ArticlePDF Available

The quality and duration of sleep are related to hedonic hunger: a cross-sectional study in university students

January 2021
Sleep and Biological Rhythms 19(20)

DOI:10.1007/s41105-020-00303-8

Authors:

Murat Açık

Fırat University

Funda pinar Cakiroglu

Ankara University

There is a relationship between extreme sleep duration and increased food intake. Some studies have reported that there was no change in the homeostatic aspects of energy balance, despite the increased nutrient intake, and in this case, the hedonic aspects may be effective. The purpose of this study was to examine the associations of hedonic hunger with sleep quality and duration among university students. This cross-sectional study was carried out on university students. An online questionnaire was applied in the study, in which 1144 participants were included. The Pittsburgh Sleep Quality Index (PSQI) was used to assess sleep quality and duration, and the Power of Food Scale (PFS) and Palatable Eating Motives Scale (PEMS) for detecting hedonic hunger states. Multiple linear regression analysis was performed to evaluate the relationship between sleep quality and duration and hedonic hunger by modelling. Total PFS and PEMS scores were positively associated with PSQI scores after controlling for all possible confounding factors [β (95%CI) = 0.04 (0.03–0.05), p < 0.05; 0.06 (0.05–0.07), p < 0.05, respectively). The relationship between the subdimensions of the PFS, PEMS and PSQI remained statistically significant, except for food availability (PFS) and social motive (PEMS).After fully adjusting, the odds of having a high PEMS score increased in individuals with short or long sleep duration, but not PFS score [OR (95%CI) = 1.40 (1.09–1.83), p = 0.012; 0.98 (0.70–1.21), p = 0.878, respectively]. While a positive relationship was found between increased hedonic hunger and poor sleep quality, an inverse relationship was observed between ideal sleep duration and hedonic hunger. The findings suggest that improving sleep quality and duration can help reduce hedonic hunger, which increases the tendency to unhealthy and delicious foods and plays a role in weight gain.

Content uploaded by Murat Açık

Content may be subject to copyright.

Vol.:(0123456789)

1 3

Sleep and Biological Rhythms (2021) 19:163–172

https://doi.org/10.1007/s41105-020-00303-8

ORIGINAL ARTICLE

The quality andduration ofsleep are related tohedonic hunger:

across‑sectional study inuniversity students

MuratAçik1 · AyşeNurSongürBozdağ1· FundaPınarÇakiroğlu1

Received: 19 June 2020 / Accepted: 24 December 2020 / Published online: 16 January 2021

Abstract

There is a relationship between extreme sleep duration and increased food intake. Some studies have reported that there was

no change in the homeostatic aspects of energy balance, despite the increased nutrient intake, and in this case, the hedonic

aspects may be eﬀective. The purpose of this study was to examine the associations of hedonic hunger with sleep quality and

duration among university students. This cross-sectional study was carried out on university students. An online question-

naire was applied in the study, in which 1144 participants were included. The Pittsburgh Sleep Quality Index (PSQI) was

used to assess sleep quality and duration, and the Power of Food Scale (PFS) and Palatable Eating Motives Scale (PEMS)

for detecting hedonic hunger states. Multiple linear regression analysis was performed to evaluate the relationship between

sleep quality and duration and hedonic hunger by modelling. Total PFS and PEMS scores were positively associated with

PSQI scores after controlling for all possible confounding factors [β (95%CI) = 0.04 (0.03–0.05), p < 0.05; 0.06 (0.05–0.07),

p < 0.05, respectively). The relationship between the subdimensions of the PFS, PEMS and PSQI remained statistically

signiﬁcant, except for food availability (PFS) and social motive (PEMS).After fully adjusting, the odds of having a high

PEMS score increased in individuals with short or long sleep duration, but not PFS score [OR (95%CI) = 1.40 (1.09–1.83),

p = 0.012; 0.98 (0.70–1.21), p = 0.878, respectively]. While a positive relationship was found between increased hedonic

hunger and poor sleep quality, an inverse relationship was observed between ideal sleep duration and hedonic hunger. The

ﬁndings suggest that improving sleep quality and duration can help reduce hedonic hunger, which increases the tendency to

unhealthy and delicious foods and plays a role in weight gain.

Keywords Sleep quality· Hedonic hunger· Sleep duration· Appetite· University students

Introduction

Sleep is a rapidly reversible recurring state of inactivity

associated with diminished responsiveness to the external

environment [1]. Short and long sleep duration have been

suggested to be risk factors for chronic diseases [2]. Pro-

spective cohort studies especially point out that extremes of

sleep duration (short and long) are associated with adverse

health outcomes, including increased risk of obesity and

non-communicable disease (i.e. diabetes and cardiovascular

diseases etc.) [3, 4]. Moreover, there is much growing evi-

dence investigating the relationship between sleep quality or

time and obesity, including extensive meta-analysis reports

in recent years [5–7].

Epidemiologic studies have demonstrated connections

between sleep duration and diet. Sleep deprivation can

modify nutritional habits [8], and adequate sleep is posi-

tively associated with health-related behaviour, such as the

adoption of healthy eating habits. Recent studies observed

a U-shaped association between sleep duration and eating

behaviours and dietary quality among women; women with

short or long sleep duration were more likely to eat dur-

ing unconventional hours and replace meals with snacks

than women with adequate sleep duration. Decreasing or

increasing sleep duration decays the control of homeostatic,

increasing food intake (especially, intake of foods rich in fat

and carbohydrates) [9, 10]. The underlying mechanism is

that short sleep duration is associated with decreased leptin

or increased ghrelin levels [11, 12]. However, Calvin etal.

[13] did not detect any change in the leptin and ghrelin levels

* Murat Açik

acikm@ankara.edu.tr

1 Faculty ofHealth Sciences, Department ofNutrition

andDietetics, Ankara University, Tepebaşi Neighborhood,

Fatih Street, Keçiören, Ankara, Turkey

164 Sleep and Biological Rhythms (2021) 19:163–172

1 3

of individuals with short sleep duration, and they observed

an increase in the food intake of these participants. In recent

systematic and meta-analysis reports, it was suggested that

the hedonic system, which increases in short sleep duration,

seems to play a crucial role in food intake, and increased

food intake was observed in cases where the homeostatic

control did not change [14]. This may be attributed to het-

erogeneity in the study designs, but other researchers suggest

that these results are due to increased hedonic system activ-

ity that may contribute to increased food intake independent

of hormonal change [13, 15]. On the other hand, it is known

that there is a negative relationship between prolonged sleep

time and gut hormone glucagon-like peptide 1, which pro-

vides satiety due to long sleep duration, although there are

not many studies on weight gain or food intake. Moreover,

it activates sedentary behaviours in individuals with long

sleep duration and as a result it plays a role in appetite regu-

lation, energy expenditure and body composition, due to the

decrease in leptin, the anorexigenic adipokine [16].

Hedonic eating is satisﬁed by the intake of highly palat-

able foods, which are typically made tasty by their higher

fat, sugar, and salt contents and hence also tend to be dense

in calories [17]. However, it is diﬃcult to evaluate hedonic

hunger in those with eating behaviour disorders and under-

weight individuals. Individuals, are underweight and have

eating behaviour disorders such as anorexia nervosa, have

low body weight and a signiﬁcant negative energy balance.

Therefore, they may have higher or lower hedonic hunger

scores [18]. It was demonstrated that the eﬀects of substance

abuse on the limbic system are similar to the eﬀects of the

consumption of delicious foods (especially high fat and/or

sugar) on the limbic system [18]. Neuroimaging studies sup-

port the evidence that some individuals may be obese due

to a disorder in their dopaminergic pathways [19, 20]. It

was suggested, in experimental studies, that the activation

of neurons in the limbic system of participants, responsible

for the reward, increased after short-term sleep manipula-

tion. Thus, the increase in food intake associated with sleep

duration seems to be mainly driven by hedonic and nonhor-

monal factors [21, 22]. However, this view only is limited to

short-term intervention studies. Therefore, epidemiological

studies are also needed to conﬁrm such a relationship in the

large population.

To the best of our knowledge, there are no studies in the

literature that examine the eﬀect of sleep duration or quality

on hedonic hunger. However, a few studies have explored

the possible inﬂuence of eating behaviour constructs, as an

independent and strong predictor of weight gain, short sleep

duration and weight status. Chaput etal. [23] have inves-

tigated the relationship between sleep duration or quality

and eating behaviour. In a 6-year longitudinal study, they

observed that short sleep with a high disinhibition eating

behaviour trait had a greater increase in body weight. In a

recent study, Blumﬁeld etal. [24] examined if eating behav-

iour mediates the relationship between sleep and body mass

index (BMI) in a large sample of American adults. It was

found that disinhibited eating behaviour mediated the rela-

tionship between sleep quality and weight status. Results

of this research suggest that improving sleep quality may

beneﬁt weight loss by helping to reduce an individual’s sus-

ceptibility to overeating. In both studies, the relationship

between sleep duration or quality and three-factor eating

attitude-behaviour was examined. Although eating behav-

iour is an independent and powerful marker of weight gain

in adults, it cannot provide detailed information about the

hedonic hunger status and psychological eﬀects in various

delicious food environments. Since both sleep and hedonic

hunger are two important factors in weight management,

it is important to explain the relationship between them.

In addition, when such studies are conducted in university

students, who are considered a high-risk group in terms of

sleep behaviours and eating habits, more eﬀective strategies

can be developed in the struggle against obesity and adop-

tion of healthy life habits in young people. To this end, we

conducted a study to see if hedonic hunger is associated with

sleep quality and to what extent hedonic hunger is related to

sleep quality, depending on factors that can aﬀect it.

Materials andmethods

Participants

Data were obtained from a cross-sectional study that was

conducted in a convenience sample of students studying at

six universities in Ankara. In the research, an online ques-

tionnaire was used as a data collection tool. This online

questionnaire was shared on the social media accounts of

universities in Ankara province and reached the participants

during October 2020. Four weeks following the students’

invitations, a convenient sample of 1250 participants was

reached, allowing the survey to be closed. Informed consent

was obtained prior to beginning the online survey. Partici-

pants who provided exclusion criteria (i.e. the presence of

any major disease, nutrition and dietetic or sports science

students and underweight individuals (BMI < 18.5kg/m2))

were excluded (n = 106) to give a total sample size of 1144

adults.

Participants were full-time students, aged 18–28years,

and enrolled at six universities. Eligibility criteria for this

study included the following: BMI ≥ 18.5kg/m2; not a

nutrition major or sports science major graduate or student;

depression, anxiety, or other psychiatric disorders; cancer;

history of obstructive sleep apnoea; free from health condi-

tions that could interfere with diet and exercise changes;

taking part in another study; and not pregnant or lactating.

165Sleep and Biological Rhythms (2021) 19:163–172

1 3

Questionnaires

An online, self-administered questionnaire was used to

obtain data about demographic and personal characteris-

tics, physical activity level, sleep time-quality and hedonic

hunger. These ﬁve constructs were assessed using previously

validated questionnaires, and the ﬁnal questionnaire required

approximately 25min to complete.

Demographic andpersonal characteristics

Participants were asked to report their gender, age, smok-

ing status, alcohol consumption, disease status, height and

weight. The BMI was calculated as weight (kg) divided by

height squared (m2).

Physical activity

The International Physical Activity Questionnaire (IPAQ)

was developed by Craig etal. [25] to determine the physi-

cal activity levels of participants aged 15–65years. This

questionnaire provides information about the time spent

sitting, walking, and in moderate to vigorous activities. In

order to evaluate all activities, it is taken as a criteria that

each activity is performed for at least 10min at a time. Data

collected with the IPAQ can be reported as a continuous

measure of total physical activity and reported in metabolic

equivalent of task (MET) minutes. In calculating the score

related to physical activity, the weekly duration (minutes)

of each activity and the metabolic equivalent (MET) energy

values created for the IPAQ are multiplied [25]. The Turk-

ish validity and reliability study of the IPAQ was done by

Öztürk [26].

Sleep assessment

Sleep quality was measured with the Pittsburgh Sleep Qual-

ity Index (PSQI), which was used as a subjective measure

of sleep quality and disturbances of the patient’s sleep. The

PSQI is a 19-item self-reported questionnaire grouped into

seven score components (subjective sleep quality, sleep

latency, sleep duration, habitual sleep eﬃciency, sleep dis-

turbances, use of sleeping medication and daytime dysfunc-

tion). In addition, ﬁve questions of the PSQI are answered

by the individual’s sleep partner, if applicable. However,

these questions are not included in the scoring. The PSQI

questions consist of a 4-point Likert structure and are scored

ranging from “0” to “3”. The PSQI yields a score from 0

(good quality) to 21 (poor quality). If the total score of the

PSQI is > 5, it indicates poor sleep quality. Those with a

score ≤ 5 were classiﬁed as good sleepers [27]. The ques-

tionnaire was validated previously for use in Turkey and

achieved a Cronbach’s alpha of 0.80 for 114 individuals [28].

Hedonic Hunger

Power offood scale

The Power of Food Scale (PFS) assesses the psychological

impact of living in food-abundant environments. It measures

appetite for, rather than consumption of, palatable foods at

three levels of food proximity (food available, food present,

and food tasted) [29]. The questionnaire originally started

with 21 items, but it was decreased to 15 items after the

validity and reliability analyses in a Turkish population. The

validity and reliability study of the survey carried out on 363

university students was conducted by Hayzaran etal., and

a Cronbach’s alpha of 0.85 for the Turkish population was

observed [30]. The current PFS consists of 15 items rated

on a 5-point Likert scale, with an aggregate score and three

subscales: (1) food available, but not present (e.g. ‘‘I seem

to have food on my mind a lot’’); (2) food present, but not

tasted (e.g. ‘‘If I see or smell a food I like, I get a power-

ful urge to have some’’); and (3) food tasted, but not eaten

(e.g. ‘‘When I eat delicious food, I focus a lot on how good

it tastes’’). It means that when the individual’s scale score

increases, the inﬂuence of the PFS (hedonic hunger) on the

individual increases [29].

Palatable eating motives scale

The Palatable Eating Motives Scale (PEMS) scale, devel-

oped by Burgess etal. and composed of 20 items, was

reduced to 19 items after validity and reliability analyses

were performed [31]. The PEMS scale was developed to

identify the reasons why individuals consume delicious food

and drink. The scale includes four subfactors: social, cop-

ing, reward enhancement, and conformity motives. Social

motives include questions about consuming palatable foods

or beverages for social reasons (e.g. consumption to enjoy a

party, eating behaviour shown to socialize more with friends,

etc.). Coping motive includes items related to the consump-

tion of palatable foods or beverages to deal with negative

feelings (such as anxiety, eating behaviour in a sad situa-

tion, depression, or eating behaviour related to frustration

and anger). Reward enhancement motive involves ques-

tions about consuming palatable foods or drinks to enhance

positive experiences, emotions or inherent satisfying traits

that are not related to social situations (e.g. eating behav-

iour towards foods that give pleasure or make the individual

happy when consumed). Conformity motive involves ques-

tions about the consumption of palatable foods in response

to external pressures (e.g. conformity to social environments

with friends and eating behaviour shown to avoid being

excluded) [31]. Responses can range from 1 (Never/Almost

Never) to 5 (Almost Always/Always) and are scored as the

mean response for each motive. The validity and reliability

166 Sleep and Biological Rhythms (2021) 19:163–172

1 3

study of the survey carried out on 363 university students

was conducted by Hayzaran etal. [30]. The reliability coef-

ﬁcient of the PEMS (Cronbach’s alpha) has been calculated

as 0.88, with the four subscales of the PEMS approved in

conﬁrmatory factor analysis changes between 0.75 and 0.89

[30].

Statistical analysis

We calculated the mean ± standard deviation values for

continuous variables and the frequency (percentage) val-

ues for nominal variables. We used independent t tests to

compare continuous variables and Pearson’s Chi squared

test to compare categorical data between groups by gender.

After the necessary criteria were provided, linear regression

analyses were performed to calculate the β-coeﬃcient of

the association between BMI and hedonic hunger status by

gender. We also calculated crude values of the linear regres-

sion relationship between hedonic hunger and sleep quality.

After the confounding factors “age, BMI and smoking” were

included in the model, multiple linear regression was used to

determine the association between hedonic hunger and sleep

quality. As emphasized in the introduction, long and short

sleep duration share similar physiological characteristics.

Therefore, short and longed sleep duration groups (< 7h/d

or > 9h/d, respectively) were combined and logistic regres-

sion analysis was used to determine the eﬀect on hedonic

hunger compared to ideal sleep duration (7–9h/d). The other

main reason for combining groups is that the number of indi-

viduals with long sleep duration (n = 52) is quite low. Then,

we created multiple logistic regression analyses to assess the

relationship between the PFS and PEMS scores and good/

short or long sleep duration and to determine whether the

odds of good/short or long sleep duration were associated

with hedonic hunger by gender. Moreover, model 1 was

adjusted for age and BMI. Model 2 additionally adjusted

for physical activity and smoking. A p < 0.05 was considered

statistically signiﬁcant. All statistical analyses were calcu-

lated using the Statistical Package for the Social Sciences

(SPSS) version 21 for Windows (SPSS Inc., Chicago, IL).

Results

Descriptive statistics are presented in Table1. In all,

25.3% of the 1144 participants in the study were male and

74.7% were female. The mean age of the participants was

22.0 ± 2.7years, and the mean BMI was 24.6 ± 3.6kg/m2

(male: 25.6 ± 3.4 kg/m2, female: 24.2 ± 3.6 kg/m2). The

majority (57.5%) of the participants were normal, 34.7%

were overweight and 7.8% were obese. Females were more

likely to be younger (p = 0.003) and have a lower BMI

(p < 0.001), smoking (p < 0.001), alcohol consumption

and physical activity score (p < 0.001) compared to males

(Table1). While the average sleep quality scores of the

participants were 6.1 ± 2.6, the sleep quality was worse

in females (6.3 ± 2.6) compared to males (5.8 ± 2.7)

(p < 0.001). However, there was no signiﬁcant diﬀerence

between females and males in sleep duration (6.9 ± 1.4h

for females and 6.9 ± 1.4h for males; p = 0.748) and time

in bed (p = 0.253).

Table1 also lists the mean scores for females and males

on the PFS and the PEMS subscales. The scores of PFS

and its subscales were found to be higher in women than

men (p < 0.001). The coping motive (PEMS subscale) scores

were signiﬁcantly higher in females, while the conformity

motive (PEMS subscale) scores were signiﬁcantly higher in

males (p < 0.05).

We showed the association of the PEMS and PFS aggre-

gated and their subscales scores with BMI in the female

and male groups using linear regression (Table2). For the

total group, high PFS aggregated score [β (95%CI) = 0.07

(0.06–0.09), p < 0.001], “food available” [β (95%CI) = 0.18

(0.15–0.22), p < 0.001), “food present” [β (95%CI) = 0.22

(0.17–0.27), p < 0.001] and “food tasted” [β (95%CI) = 0.14

(0.10–018), p < 0.001] scores were independently associated

with high BMI, and explained 9.4% of food available, 7.2%

of food present and 4.1% of food tasted. This association

remained signiﬁcant in both groups when classiﬁed accord-

ing to gender. In the total group, PEMS scores were posi-

tively associated with BMI [β (95%CI) = 0.05 (0.04–0.07),

p < 0.001] and explained 4.6% of its. Moreover, there was

a signiﬁcant relationship between poorer sleep quality and

higher “coping motive and reward motive” (PEMS sub-

scales), but not “social motive” or “conformity motive”.

While positive linear trends for only “social motive” scores

with BMI did not remain statistically significant in the

female group and for both “social motive” and “conformity

motive” in the male group.

After modelling was performed with age, BMI, smok-

ing and physical activity, which were thought to aﬀect

the PFS and PEMS subscales, multiple linear regres-

sion analysis was conducted to determine the eﬀect of

hedonic hunger on sleep quality (Table3). The PFS total

scores were positively associated with the PSQI scores

after being fully adjusted [β (95%CI) = 0.04 (0.03–0.05),

p < 0.05] and coeﬃcient of determination was found to

be 20.0%. Moreover, “food present” and “food tasted”

scores were positively associated with the PSQI score after

adjusting for age, BMI, smoking and physical activity [β

(95% CI) = 0.15 (0.10–0.17), p < 0.05 for food present;

β (95% CI) = 0.14 (0.10–0.16), p < 0.05 for food tasted].

The PEMS scores were positively associated with the

PSQI scores after controlling for all possible confound-

ing factors [β (95%CI) = 0.06 (0.05–0.07), p < 0.05] and

explained 25.6% of PEMS scores. “Coping”, “reward”

167Sleep and Biological Rhythms (2021) 19:163–172

1 3

and “conformity” motives were positively associated with

the PSQI scores in the crude and fully adjusted models.

Variables with signiﬁcant correlation between the PFS

or PEMS subscales and the PSQI in total participants

remained signiﬁcant in both groups according to gender.

In males, higher current PFS scores were associated with

Table 1 General characteristics,

sleep attitudes and hedonic

hunger states of the university

students by gender

BMI Body Mass Index; MET: Metabolic Equivalent of Task; PEMS Palatable Eating Motives Scale; PFS

Power of Food Scale; PSQI Pittsburgh Sleep Quality Index

Numeric variables are presented as mean ± standard deviation (except for physical activity). Nominal vari-

ables are shown as percentage (frequency); p values were derived by independent t test and Pearson’s Chi

squared, respectively

a Data are presented as median (25 and 75th quartile) values; p values were derived by Mann Whitney U

test

b Deﬁned by the Pittsburgh Sleep Quality Index (PSQI). A global PSQI score > 5 indicates poor sleep qual-

ity

*p < 0.01; **p < 0.001

Characteristics Total (n = 1144) Females (n = 778) Males (n = 366) p value

Age 22.0 ± 2.7 21.8 ± 2.5 22.4 ± 3.1 0.003*

Smoking status, n (%)

Never 833 (72.8) 626 (80.5) 207 (56.6)

Former 97 (8.5) 55 (7.1) 42 (11.4) < 0.001**

Current 214 (18.7) 97 (12.4) 117 (32.0)

Alcohol consumption, % (n)

Non-drinker 1020 (89.2) 723 (92.9) 297 (81.1) < 0.001**

Yes 124 (10.8) 55 (7.1) 69 (18.9)

BMI (kg/m2)24.6 ± 3.6 24.2 ± 3.6 25.6 ± 3.4 < 0.001**

Normal 658 (57.5) 498 (64.0) 160 (43.7)

Overweight 397 (34.7) 223 (28.7) 174 (47.5) < 0.001**

Obese 89 (7.8) 57 (7.3) 32 (8.8)

Physical activity (MET-

mins)a (Median (25-75th)

742 (396–1386) 693 (330–1314) 1072 (532–2133) < 0.001**

Disease status, %(n)

No 981 (86.0) 641 (82.7) 340 (92.9)

Respiratory diseases 35 (3.1) 30 (3.9) 5 (1.4)

Endocrine diseases 27 (2.4) 25 (3.2) 2 (0.5)

Digestive diseases 24 (2.1) 21 (2.7) 3 (0.8)

Sleeping behaviour

Sleep qualityb6.1 ± 2.6 6.3 ± 2.6 5.8 ± 2.7 0.005*

Sleep quality ≤ 5 496 (43.4) 317 (40.7) 179 (48.9) 6.751

Sleep quality > 5 648 (56.6) 461 (59.3) 187 (51.1) 0.009*

Time in bed (clock time, h) 7.5 ± 1.5 7.5 ± 1.4 7.6 ± 1.6 0.253

Total sleep duration (h) 6.9 ± 1.4 6.9 ± 1.4 6.9 ± 1.4 0.748

Total sleep duration < 7h 474 (41.4) 313 (40.2) 161 (44.0) 1.448

Total sleep duration ≥ 7h 670 (58.6) 465 (59.8) 205 (56.0) 0.229

Hedonic hunger

Total PFS 47.9 ± 13.6 49.6 ± 13.3 44.5 ± 13.6 < 0.001**

Food available 17.5 ± 5.9 18.2 ± 5.9 16.1 ± 5.6 < 0.001**

Food present 14.0 ± 4.3 14.4 ± 4.2 12.9 ± 4.3 < 0.001**

Food tasted 17.0 ± 5.1 17.0 ± 4.9 15.4 ± 5.3 < 0.001**

Total PEMS 50.0 ± 13.7 50.3 ± 13.3 49.5 ± 14.4 0.362

Social motive 12.4 ± 4.3 12.4 ± 4.2 12.5 ± 4.3 0.954

Coping motive 13.3 ± 4.6 13.7 ± 4.6 12.5 ± 4.3 0.001*

Reward motive 14.0 ± 4.8 14.0 ± 4.8 13.8 ± 4.9 0.503

Conformity motive 9.6 ± 3.0 9.3 ± 2.8 10.0 ± 3.4 0.002*

168 Sleep and Biological Rhythms (2021) 19:163–172

1 3

worse sleep quality compared to females after adjustment

for confounders [β (95%CI) = 0.06, (0.04–0.08), p < 0.05

for males; β (95% CI) = 0.04, (0.02–0.05), p < 0.05 for

females], and coeﬃcient of determinations were found

to be 22.2% and 20.8%, respectively. When classiﬁed by

gender, the PEMS and PSQI scores remained signiﬁcant

after being fully adjusted in both groups. In both females

and males, there was a signiﬁcant relationship between

poorer sleep quality and higher “food present and food

tasted” (PFS subscales), “coping motive, reward motive

Table 2 Linear regression analysis of the relationship between scores of PFS or PEMS subscales and BMI by gender

BMI Body Mass Index; PEMS Palatable Eating Motives Scale; PFS Power of Food Scale; SE Standard Error

Beta coeﬃcient, standard error, 95% conﬁdence interval, R squared and p value calculated from linear regression analysis

*p < 0.01; **p < 0.001

Total Female Male

βSE 95% CI R2βSE 95% CI R2βSE 95% CI R2

PFS

Food available 0.18 0.01 0.15–0.22 0.094** 0.20 0.02 0.16–0.24 0.113** 0.22 0.03 0.16–0.28 0.136**

Food present 0.22 0.02 0.17–0.27 0.072** 0.24 0.02 0.19–0.30 0.085** 0.27 0.03 0.19–0.34 0.118**

Food tasted 0.14 0.02 0.10–0.18 0.041** 0.15 0.02 0.10–0.20 0.045** 0.17 0.03 0.11–0.24 0.078**

PFS total 0.07 0.01 0.06–0.09 0.084** 0.08 0.01 0.06–0.10 0.098** 0.09 0.01 0.06–0.11 0.136**

PEMS

Social motive 0.03 0.02 −0.01–0.08 0.002 0.04 0.03 −0.01–0.10 0.003 0.03 0.04 -0.05–0.11 0.001

Coping motive 0.21 0.02 0.16–0.25 0.075** 0.25 0.02 0.20–0.30 0.107** 0.18 0.03 0.11–0.26 0.063**

Reward motive 0.16 0.02 0.12–0.20 0.047** 0.15 0.02 0.10–0.20 0.041** 0.19 0.03 0.12–0.26 0.074**

Conformity motive 0.11 0.05 0.01–0.21 0.009 0.13 0.04 0.04–0.22 0.011*0.07 0.05 -0.02–0.17 0.006

PEMS total 0.05 0.01 0.04–0.07 0.046** 0.06 0.01 0.04–0.08 0.053** 0.05 0.01 0.02–0.07 0.044**

Table 3 Multiple linear regression analysis of associations between scores of PFS or PEMS subscales and PSQI by gender

PEMS Palatable Eating Motives Scale; PFS Power of Food Scale; PSQI Pittsburgh Sleep Quality Index; SE Standart Error

Beta coeﬃcient, standart error, 95% conﬁdence interval, R squared and p value calculated from a multiple linear regression analysis

a All models adjusted for age, BMI, physical activity and smoking

*p < 0.05

Total Female Male

β SE 95% CI R2 β SE 95% CI R2 β SE 95% CI R2

PFS

Food available Crude 0.04 0.01 0.01–0.08 0.007 0.02 0.02 −0.13–0.05 0.002 0.03 0.01 0.01–0.13 0.021

Adjusteda0.06 0.01 0.03–0.14 0.010 0.04 0.01 0.01–0.10 0.005 0.05 0.02 0.02–0.17 0.035

Food present Crude 0.19 0.01 0.16–0.23 0.101*0.17 0.02 0.12–0.21 0.075*0.23 0.03 0.17–0.30 0.143*

Adjusted 0.15 0.01 0.10–0.17 0.196*0.10 0.02 0.06–0.14 0.203*0.18 0.03 0.12–0.24 0.220*

PFS total Crude 0.06 0.01 0.05–0.07 0.112*0.05 0.01 0.04–0.07 0.085*0.08 0.01 0.06–0.10 0.156*

Adjusted 0.04 0.01 0.03–0.05 0.200*0.04 0.01 0.02–0.05 0.208*0.06 0.01 0.04–0.08 0.222*

PEMS

Social motive Crude 0.06 0.02 −0.01–0.10 0.006 0.03 0.02 −0.01–0.08 0.003 0.07 0.02 0.01–0.16 0.012

Adjusted 0.05 0.01 0.01–0.08 0.012 0.03 0.02 −0.01–0.07 0.010 0.08 0.03 0.02–0.14 0.015

Coping motive Crude 0.25 0.01 0.22–0.28 0.194*0.24 0.01 0.21–0.28 0.194*0.25 0.03 0.20–0.31 0.182*

Adjusted 0.21 0.02 0.18–0.24 0.269*0.19 0.02 0.16–0.22 0.281*0.22 0.03 0.16–0.27 0.270*

Reward motive Crude 0.19 0.01 0.16–0.22 0.125*0.17 0.01 0.14–0.21 0.107*0.22 0.02 0.17–0.28 0.163*

Adjusted 0.16 0.02 0.13–0.18 0.221*0.14 0.02 0.10–0.17 0.247*0.19 0.03 0.13–0.24 0.246*

Conformity motive Crude 0.25 0.02 0.21–0.30 0.087*0.26 0.03 0.20–0.33 0.082*0.27 0.03 0.19–0.34 0.115*

Adjusted 0.22 0.03 0.18–0.27 0.210*0.22 0.03 0.16–0.28 0.234*0.26 0.04 0.18–0.33 0.248*

PEMS total Crude 0.08 0.01 0.06–0.08 0.166*0.07 0.01 0.06–0.09 0.151*0.08 0.01 0.06–0.10 0.194*

Adjusted 0.06 0.01 0.05–0.07 0.256*0.06 0.01 0.05–0.07 0.266*0.07 0.01 0.06–0.09 0.286*

169Sleep and Biological Rhythms (2021) 19:163–172

1 3

and conformity motive” (PEMS subscales), but not “food

available” and “social motive”.

Table4 shows the subgroup analysis of multiple logistic

regression between the hedonic hunger score and sleep dura-

tion according to gender. After adjusting for potential con-

founding factors (age and BMI) in model 2, the odds of hav-

ing a high PEMS score increased in individuals with short

or long sleep duration, but not the PFS score. According to

model 2, participants who slept for less than 7h or more

than 9h had 40% higher odds for increased PEMS scores

compared to those who slept for 7–9h (reference group).

Moreover, poor sleep duration was associated with a 1.79

higher odds ratio for a high PEMS score in the male group.

In the female group, inadequate sleep duration was posi-

tively associated with the PEMS score in the crude model

or model 1 (adjusting for age and BMI), but the association

was not signiﬁcant in the fully adjusted models (model 2).

While there was no relationship between the PFS and sleep

duration in the female group, this association was found only

in the crude model of the male group (odds ratio (OR) (95%

CI) = 1.72 (1.09–2.72), p = 0.020).

Discussion

With a large sample of young people, this is the first

research to examine the relationship between sleep quality

and duration and hedonic hunger by potential confound-

ing. General results have shown that poor sleep quality

or inadequate sleep duration may be associated with high

hedonic hunger and most of its subdimensions.

As is known, the PFS and PEMS have been developed

to detect the factors aﬀecting the hedonic hunger impulse.

In particular, the PFS is used to assess the psychologi-

cal eﬀects of various delicious foods, while the PEMS

was developed to determine the reasons for consuming

delicious foods and drinks. In this study, hedonic hunger

scores were found statistically signiﬁcant between women

and men, especially in the PFS and its subcomponent

scores. Since eating disorders are more common in women,

most of the studies on hedonic hunger have been in women

and sometimes in both genders. In the majority of stud-

ies, it was determined that women are more sensitive to

Table 4 Logistic regression

models for the association

between sleep duration and

PEMS and PFS score according

to gender

PEMS Palatable Eating Motives Scale; PFS Power of Food Scale

All OR’s and 95% CI’s with p value were given for those with short or long sleep duration compared to

other with good sleep

Model 1: adjusted for age and BMI. Model 2: additionally adjusted for smoking and physical activity

Short or long sleep duration was deﬁned by less than 7h or over 9h

Good sleep duration for female and male group (n) = 422 and 186, respectively. Short or long sleep dura-

tion for female and male group (n) = 356 and 180, respectively

*p < 0.05; **p < 0.01; ***p < 0.001

Total pFemale pMale p

PFS

Crude