or other tests. Strengths of the present study include the use of objective measures, the relatively large sample size and the inclusion of active energy expenditure per minute as a PA estimate. The present study replicated the findings of Galioto et al. (2014) (29) by using a different methodology in a larger sample from a different country suggesting that the finding 15

ACCEPTED MANUSCRIPT

Physical activity and cognition

concerning lacking association between PA and cognitive function in pre-bariatric surgery patients is robust, which contributes to the literature.

PT

Although the low level of PA likely accounts for the missing positive link between cognition and PA, this restriction is reflective of the population under study and has clinical

SC RI

implications. It is questionable whether weight loss due to bariatric surgery corresponds with an increase in daily PA. According to recent research, most adults who undergo bariatric surgery remain insufficiently physically active after surgery (76,77). Considering these results

NU

and the findings of the present study PA should be assessed not only pre- but also post-

MA

surgery, with referral to PA counseling as appropriate in order to improve surgery-related outcomes.

There are limitations that have to be kept in mind when interpreting the outcomes. We

ED

cannot rule out that we underestimated the PA (e.g. with regard to the number of steps per

PT

day) given that we considered only steps within wear time. PA outside accelerometer wear time was unknown and could not be considered in the correlation analyses that may have

CE

biased the results. Furthermore, the study was conducted in pre-bariatric surgery patients,

AC

preventing the generalizability of our findings to obesity in general. In summary, our findings suggest the absence of an association between PA and cognitive performance in morbidly obese patients. Future research should address doseresponse questions in the area of PA and cognitive function in patients with obesity across a broader BMI range including patients before and after bariatric surgery as well as individuals in non-clinical settings.

Conflict of interest The authors have no competing interests to report.

Acknowledgements 16

ACCEPTED MANUSCRIPT

Physical activity and cognition

This work was supported by a grant of the Federal Ministry of Education and Research within the German Obesity Competence Network (01GI1323). We wish to thank our colleagues Eva

PT

Malenka Voth, M.D., Ekaterini Georgiadou, Ph.D., and Laszlo Gaal, Ph.D., for their support during the course of the study, and the three anonymous reviewers for their constructive

SC RI

remarks.

References

NU

(1) Elias MF, Elias PK, Sullivan LM, Wolf PA, D'Agostino RB. Lower cognitive function in

MA

the presence of obesity and hypertension: the Framingham heart study. Int J Obes Relat Metab Disord 2003 Feb;27(2):260-268.

(2) Gunstad J, Lhotsky A, Wendell CR, Ferrucci L, Zonderman AB. Longitudinal

ED

examination of obesity and cognitive function: results from the Baltimore longitudinal study

PT

of aging. Neuroepidemiology 2010;34(4):222-229. (3) Gunstad J, Paul RH, Cohen RA, Tate DF, Spitznagel MB, Gordon E. Elevated body mass

CE

index is associated with executive dysfunction in otherwise healthy adults. Compr Psychiatry

AC

2007 Jan-Feb;48(1):57-61.

(4) Verdejo-Garcia A, Perez-Exposito M, Schmidt-Rio-Valle J, Fernandez-Serrano MJ, Cruz F, Perez-Garcia M, et al. Selective alterations within executive functions in adolescents with excess weight. Obesity (Silver Spring) 2010 Aug;18(8):1572-1578. (5) Gunstad J, Müller A, Stanek K, Spitznagel MB. Cognitive dysfunction in obesity: implications for bariatric surgery patients. In: Mitchell JE, de Zwaan M, editors. Psychosocial assessment and treatment of bariatric surgery patients New York: Routledge, Taylor & Francis Group; 2011. p. 99-114. (6) Lokken KL, Boeka AG, Yellumahanthi K, Wesley M, Clements RH. Cognitive performance of morbidly obese patients seeking bariatric surgery. Am Surg 2010 Jan;76(1):55-59. 17

ACCEPTED MANUSCRIPT

Physical activity and cognition

(7) Elias MF, Elias PK, Sullivan LM, Wolf PA, D'Agostino RB. Obesity, diabetes and cognitive deficit: The Framingham Heart Study. Neurobiol Aging 2005 Dec;26 Suppl 1:11-

PT

16. (8) Kiunke W, Brandl C, Georgiadou E, Gruner-Labitzke K, Horbach T, Kohler H, et al.

SC RI

Performance in neurocognitive tasks in obese patients. Does somatic comorbidity matter? Front Psychiatry 2013;4:84.

(9) McCrimmon RJ, Ryan CM, Frier BM. Diabetes and cognitive dysfunction. Lancet 2012

NU

Jun 16;379(9833):2291-2299.

MA

(10) Moriarty O, McGuire BE, Finn DP. The effect of pain on cognitive function: a review of clinical and preclinical research. Prog Neurobiol 2011 Mar;93(3):385-404. (11) Naegele B, Launois SH, Mazza S, Feuerstein C, Pepin JL, Levy P. Which memory

ED

processes are affected in patients with obstructive sleep apnea? An evaluation of 3 types of

PT

memory. Sleep 2006 Apr;29(4):533-544.

(12) Pedersen M, Pedersen KK, Bruunsgaard H, Krabbe KS, Thomsen C, Faerch K, et al.

CE

Cognitive functions in middle aged individuals are related to metabolic disturbances and

AC

aerobic capacity: a cross-sectional study. PLoS One 2012;7(12):e51132. (13) Rincon F, Wright CB. Vascular cognitive impairment. Curr Opin Neurol 2013 Feb;26(1):29-36.

(14) Cserjesi R, Luminet O, Poncelet AS, Lenard L. Altered executive function in obesity. Exploration of the role of affective states on cognitive abilities. Appetite 2009 Apr;52(2):535539. (15) Luppino FS, de Wit LM, Bouvy PF, Stijnen T, Cuijpers P, Penninx BW, et al. Overweight, obesity, and depression: a systematic review and meta-analysis of longitudinal studies. Arch Gen Psychiatry 2010 Mar;67(3):220-229.

18

ACCEPTED MANUSCRIPT

Physical activity and cognition

(16) Joseph RJ, Alonso-Alonso M, Bond DS, Pascual-Leone A, Blackburn GL. The neurocognitive connection between physical activity and eating behaviour. Obes Rev 2011

PT

Oct;12(10):800-812. (17) Lipnicki DM, Gunga HC, Belavy DL, Felsenberg D. Bed rest and cognition: effects on

SC RI

executive functioning and reaction time. Aviat Space Environ Med 2009 Dec;80(12):10181024.

(18) Etnier JL, Nowell PM, Landers DM, Sibley BA. A meta-regression to examine the

NU

relationship between aerobic fitness and cognitive performance. Brain Res Rev 2006 Aug

MA

30;52(1):119-130.

(19) Lambourne K. The relationship between working memory capacity and physical activity rates in young adults. J Sports Sci Med 2006 Mar 1;5(1):149-153.

ED

(20) Li L, Men WW, Chang YK, Fan MX, Ji L, Wei GX. Acute aerobic exercise increases

PT

cortical activity during working memory: a functional MRI study in female college students. PLoS One 2014 Jun 9;9(6):e99222.

CE

(21) Smith PJ, Blumenthal JA, Hoffman BM, Cooper H, Strauman TA, Welsh-Bohmer K, et

AC

al. Aerobic exercise and neurocognitive performance: a meta-analytic review of randomized controlled trials. Psychosom Med 2010 Apr;72(3):239-252. (22) Hillman CH, Erickson KI, Kramer AF. Be smart, exercise your heart: exercise effects on brain and cognition. Nat Rev Neurosci 2008 Jan;9(1):58-65. (23) Guiney H, Lucas SJ, Cotter JD, Machado L. Evidence Cerebral Blood-Flow Regulation Mediates Exercise-Cognition Links in Healthy Young Adults. Neuropsychology 2014 Jul 28. (24) Ogoh S, Tsukamoto H, Hirasawa A, Hasegawa H, Hirose N, Hashimoto T. The effect of changes in cerebral blood flow on cognitive function during exercise. Physiol Rep 2014 Sep 28;2(9):10.14814/phy2.12163. Print 2014 Sep 1.

19

ACCEPTED MANUSCRIPT

Physical activity and cognition

(25) Voss MW, Heo S, Prakash RS, Erickson KI, Alves H, Chaddock L, et al. The influence of aerobic fitness on cerebral white matter integrity and cognitive function in older adults:

PT

results of a one-year exercise intervention. Hum Brain Mapp 2013 Nov;34(11):2972-2985. (26) Erickson KI, Voss MW, Prakash RS, Basak C, Szabo A, Chaddock L, et al. Exercise

SC RI

training increases size of hippocampus and improves memory. Proc Natl Acad Sci U S A 2011 Feb 15;108(7):3017-3022.

(27) Bhurosy T, Jeewon R. Overweight and obesity epidemic in developing countries: a

NU

problem with diet, physical activity, or socioeconomic status? ScientificWorldJournal

MA

2014;2014:964236.

(28) Bond DS, Unick JL, Jakicic JM, Vithiananthan S, Pohl D, Roye GD, et al. Objective assessment of time spent being sedentary in bariatric surgery candidates. Obes Surg 2011

ED

Jun;21(6):811-814.

PT

(29) Galioto R, King WC, Bond DS, Spitznagel MB, Strain G, Devlin M, et al. Physical activity and cognitive function in bariatric surgery candidates. Int J Neurosci 2014

CE

Dec;124(12):912-918.

AC

(30) King WC, Bond DS. The importance of preoperative and postoperative physical activity counseling in bariatric surgery. Exerc Sport Sci Rev 2013 Jan;41(1):26-35. (31) King WC, Belle SH, Eid GM, Dakin GF, Inabnet WB, Mitchell JE, et al. Physical activity levels of patients undergoing bariatric surgery in the Longitudinal Assessment of Bariatric Surgery study. Surg Obes Relat Dis 2008 Nov-Dec;4(6):721-728. (32) Galioto R, Gunstad J, Heinberg LJ, Spitznagel MB. Adherence and weight loss outcomes in bariatric surgery: does cognitive function play a role? Obes Surg 2013 Oct;23(10):17031710. (33) Niemeier HM, Phelan S, Fava JL, Wing RR. Internal disinhibition predicts weight regain following weight loss and weight loss maintenance. Obesity (Silver Spring) 2007 Oct;15(10):2485-2494. 20

ACCEPTED MANUSCRIPT

Physical activity and cognition

(34) van Hout GC, Jakimowicz JJ, Fortuin FA, Pelle AJ, van Heck GL. Weight loss and eating behavior following vertical banded gastroplasty. Obes Surg 2007 Sep;17(9):1226-

PT

1234. (35) Spitznagel MB, Galioto R, Limbach K, Gunstad J, Heinberg L. Cognitive function is

SC RI

linked to adherence to bariatric postoperative guidelines. Surg Obes Relat Dis 2013 JulAug;9(4):580-585.

(36) Chang YK, Labban JD, Gapin JI, Etnier JL. The effects of acute exercise on cognitive

NU

performance: a meta-analysis. Brain Res 2012 May 9;1453:87-101.

MA

(37) Bond DS, Thomas JG, Unick JL, Raynor HA, Vithiananthan S, Wing RR. Self-reported and objectively measured sedentary behavior in bariatric surgery candidates. Surg Obes Relat Dis 2013 Jan-Feb;9(1):123-128.

ED

(38) Ward DS, Evenson KR, Vaughn A, Rodgers AB, Troiano RP. Accelerometer use in

PT

physical activity: best practices and research recommendations. Med Sci Sports Exerc 2005 Nov;37(11 Suppl):S582-8.

CE

(39) M. Mathie. Monitoring and Interpreting Human Movement Patterns Using a Triaxial

2003.

AC

Accelerometer. New South Wales: University of New South Wales, Faculty of Engineering;

(40) Marschollek M, Goevercin M, Wolf KH, Song B, Gietzelt M, Haux R, et al. A performance comparison of accelerometry-based step detection algorithms on a large, nonlaboratory sample of healthy and mobility-impaired persons. Conf Proc IEEE Eng Med Biol Soc 2008;2008:1319-1322. (41) Marschollek M. A semi-quantitative method to denote generic physical activity phenotypes from long-term accelerometer data--the ATLAS index. PLoS One 2013 May 8;8(5):e63522.

21

ACCEPTED MANUSCRIPT

Physical activity and cognition

(42) Marschollek M. Clustering physical activity phenotypes using the ATLAS index on accelerometric data from an epidemiologic cohort study. Stud Health Technol Inform

PT

2014;205:763-767. (43) Tudor-Locke C, Bassett DR,Jr. How many steps/day are enough? Preliminary pedometer

SC RI

indices for public health. Sports Med 2004;34(1):1-8.

(44) Grams L, Tegtbur U, Kück M, Gützlaff E, Marschollek M, Kerling A. Energy Expenditure Measurements under Controlled Conditions - A Comparison of Accelerometry,

NU

Multisensorsystem and Mobile Spiroergometry. DZSM 2011;62(6):160-165.

MA

(45) Johannsen DL, Calabro MA, Stewart J, Franke W, Rood JC, Welk GJ. Accuracy of armband monitors for measuring daily energy expenditure in healthy adults. Med Sci Sports Exerc 2010 Nov;42(11):2134-2140.

ED

(46) Jakicic JM, Marcus M, Gallagher KI, Randall C, Thomas E, Goss FL, et al. Evaluation of

PT

the SenseWear Pro Armband to assess energy expenditure during exercise. Med Sci Sports Exerc 2004 May;36(5):897-904.

CE

(47) Bond DS, Jakicic JM, Unick JL, Vithiananthan S, Pohl D, Roye GD, et al. Pre- to

AC

postoperative physical activity changes in bariatric surgery patients: self report vs. objective measures. Obesity (Silver Spring) 2010 Dec;18(12):2395-2397. (48) Unick JL, Bond DS, Jakicic JM, Vithiananthan S, Ryder BA, Roye GD, et al. Comparison of two objective monitors for assessing physical activity and sedentary behaviors in bariatric surgery patients. Obes Surg 2012 Mar;22(3):347-352. (49) Josbeno DA, Kalarchian M, Sparto PJ, Otto AD, Jakicic JM. Physical activity and physical function in individuals post-bariatric surgery. Obes Surg 2011 Aug;21(8):1243-1249. (50) Trost SG, McIver KL, Pate RR. Conducting accelerometer-based activity assessments in field-based research. Med Sci Sports Exerc 2005 Nov;37(11 Suppl):S531-43. (51) Heil DP, Brage S, Rothney MP. Modeling physical activity outcomes from wearable monitors. Med Sci Sports Exerc 2012 Jan;44(1 Suppl 1):S50-60. 22

ACCEPTED MANUSCRIPT

Physical activity and cognition

(52) Schulze M, Pek C, Langenberg S, Bartsch M, Crosby RD, Marschollek M, et al. Does obesity bias sensor-based activity monitoring in daily life? (manuscript submitted for

PT

publication). (53) Bechara A, Damasio AR, Damasio H, Anderson SW. Insensitivity to future

SC RI

consequences following damage to human prefrontal cortex. Cognition 1994 Apr-Jun;50(13):7-15.

(54) Berch DB, Krikorian R, Huha EM. The Corsi block-tapping task: methodological and

NU

theoretical considerations. Brain Cogn 1998 Dec;38(3):317-338.

MA

(55) Corsi PM. Human memory and the medial temporal region of the brain. Dissertation Abstracts International 1972;34 (02):819B.

(56) Smirni P, Villardita C, Zappala G. Influence of different paths on spatial memory

ED

performance in the Block-Tapping Test. J Clin Neuropsychol 1983 Dec;5(4):355-359.

PT

(57) Kessels RP, van Zandvoort MJ, Postma A, Kappelle LJ, de Haan EH. The Corsi BlockTapping Task: standardization and normative data. Appl Neuropsychol 2000;7(4):252-258.

CE

(58) Rey A. L'examen clinique en psychologie. Paris: Presses Universitaires de France 1964.

AC

(59) Lezak MD. Neuropsychological Assessment. 2nd ed. New York: Oxford University Press; 1983.

(60) Powell JB, Cripe LI, Dodrill CB. Assessment of brain impairment with the Rey Auditory Verbal Learning Test: a comparison with other neuropsychological measures. Arch Clin Neuropsychol 1991;6(4):241-249. (61) Kroenke K, Spitzer RL, Williams JB. The PHQ-9: validity of a brief depression severity measure. J Gen Intern Med 2001 Sep;16(9):606-613. (62) Gilbody S, Richards D, Brealey S, Hewitt C. Screening for depression in medical settings with the Patient Health Questionnaire (PHQ): a diagnostic meta-analysis. J Gen Intern Med 2007 Nov;22(11):1596-1602.

23

ACCEPTED MANUSCRIPT

Physical activity and cognition

(63) Hilbert A, Tuschen-Caffier B. Eating Disorder Examination-Questionnaire. Deutsche Übersetzung. Münster: Verlag für Psychotherapie; 2006.

PT

(64) Hilbert A, de Zwaan M, Braehler E. How frequent are eating disturbances in the population? Norms of the eating disorder examination-questionnaire. PLoS One

SC RI

2012;7(1):e29125.

(65) Cohen J, Cohen P, West SG, & Aiken LS. Applied multiple regression/correlation analysis for the behavioral sciences. 3rd ed. Mahwah, NJ: Lawrence Erlbaum; 2003.

NU

(66) Kerr J, Marshall SJ, Patterson RE, Marinac CR, Natarajan L, Rosenberg D, et al.

MA

Objectively measured physical activity is related to cognitive function in older adults. J Am Geriatr Soc 2013 Nov;61(11):1927-1931.

(67) Xu L, Jiang CQ, Lam TH, Zhang WS, Thomas GN, Cheng KK. Dose-response relation

ED

between physical activity and cognitive function: guangzhou biobank cohort study. Ann

PT

Epidemiol 2011 Nov;21(11):857-863.

(68) Davis CL, Tomporowski PD, McDowell JE, Austin BP, Miller PH, Yanasak NE, et al.

CE

Exercise improves executive function and achievement and alters brain activation in

AC

overweight children: a randomized, controlled trial. Health Psychol 2011 Jan;30(1):91-98. (69) Geda YE, Roberts RO, Knopman DS, Christianson TJ, Pankratz VS, Ivnik RJ, et al. Physical exercise, aging, and mild cognitive impairment: a population-based study. Arch Neurol 2010 Jan;67(1):80-86. (70) Cinini SM, Barnabe GF, Galvao-Coelho N, de Medeiros MA, Perez-Mendes P, Sousa MB, et al. Social isolation disrupts hippocampal neurogenesis in young non-human primates. Front Neurosci 2014 Mar 27;8:45. (71) Stranahan AM, Khalil D, Gould E. Social isolation delays the positive effects of running on adult neurogenesis. Nat Neurosci 2006 Apr;9(4):526-533. (72) Greve J. Obesity and labor market outcomes in Denmark. Econ Hum Biol 2008 Dec;6(3):350-362. 24

ACCEPTED MANUSCRIPT

Physical activity and cognition

(73) Lindeboom M, Lundborg P, van der Klaauw B. Assessing the impact of obesity on labor market outcomes. Econ Hum Biol 2010 Dec;8(3):309-319.

PT

(74) Giel KE, Zipfel S, Alizadeh M, Schaffeler N, Zahn C, Wessel D, et al. Stigmatization of obese individuals by human resource professionals: an experimental study. BMC Public

SC RI

Health 2012 Jul 16;12:525-2458-12-525.

(75) Hilbert A, Rief W, Braehler E. Stigmatizing attitudes toward obesity in a representative population-based sample. Obesity (Silver Spring) 2008 Jul;16(7):1529-1534.

NU

(76) Berglind D, Willmer M, Eriksson U, Thorell A, Sundbom M, Udden J, et al.

Obes Surg 2015 Jan;25(1):119-125.

MA

Longitudinal assessment of physical activity in women undergoing Roux-en-Y gastric bypass.

(77) King WC, Hsu JY, Belle SH, Courcoulas AP, Eid GM, Flum DR, et al. Pre- to

ED

postoperative changes in physical activity: report from the longitudinal assessment of bariatric

AC

CE

PT

surgery-2 (LABS-2). Surg Obes Relat Dis 2012 Sep-Oct;8(5):522-532.

25

ACCEPTED MANUSCRIPT

Physical activity and cognition

AC

CE

PT

ED

MA

NU

SC RI

PT

Figure 1: Study flow

Note. SWA = Sense Wear Armband

26

ACCEPTED MANUSCRIPT

Physical activity and cognition

Table 1 Parameters of physical activity and neurocognitive performance in bariatric surgery

PT

candidates (N = 71)

M (SD)

SC RI

Parameters of physical activity Composite physical activity score

18.21 (14.21)

Mean steps per minute within wear time

5.68 (3.02)

37.63 (11.72)

Physical activity within wear time [percent]

NU

Step frequency [steps/min]

MA

Active energy expenditure per minute within wear time [kcal] within wear time [kcal]

ED

Neurocognitive performance

8.45 (6.22) 0.55 (0.48)

M (SD) 17.1 (22.2)

Corsi Block Tapping Test

33.0 (16.7)

PT

Iowa Gambling Task [net score]

55.0 (7.3)

CE

Wordlist [(block span) x (number of correct trials)]

AC

[total number of correctly repeated words]

27

ACCEPTED MANUSCRIPT Physical activity and cognition Table 2

number of somatic comorbidities, and depressive symptoms (N = 68)

Physical activity

IGT

CBT

Composite PA score

-0.21*

-0.05

Mean steps per minute within wear time

-0.21*

Step frequency [steps/min]

-0.09

Physical activity within wear time

-0.19

Active energy expenditure per minute within wear time [kcal]

-0.18

Age

Number of BMI

comorbid somatic

Depressive

disordersa

symptomsb

Wordlist -0,02

-0,40**

-0,05

0,07

-0.01

-0.17

-0,06

-0,41**

-0,12

0,11

0.17

-0.08

-0,09

-0,27*

-0,10

-0,06

-0.08

-0.09

-0,03

-0,35**

-0,07

0,11

-0.11

-0,11

-0,22*

-0,09

0,03

PT ED

-0.10

AC CE

[percent]

MA

NU

Neurocognitive performance

SC RI

PT

One-tailed Spearman rank correlations between parameters of physical activity and measures of neurocognitive performance, age, BMI,

-0.01

Note. PA = physical activity, IGT = Iowa Gambling Task, CBT = Corsi Block Tapping Test, BMI = body mass index. a including hypertension, diabetes, sleep apnea, chronic pain, dyslipidemia. b Patient Health Questionnaire-9. *p < 0.05, **Bonferroni corrected p < 0.01 (0.05/5).

28

ACCEPTED MANUSCRIPT

Physical activity and cognition

Figure 2 Results of the regression analysis predicting the interaction effect of depression and

PT

physical activity on CBT performance 37

SC RI

36

34 33 32

NU

CBT performance

35

31

29 low Physical Activity

MA

30

high Physical Activity

ED

low Depression high Depression

Note. This interaction indicates that patients with high depression scores performed worse on

AC

CE

Block Tapping Test

PT

the CBT, regardless of whether they showed low or high physical activity. CBT = Corsi

29

ACCEPTED MANUSCRIPT

Physical activity and cognition

AC

CE

PT

ED

MA

NU

SC RI

PT

Conflict of interest The authors have no competing interests to report.

30

ACCEPTED MANUSCRIPT

Physical activity and cognition

Highlights: Physical activity is not associated with cognitive performance in obese prebariatric surgery patients

We examined physical activity and cognitive performance in extreme obesity



Pre-bariatric surgery patients were considerably low active in daily life



Physical activity was not related to cognitive performance



Physical activity was not related to age, somatic comorbidity or depression



Lower physical activity was related to higher body mass index

AC

CE

PT

ED

MA

NU

SC RI

PT



31