Effects of Acute Dance and Aerobic Exercise on

Drug Craving and Food Reward in Women with

CLINICAL SCIENCES
Methamphetamine Dependence
YU ZHOU1, GRAHAM FINLAYSON2, XUDONG LIU1, QICHEN ZHOU1, TIANZE LIU3, and CHENGLIN ZHOU1
1
School of Psychology, Shanghai University of Sport, Shanghai, PEOPLE’S REPUBLIC OF CHINA; 2School of Psychology,
Faculty of Medicine and Health, University of Leeds, Leeds, UNITED KINGDOM; and 3People’s Liberation Army Second
Military Medical University Naval Medical University, Department of Orthopedics, Changhai Hospital, Shanghai, PEOPLE’S
REPUBLIC OF CHINA
Downloaded from https://journals.lww.com/acsm-msse by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsIHo4XMi0hCywCX1AWnYQp/IlQrHD3i3D0OdRyi7TvSFl4Cf3VC1y0abggQZXdtwnfKZBYtws= on 01/11/2022

ABSTRACT
ZHOU, Y., G. FINLAYSON, X. LIU, Q. ZHOU, T. LIU, and C. ZHOU. Effects of Acute Dance and Aerobic Exercise on Drug Craving and
Food Reward in Women with Methamphetamine Dependence. Med. Sci. Sports Exerc., Vol. 53, No. 11, pp. 2245-2253, 2021. Introduction:
Drug dependence causes an overestimation of drug-related stimuli and an underestimation of non–drug-related stimuli, such as food. The pur-
pose of this study was to investigate the effects of acute moderate-intensity dance and aerobic exercise on drug craving, appetite, prefrontal
neural activation to food cues, and food reward in women with methamphetamine MA dependence. Methods: Thirty-nine women who
met the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition MA dependence criteria participated in the experiment and were
randomly assigned to either a dance (n = 20) or exercise (n = 19) group. A moderate-intensity (65%–75% maximum heart rate) 35-min dance
or treadmill intervention counterbalanced with a reading control session was conducted. After the intervention or control, subjective drug crav-
ing was measured before and after exposure to drug-related cues. Visual analog scales were used to measure subjective feelings of appetite.
Participants then completed a visual food cue paradigm while using functional near-infrared spectroscopy to monitor prefrontal blood oxygen
changes. Finally, the Leeds Food Preference Questionnaire was used to measure reward responses to different categories of food. Results: The
results showed that the dance and exercise interventions reduced subjective craving for drugs after being exposed to drug cues (P = 0.019).
Implicit wanting (P < 0.001) and relative preferences (P = 0.001) for high-calorie savory foods were all increased after interventions relative
to control. Compared with the control session, the left dorsolateral prefrontal cortex (P = 0.020) was activated when viewing high-calorie foods
after moderate-intensity aerobic exercise. Conclusions: The current results support the use of moderate-intensity exercise as a therapeutic
intervention to restore the balance between drug and nondrug rewards by decreasing cue-induced MA craving and increasing food reward.
Key Words: ADDICTION, DRUG REWARD, EXERCISE, FNIRS, NATURAL REWARD

D
rug addiction is a major problem worldwide, contrib- be attracted to MA for weight loss and to control symptoms
uting significantly to the global burden of disease. of depression (2). Addiction to MA is a major public health
According to the United Nations Office on Drugs problem; currently, there are no pharmacotherapies for MA
and Crime’s 2017 report, methamphetamine (MA) is the sec- use disorder that have been approved for use by the US
ond most commonly used drug worldwide (1). Unlike with Food and Drug Administration or the European Medicines
cocaine and heroin, where a high proportion of users are Agency (3). Longitudinal studies of abstinence show convincing
men, women use MA at rates almost equal to men. Surveys changes in the neurochemical markers (i.e., dopamine
among women suggest that they are more likely than men to transporter binding, glucose metabolism) and gray-matter
structure of MA-dependent subjects compared with control
subjects who were tested at similar intervals (4). Growing
Address for correspondence: Chenglin Zhou, Ph.D., School of Psychology, Shanghai evidence demonstrates that MA compromises prefrontal
University of Sport, 650 Qingyuan Ring Road, Yangpu District, Shanghai, cortex (PFC) activity, resulting in persisting cognitive function
200438, People’s Republic of China; E-mail: [email protected]. impairments, such as attentional and memory (5).
Submitted for publication January 2021.
Accepted for publication June 2021. Therefore, measures must be taken to diminish cognitive
Supplemental digital content is available for this article. Direct URL citations impairment during drug withdrawal. A number of studies have
appear in the printed text and are provided in the HTML and PDF versions reported a reduction in dopamine D2 receptors and a reduction
of this article on the journal’s Web site (www.acsm-msse.org). in dopamine release in the striatum of drug-dependent
0195-9131/21/5311-2245/0 humans, which can persist for months after detoxification; this
MEDICINE & SCIENCE IN SPORTS & EXERCISE® has been reported for a variety of addictive drugs (nicotine,
Copyright © 2021 by the American College of Sports Medicine cocaine, and alcohol) including MA (6). These persistent
DOI: 10.1249/MSS.0000000000002723 neuroadaptive changes may lead to reduced sensitivity to

2245

Copyright © 2021 by the American College of Sports Medicine. Unauthorized reproduction of this article is prohibited.
 nondrug reinforcers (7) and may even impair the ability to requirements for physical activity (25). This unique
respond adequately to non–drug-related rewards even during combination of properties makes dance a potentially powerful
abstinence (8). Individuals dependent on substances show a interventional approach. However, it is unclear whether dance
CLINICAL SCIENCES

decreased response in the striatum to natural rewards and interventions are equally as effective as other forms of
thus seem to search for alternative stimuli (drugs) to physical activity such as aerobic exercise.
maintain their equilibrium (6,9–11). Brain regions associated Therefore, the purpose of this study was to investigate the
with substance abuse and natural reward overlap (9,12), for effects of acute moderate-intensity dance and aerobic exercise
example, with the food reward region (13), which supports on changes in drug craving after drug-related cue exposure,
the hypothesis that drug dependence “hijacks” the natural food reward, and appetite in female MA-dependent subjects
reward pathways, leading to overestimation of drug-related using functional near-infrared spectroscopy (fNIRS) and be-
rewards and underestimation of nondrug-related rewards havioral approaches. Based on existing research and literature,
(14,15). Most neuroimaging studies of human addiction we hypothesized that moderate-intensity dance and aerobic
have focused on examining neural responses to drug-related exercise can reduce drug craving and increase food reward
cues. These studies have typically reported that drug-related and appetite in MA-dependent women.
cues trigger activations in brain regions such as the striatum
and medial PFC, which normally activate in response to METHODS
non–drug-related rewards in healthy subjects (8,14,16). To Participants
the degree that drugs and food activate common reward
circuitry in the brain, drugs offer powerful tools for Forty-four of 120 eligible volunteers were recruited from
understanding the neural circuitry that mediates food- the Drug Rehabilitation Bureau of the Mo Ganshan in
motivated habits and how this circuitry may be hijacked to Zhejiang province and were randomly assigned into either
cause appetitive behaviors to go awry (17). To support the dance or aerobic exercise group; 39 females completed
successful recovery from MA dependence, strategies are the entire trial (Fig. 1). Participants were eligible if they
needed to increase responsiveness to non–drug-related rewards met the criteria for MA dependence according to the Structured
such as food that could be measured by food-cue paradigm. Diagnostic Interview from the Diagnostic and Statistical Man-
Researchers have found that exercise can help to reduce ual of Mental Disorders, Fifth Edition; had been incarcerated
drug abuse (18). It has been shown that structured exercise and actively received detoxification treatment for no more than
training can ameliorate striatal dopamine D2/D3 receptor 3 months; were weight stable (<±3 kg change over last
deficits (19) and decreased MA craving and use in individuals 3 months); were age from 18 to 45 yr; had no history of gastro-
with MA dependence (20). Regular aerobic exercises could intestinal surgery or disorder; and were nondiabetic. Women do
improve global cognition in older adults with mild cognitive not exercise regularly in the drug rehabilitation center, and their
impairment, such as aerobic dance, a type of commonly designated meal, work, rest, and sleep times are very similar.
practiced aerobic exercise. In addition, it also benefits This study was conducted in accordance with the guidelines
executive function (21). However, its effects remain debatable laid down in the Declaration of Helsinki and was approved by
in improving cognitive function in patients with drug addiction the ethics committee of Shanghai University of Sport (No.
because the neurobiology of physical exercise in MA users is 102772019RT044). All participants gave their written informed
not fully elucidated and is essentially derived from animal consent before participating.
studies, and further clinical studies are profoundly needed to Study Design
confirm reproducibility of previous findings in humans and to
dissect the neurobiological basis of physical exercise benefits In a mixed design, each group (dance or exercise) completed ba-
(22). Dancing is increasingly used as an intervention, the sic data collection at baseline. Then drug craving measurements,
changes of rhythm and the spatial awareness of the body appetite, food reward, and neural response to food were taken after
during dance movement that activate place cells and grid intervention (dance or treadmill exercise) or after rest (crossover
cells (23) and delay the degeneration of nerve cells (24), design). Testing took place 1 wk apart in counterbalanced or-
making it a promising neuroplasticity-inducing tool (25). At der (Fig. 2). Eating was forbidden 2 h before testing.
the same time, dance has been established as a therapeutic tool
Dance Treatment
for the treatment of various populations such as Parkinson’s
disease (26); dementia (27); patients with serious mental illness Dance sessions were designed and directed by licensed in-
(28), especially people with stress (29) and depression (30); structors. The specially designed dance routine for this study
obese adults (31); and overweight children (32). As a was taught to the participants by the dance instructor to ensure
cognitively impaired population, research in drug-dependent that all participants in the dance group could perform the dance
samples should also explore whether dance has a beneficial correctly before the intervention. Participants performed this
effect. Compared with activities such as physical exercise or dance routine in synchronization with a musical phrase of
playing an instrument, dance comprises rhythmic motor eight rhythmical meters with a general value of 4/4 each. Each
coordination, balance and memory, emotions, social interaction, session started with a 5-min warm-up consisting of walking or
acoustic stimulation, and musical experience apart from its stretching exercises. After warming up, the core component of

2246 Official Journal of the American College of Sports Medicine http://www.acsm-msse.org

Copyright © 2021 by the American College of Sports Medicine. Unauthorized reproduction of this article is prohibited.
 CLINICAL SCIENCES
FIGURE 1—Flowchart of participant enrollment in the study.

the sessions focused on balance, coordination, and flexibility Subjective Measures of Craving
and lasted for 25 min. The intensity of dance training was
A visual analog scale (VAS) was used as a single-item,
intended to match with that of treadmill training. It was calculated
self-reported craving scale (34). The participants’ drug craving
as approximately 65%–75% of estimated maximum heart rate
assessments were required to mark a point on a 10-cm line,
(calculated as 206.9 − 0.67  age (33)), which was monitored
with the left end (0) indicating “not at all,” and the other end
using a Suunto smart sensor (Suunto Oy, Vantaa, Finland). All
(10) representing “extremely intense.” Participants were
patients had a cooling-off period after the dance, which included
instructed to pay close attention to the drug-related cues, and
slow shoulder movements, stepping, and breathing exercises.
then craving levels were scored before (baseline craving) and
after (cue-induced craving) drug-related cue exposure. The
Exercise Treatment cues, provided by the drug rehabilitation bureau, included
The supervised aerobic exercise training program was per- MA and the tools to use it, which are more effective than
formed using a treadmill that can monitor heart rate automati- drug-related images.
cally (Yijian JD618, Zhejiang, China). Participants were
instructed to warm-up for 5 min and to cool down for 5 min. Subjective Appetite Sensations
The experimental exercise period was 25 min. For this period, Subjective appetite sensations were measured immediately af-
the participant’s heart rate was adjusted within 65%–75% of ter intervention or rest by VAS ranging from “not at all” to “ex-
estimated maximum heart rate. tremely” on an electronic appetite self-rating system (35).
Participants rated sensations of desire to eat, fullness, and hunger.
Control Treatment
Visual Food Cue Paradigm
The participants in the control rest condition performed no
exercise. Instead, they sat in a quiet room for 35 min and read The visual food cue paradigm was adapted from a previous
about exercise- and fitness-related materials. study (36) using high-quality, full-color photographs with

FIGURE 2—Schematic diagram of study design. LFPQ, Leeds Food Preference Questionnaire; VAS, visual analog scale, to measure subjective appetite
sensations; VFCP, visual food cue paradigm.

EXERCISE, FOOD REWARD, AND MA DEPENDENCE Medicine & Science in Sports & Exercise® 2247

Copyright © 2021 by the American College of Sports Medicine. Unauthorized reproduction of this article is prohibited.
CLINICAL SCIENCES

FIGURE 3—Schematic representation of the visual food cue paradigm procedure. C indicates control; L, low-calorie food; H, high-calorie food; and +, fix-
ation cross.

fNIRS. Participants were asked to look at each image carefully (39). Motion artifacts were kept at a minimum by requiring
while undergoing fNIRS. The images were displayed on a participants to remain stationary during detection.
computer screen 30 cm from the participants’ eyes. The The fNIRS data were evaluated using Homer 2 software
control images were nonfood objects (such as flowers) from (MGH-Martinos Center for Biomedical Imaging, Boston,
nature. They were very similar in shape, color, and texture. MA) using MATLAB (MathWorks, Natick, MA). When the
Low-calorie food images were fruits and vegetables, whereas signal changed by more than 10% of the SD, motion artifacts
high-calorie food images were of fried chicken, ice cream, were detected within 0.5 s. In Homer 2, wavelet filtering was
and so on. According to the sequence shown in Figure 3, each used to detect and remove these artifacts (40). The fNIRS
image displays 7 alternating blocks of 10 food and nonfood signals were preprocessed: baseline drift was removed using a
images, respectively, for 3 s. The blocks were delimited by a high-pass filter with a cutoff frequency of 0.01 Hz, and a
10-s fixation cross (+). The whole scan lasted 270 s. low-pass filter with a frequency of 0.1 Hz was used to reduce
the impact on the heartbeat, respiration, blood pressure,
fNIRS Data Acquisition and Processing and skin blood flow signals. The changes of hemoglobin
concentration were calculated by the modified Beer–
In this study, a multichannel, continuous-wave, fNIRS in-
Lambert law. The average response of each participant to
strument (NIRScout; NIRx Medical Technologies LLC,
images at the 20 channels before and after exercise was
Minneapolis, MN) was used to record data. We acquired
obtained by block average. Because oxygenated hemoglobin
760- and 850-nm dual-wavelength near-infrared light to
signals have a better signal-to-noise ratio than deoxygenated
measure the relative concentration changes of oxyHb and
hemoglobin signals (41,42), only oxygenated hemoglobin
deoxyHb (37) based on the modified Beer–Lambert law (38),
concentration ([HbO]) data were used.
with a sampling frequency of 7.81 Hz. For the NIRS
experiment, eight sources and seven detectors (yielding 20
Food Reward Assessment
channels) were placed over the PFC region (Fig. 4B). Sensors
were positioned by aligning the bottom row of electrodes with Participants completed the Leeds Food Preference Question-
the International 10–20 sites AF7–Fp1–Fpz–Fp2–AF8 line naire (LFPQ) (43), a validated computer-based behavioral

FIGURE 4—Schematic arrangement of the near-infrared spectroscopy (NIRS) probe array (front view). The NIRS probe comprises eight sources (circular)
and seven detectors (rectangle). The midpoint of the source-detector distance is defined as the channel (Ch) location, labeled numerically (1–20) in the
schematic.

2248 Official Journal of the American College of Sports Medicine http://www.acsm-msse.org

Copyright © 2021 by the American College of Sports Medicine. Unauthorized reproduction of this article is prohibited.
procedure implemented with experimental software (E-prime TABLE 1. Participants’ demographic characteristics.
v.2.0; Psychology Software Tools, ND), which was adapted Treatment Group
and translated into Chinese, with reference to related literature Variables Total (n = 39) Dance (n = 20) Exercise (n = 19) t Test

CLINICAL SCIENCES
(44,45). In short, the LFPQ provides measures of implicit Demographic, M ± SD
Age (yr) 26.46 ± 4.08 26.45 ± 4.12 26.47 ± 4.15 0.99
wanting, relative preference, and explicit liking for a range of Height (cm) 161.67 ± 4.70 162.65 ± 4.51 160.63 ± 4.79 0.18
food images that vary along the dimensions of calorie content Weight (kg) 62.42 ± 6.66 63.75 ± 6.59 61.03 ± 6.61 0.21
BMI (kg·m−2) 23.87 ± 2.29 24.10 ± 2.40 23.64 ± 2.21 0.53
and sweet taste. A total of 16 different foods, divided into Marriage, n (%)
four categories (high-calorie savory (HCSA), low-calorie Married 6 (15.38) 1 (7.14) 5 (25.00) 0.58
savory (LCSA), high-calorie sweet (HCSW), and low-calorie Single 26 (65.38) 16 (78.57) 10 (50.00)
Divorce 4 (11.54) 1 (7.14) 3 (16.67)
sweet (LCSW)), formed the array for this study. Live together 3 (7.69) 1 (7.14) 2 (8.33)
Implicit wanting and relative preference. In a Education, n (%)
Junior college 3 (7.69) 1 (7.14) 2 (8.33) 0.20
forced-choice paradigm, each food image was randomly pre- Senior 11 (26.92) 6 (28.57) 5 (25.00)
sented with every other image in a series of food pair trials. Junior 21 (53.85) 11 (57.14) 10 (50.00)
The participants were instructed to select the food they “most Elementary 4 (11.54) 1 (7.14) 3 (16.67)
MA data, M ± SD
want to eat now” before each trial. The parameters were set as Usage (g per time) 0.70 ± 0.96 0.85 ± 1.27 0.52 ± 0.33 0.20
96 randomized food pair trials presented and each stimulus Frequency (d·wk−1) 4.32 ± 2.56 4.76 ± 2.70 3.81 ± 2.42 0.29
Duration (yr) 4.21 ± 1.83 4.15 ± 1.66 4.27 ± 2.10 0.88
appearing 12 times. A standardized implicit wanting score
for each food category was calculated as a function of the
mean reaction time in selecting a food category adjusted for Compliance with the intervention. Their average basic
the frequency of choice for each category (46). In addition, heart rate of the participants was 78.37 bpm (SD, 9.96). Their
the frequency of selections made in each category (with a estimate maximum heart rate was 185.91 bpm (SD, 4.73), and
possible range of 0–48) was measured. their targeted heart rate ranged from 120.84 bpm (SD, 3.08) to
Explicit liking. Explicit liking was measured from the 16 139.43 bpm (SD 3.55). They finished 2.58 km (SD, 0.19)
food stimuli presented one at a time and rated according to a treadmill exercise, and their energy expenditure was 142.12
100-mm VAS anchored at each end by the statements “not calories (SD, 13.72).
at all” and “extremely.” Subjects were prompted with the Drug craving. ANOVA was used with baseline and
statement “How pleasant would it be to experience a mouthful cue-induced craving separately, which show that the main effect
of this food now?” The VAS was presented on-screen beneath and interaction were not found at baseline, whereas cue-induced
each food stimulus, and subjects used the mouse to move a craving decreased after intervention (F(1,37) = 6.06, P = 0.019),
centered cursor along the line to indicate their response. and it was higher in the exercise group when compared with
the dance group (F(1,37) = 13.61, P = 0.001; Fig. 5).
Data Analysis Subjective sensations of appetite. There was a main
All statistical analyses were performed using IBM SPSS for effect of intervention in both fullness (F(1,37) = 10.53,
Windows (version 20; SPSS Inc., Chicago, IL). Normality of P = 0.002) and hunger (F(1,37) = 14.99, P < 0.001), but no main
data distribution was checked using the kurtosis and skewness effect of group or interaction effect. No significant effect of in-
of all variables. A mixed-design ANOVA was used with tervention, group, or their interaction was observed for the
group (dance and exercise) as a between-subject variable and subjective feelings of desire to eat (Figure, Supplemental Dig-
intervention (control and intervention) as within-subject vari- ital Content 1, Subjective feeling of appetite after the control
able, and with drug craving, appetite, PFC region activity, or the intervention session, http://links.lww.com/MSS/C352).
and food reward as dependent variables. Post hoc tests with Prefrontal neural activation. Repeated-measures
Bonferroni adjustments were used to determine where sig- ANOVA results revealed significant main effects and interac-
nificant differences existed. We corrected for multiple com- tions on the HbO concentration in 2 of 20 channels while par-
parisons of 20 fNIRS channel data using the Benjamini– ticipants were viewing high-calorie and low-calorie food
Hochberg false discovery rate procedure (47). Pearson images. After applying post hoc tests and controlling for
correlations were conducted between drug craving before multiple comparisons using the Benjamini–Hochberg false
and after drug-related cues and PFC region activity response discovery rate procedure (47) (Table, Supplemental Digital
to food images to determine their relationships. Values were Content 2, Mean changes in oxyhemoglobin concentration
expressed as means ± SD unless otherwise stated. (HbO) among individuals viewing high-calorie food following
moderate-intensity treadmill exercise (A) and dance (B) or
resting control sessions, http://links.lww.com/MSS/C353),
RESULTS one channel (Ch8) remained significant (Fig. 6) in the
Baseline characteristics of participants. Sample exercise group but not in the dance group, and this channel
characteristics and tests of baseline differences in social and was located over the left dorsolateral PFC (DLPFC) (Fig. 4).
demographic background are presented in Table 1. There were There was a significant main effect of intervention (F(1,37) = 4.20,
no significant differences between the dance (n = 20) and ex- P = 0.050) and a significant interaction (F(1,37) = 5.05, P = 0.032)
ercise (n = 19) groups on baseline characteristics. between intervention and group for Ch8. Furthermore, the

EXERCISE, FOOD REWARD, AND MA DEPENDENCE Medicine & Science in Sports & Exercise® 2249

Copyright © 2021 by the American College of Sports Medicine. Unauthorized reproduction of this article is prohibited.
CLINICAL SCIENCES

FIGURE 5—Drug craving before and after drug cue exposure between intervention and control sessions in two groups. Values are presented as means for
participants with SEM represented by vertical bars. *P < 0.05.

HbO concentration changes in the left DLPFC decreased cue exposure were positively correlated with DLPFC
significantly after treadmill exercise (P = 0.020) but not the (Ch8, Ch9, Ch10) activities when viewing high-calorie food
dance group (P = 0.880). (Table 2). We performed the correlation between carvings
Food reward. On comparison of the food reward mea- and neural activation in all prefrontal regions (Table, Supple-
sures after intervention compared with control, the ANOVA mental Digital Content 4, Correlations between drug craving
showed that intervention resulted in a greater implicit wanting (before and after drug-related cue exposure) and neural re-
(F(1,37) = 22.53, P < 0.001) and relative preference (F(1,37) = 13.57, sponse to high-calorie food at baseline, http://links.lww.com/
P = 0.001) for HCSA, whereas it was a lower relative prefer- MSS/C355) and found our results inconclusive.
ence (F1,37) = 7.25, P = 0.011) for HCSW. An interaction ef-
fect between intervention and group (F(1,37) = 5.60,
P = 0.019) was founded in explicit liking for LCSA, which
DISCUSSION
showed that compared with the dance intervention, treadmill This study examined the acute effects of moderate-intensity
exercise increased explicit liking for LCSA (P = 0.009). No dance and aerobic exercise on drug craving, appetite, neural
other significant differences between intervention compared response to foods, and food reward in women with MA depen-
with control, intensity, or interactions between intervention dence. Our findings suggest that cue-induced MA craving was
and group were noted in the food reward variables for any significantly reduced after dance and aerobic exercise, and that
food category (Figure, Supplemental Digital Content 3, Im- both groups increased subjective sensations of hunger and de-
plicit wanting (A), relative preference (B) and explicit liking creased fullness. Furthermore, HbO concentration changes of
(C) for four categories of food between intervention and con- the left DLPFC when viewing images of high-calorie foods
trol sessions grouped by dance and exercise, http://links.lww. was decreased after moderate-intensity aerobic exercise but
com/MSS/C354). not after the dance intervention compared with the control ses-
Correlation between cravings and prefrontal neu- sion. After intervention, both groups increased their implicit
ral activation. Drug cravings before and after drug-related wanting for HCSA and reduced their relative preference for

FIGURE 6—Changes in oxyhemoglobin concentration (HbO) for channel 8 (Ch8) in both dance (A) and exercise (B) groups after intervention or control
sessions during visual food cue paradigm when viewing images of high-calorie foods.

2250 Official Journal of the American College of Sports Medicine http://www.acsm-msse.org

Copyright © 2021 by the American College of Sports Medicine. Unauthorized reproduction of this article is prohibited.
TABLE 2. Correlations between drug craving (before and after drug-related cue exposure) activation in female users of MA was significantly lower
and neural response to high-calorie food in DLPFC after rest.
when viewing pictures of high-calorie foods after the
fNIRS Drug Craving
intervention in the aerobic exercise group. Previous researches

CLINICAL SCIENCES
Before Cue Exposure After Cue Exposure
in men who were MA dependent also found that moderate-
Area Channel r P r P
intensity exercise increased their implicit wanting and relative
2 0.129 0.441 0.051 0.760
5 0.113 0.491 0.049 0.766 preference for HCSA foods (45); however, it increased
8 0.378* 0.021 0.334* 0.043 orbital frontal cortex (OFC) but not decreased DLPFC
9 0.324* 0.047 0.283 0.085
10 0.352* 0.030 0.340* 0.037
activity associated with high-calorie food images (53). The
15 0.253 0.131 0.229 0.172 DLPFC has previously been shown to promote satiety by
17 0.050 0.770 −0.004 0.980 transmitting inhibition signals to areas of the brain that
*P < 0.05. stimulate appetite, such as the insular cortex and OFC (54).
DLPFC, dorsolateral prefrontal cortex.
Le et al. (55) suggested that the decrease of DLPFC
activation might be related to an increase in OFC activation,
HCSW. The exercise group also showed a significant increase food consumption, enhanced food reward, and weight gain.
in their preference for LCSA compared with the dance group. Drug craving is positively correlated with DLPFC activa-
These results were consistent with our hypothesis that the ex- tion when viewing high-calorie foods, which indicates that
ercise intervention would decrease drug craving and increase people with strong drug cravings are better at resisting
food reward and appetite, suggesting an improvement in re- high-calorie food cues, leading to less desire to seek and con-
ward regulation. sume high-calorie foods. This result supports the hypothesis
It was also found that craving in both groups was signifi- that individuals with substance use disorder become more re-
cantly reduced after exposure to the drug cue after interven- sponsive to drug cues and less responsive to natural rewards
tion. This is consistent with previous studies showing that an (56). At the same time, the decreased DLPFC activation to
acute exercise intervention can reduce MA drug craving high-calorie food images after exercise was accompanied by
(48). It was also found that MA use decreased in individuals the decreased drug craving before and after cue exposure,
with MA dependence after exercise (20). One potential which could be interpreted that exercise simultaneously
explanation for the reduction in craving after exercise is the reduced drug craving and inhibition control for high-calorie
transient hypofrontality hypothesis, which involves the food, which led to less responsive to drug cues and more
availability of mental and physical resources (49). We also responsive to food rewards. Moreover, this result supports
found that cue-induced craving seems higher in the exercise the potential utility of evaluating food reward outcomes in
group when compared with the dance group, although we clinical research in order to evaluate treatment outcomes for
have divided the participants randomly and there were no MA use disorder (57). As an effective natural reward, food
differences between the two groups’ baseline information. has been used to assess the sensitivity and function of the
The MA data were collected from the participants by reward system in drug addiction (58,59). As we can see that
recalling their MA use experience and were not accurate; the correlations between drug cravings and DLPFC activation
therefore, more objective tools should be used to measure were not that strong (r < 0.5), which may be caused by that
their MA craving to check the results of grouping, and we did not separate the HCSA and sweet foods, because we
enrolling more participants may diminish these differences. have found that physical activity increased the preference for
The study found that moderate-intensity aerobic exercise in- HCSA foods but decreased the preference for sweet foods in
creased hunger, but not fullness, after intervention; both women with MA dependence.
groups increased their implicit wanting and relative preference Palatable food activates brain reward circuitry through fast
for HCSA foods. In the present study, women’s relative pref- sensory inputs and through slow postingestive consequences
erence for HCSW foods was also decreased. In opiate users, (such as raising glucose concentration in blood and brain),
previous research has reported strong craving, preference for, whereas drugs activate these same pathways mostly through
and intake of sugary food stuffs either when maintained on their direct pharmacological effects on the reward circuitry
opiate agonists or when abstinent (50). The separation occurs (17), which leads the hypothesis this population may be
because sweet foods, compared with high-calorie foods, more susceptible to transpose the behavior of substance use
have a rewarding effect over addictive drugs (13). This disorder to overeating and inducing obesity when they
finding is contrary to existing research, where exercise cannot get drugs in the drug rehabilitation bureau.
was found to acutely decrease the relative preference for We found that the benefits of the two kinds of intervention
high-calorie versus low-calorie foods (51). The inconsistency are different, with aerobic exercise stimulating appetite and re-
in results could be attributable to different baseline food ducing drug craving before and after exposure compared with
reward responses between individuals who are dependent on the dance group. Compared with activities such as aerobic exer-
drugs and healthy people. Neural activity in the brain results cise on a treadmill, dance involves rhythmic motor coordina-
in increased glucose and oxygen consumption from local tion, balance, and musical experience in addition to its
capillary beds; thus, the HbO changes are index of neural requirements for physical activity (60), which may contribute
activity (52). Compared with the dance group, the DLPFC to this phenomenon. At the same time, a previous study found

EXERCISE, FOOD REWARD, AND MA DEPENDENCE Medicine & Science in Sports & Exercise® 2251

Copyright © 2021 by the American College of Sports Medicine. Unauthorized reproduction of this article is prohibited.
 that dance improved physical fitness (such as systolic and (64). Blood flow on the skin may be potential confounding
diastolic blood pressure, peak VO2, motor ability) and quality variables, which could be avoided by short-separation technique.
of life, but not more than other exercise interventions (61). In Finally, appetite changes were not measured over time after
CLINICAL SCIENCES

addition, another study showed that music training may have the intervention. We only took a single rating of hunger,
stimulatory effects on sensorimotor and auditory systems, which desire to eat, and fullness at the end of each condition, and no
may enhance cortical plasticity (62). These neuroplasticity follow-up on food intake was conducted to validate our
as well as molecular and biological changes may require behavioral and brain activation findings. Because of these
long-term interventions to develop. limitations, further randomized controlled trials with improved
In this study, there are some important limitations to con- study designs and larger sample sizes should be conducted.
sider. First, it was not possible to keep heart rate at a consistent To conclude, the present study reports the novel finding that
moderate-intensity during the dance process (60), so it was acute moderate-intensity dance and aerobic exercise may de-
difficult to accurately control the exercise intensity, which crease cue-induced MA craving and improve food reward
may have contributed to why the dance group did not and appetite responses in women with MA dependence. In
achieve the same neural effect as the exercise group. At the particular, short-term moderate-intensity aerobic exercise
same time, it is hard to compare the conditions without data may restore sensitivity to food reward by decreasing left
of exercise efforts (e.g., Borg scale, lactate concentrations, DLPFC activity among people who are MA dependent. Our
average heart rate, percentage of V̇O2 max). Second, fNIRS findings add further support to the suggestion that treatment
can only measure the activation of the prefrontal region, but of MA addiction should aim to increase the reward value of
not more internal structures such as the striatum, which non–drug-related reinforcers (6) and highlight the importance
would enrich our research results. Third, we did not separate of specific exercise programs for therapeutic interventions.
the sweet and savory caloric foods because of its original
paradigm (36); the food images we used in the visual food
The authors acknowledge the support by the National Social Sci-
cue paradigm could be divided into four categories like ence Foundation of China (grant number 17ZDA330).
LFPQ in the future to help us find the relationships between The authors have no conflict of interest to declare. Results of the
behavior and DLFC activation. Fourth, besides wavelet present study do not constitute endorsement by the American College
of Sports Medicine. Results of the study are presented clearly,
filters (63), general linear models could also be used to avoid honestly, and without fabrication, falsification, or inappropriate
bias while applying the bandpass filter as a separate step data manipulation.

REFERENCES
1. United Nations Office on Drugs and Crime (UNODC), World Drug 12. Karama S, Lecours AR, Leroux JM, et al. Areas of brain activation in
Report. 2017 United Nations Publication. Vienna: UNODC; 2017. males and females during viewing of erotic film excerpts. Hum Brain
2. Rawson RA. Current research on the epidemiology, medical and psy- Mapp. 2002;16(1):1–13.
chiatric effects, and treatment of methamphetamine use. J Food Drug 13. Dileone RJ, Taylor JR, Picciotto MR. The drive to eat: comparisons
Anal. 2013;21(4):S77–81. and distinctions between mechanisms of food reward and drug addic-
3. Chan B, Freeman M, Kondo K, et al. Pharmacotherapy for tion. Nat Neurosci. 2012;15(10):1330–5.
methamphetamine/amphetamine use disorder—a systematic review 14. Diekhof EK, Falkai P, Gruber O. Functional neuroimaging of reward
and meta-analysis. Addiction. 2019;114(12):2122–36. processing and decision-making: a review of aberrant motivational
4. Dean AC, Groman SM, Morales AM, London ED. An evaluation of and affective processing in addiction and mood disorders. Brain
the evidence that methamphetamine abuse causes cognitive decline Res Rev. 2008;59(1):164–84.
in humans. Neuropsychopharmacology. 2013;38(2):259–74. 15. Feltenstein MW, See RE. The neurocircuitry of addiction: an over-
5. Bernheim A, See RE, Reichel CM. Chronic methamphetamine view. Br J Pharmacol. 2010;154(2):261–74.
self-administration disrupts cortical control of cognition. Neurosci 16. Heinz A, Siessmeier T, Wrase J, et al. Correlation between dopamine
Biobehav Rev. 2016;69:36–48. D(2) receptors in the ventral striatum and central processing of alco-
6. Volkow ND, Fowler JS, Wang GJ, Baler R, Telang F. Imaging dopa- hol cues and craving. Am J Psychiatry. 2004;161(10):1783–9.
mine’s role in drug abuse and addiction. Neuropharmacology. 2009; 17. Volkow N, Wise R. How can drug addiction help us to understand
56:3–8. obesity? Nat Neurosci. 2005;26:5160–6.
7. Koob GF, Volkow ND. Neurocircuitry of addiction. Neuropsycho- 18. Smith MA, Lynch WJ. Exercise as a potential treatment for drug
pharmacology. 2010;35:217–38. abuse: evidence from preclinical studies. Front Psychiatry. 2012;
8. Wrase J, Schlagenhauf F, Kienast T, et al. Dysfunction of reward pro- 2:82.
cessing correlates with alcohol craving in detoxified alcoholics. 19. Robertson CL, Ishibashi K, Chudzynski J, et al. Effect of exercise
Neuroimage. 2007;35(2):787–94. training on striatal dopamine D2/D3 receptors in methamphetamine
9. Garavan H, Pankiewicz J, Bloom A, et al. Cue-induced cocaine crav- users during behavioral treatment. Neuropsychopharmacology. 2016;
ing: neuroanatomical specificity for drug users and drug stimuli. Am J 41(6):1629–36.
Psychiatry. 2000;157(11):1789–98. 20. Rawson RA, Chudzynski J, Mooney L, et al. Impact of an exercise
10. Koob GF, Le Moal M. Drug addiction, dysregulation of reward, and intervention on methamphetamine use outcomes post-residential
allostasis. Neuropsychopharmacology. 2001;24(2):97–129. treatment care. Drug Alcohol Depend. 2015;156:21–8.
11. Paulus MP, Tapert SF, Schuckit MA. Neural activation patterns of 21. Zhu Y, Zhong Q, Ji J, et al. Effects of aerobic dance on cognition in
methamphetamine-dependent subjects during decision making pre- older adults with mild cognitive impairment: a systematic review and
dict relapse. Arch Gen Psychiatry. 2005;62(7):761–8. meta-analysis. J Alzheimers Dis. 2020;74(2):679–90.

2252 Official Journal of the American College of Sports Medicine http://www.acsm-msse.org

Copyright © 2021 by the American College of Sports Medicine. Unauthorized reproduction of this article is prohibited.
22. Morais APD, Pita IR, Fontes-Ribeiro CA, Pereira FC. The neurobio- Preference Questionnaire in human experimental research. Food
logical mechanisms of physical exercise in methamphetamine addic- Qual Prefer. 2020;80:10.
tion. CNS Neurosci Ther. 2018;24(2):85–97. 45. Zhou Y, Lu Y, Jin X, Liu J, Finlayson G, Zhou C. Effects of

CLINICAL SCIENCES
23. Doeller CF, Barry C, Burgess N. Evidence for grid cells in a human moderate- and high-intensity acute aerobic exercise on food reward
memory network. Nature. 2010;463(7281):657–61. and appetite in individuals with methamphetamine dependence.
24. Hafting T, Fyhn M, Molden S, Moser M-B, Moser EI. Microstructure of Physiol Behav. 2019;211:112649.
a spatial map in the entorhinal cortex. Nature. 2005;436(7052):801–6. 46. Dalton M, Finlayson G. Psychobiological examination of liking and
25. Fong Yan A, Cobley S, Chan C, et al. The effectiveness of dance in- wanting for fat and sweet taste in trait binge eating females. Physiol
terventions on physical health outcomes compared to other forms of Behav. 2014;136:128–34.
physical activity: a systematic review and meta-analysis. Sports Med. 47. Benjamini Y, Hochberg Y. Controlling the false discovery rate: a
2018;48(4):933–51. practical and powerful approach to multiple testing. J Royal Stat
26. Abraham A, Hart A, Andrade I, Hackney ME. Dynamic neuro-cognitive Soc. 1995;57(1):289–300.
imagery improves mental imagery ability, disease severity, and mo- 48. Wang D, Zhou C, Zhao M, Wu X, Chang YK. Dose-response rela-
tor and cognitive functions in people with Parkinson’s disease. Neu- tionships between exercise intensity, cravings, and inhibitory control
ral Plast. 2018;2018:6168507. in methamphetamine dependence: an ERPs study. Drug Alcohol De-
27. Hokkanen L, Rantala L, Remes AM, Härkönen B, Viramo P, Winblad I. pend. 2016;161:331–9.
Dance and movement therapeutic methods in management of dementia: 49. Dietrich A, Audiffren M. The reticular-activating hypofrontality (RAH)
a randomized, controlled study. J Am Geriatr Soc. 2008;56(4):771–2. model of acute exercise. Neurosci Biobehav Rev. 2011;35(6):1305–25.
28. Hackney ME, Earhart GM. Social partnered dance for people with se- 50. Morabia A, Fabre J, Chee E, Zeger S, Orsat E, Robert A. Diet and opiate
rious and persistent mental illness: a pilot study. J Nerv Ment Dis. addiction: a quantitative assessment of the diet of non-institutionalized
2010;198(1):76–8. opiate addicts. Br J Addict. 2010;84(2):173–80.
29. Braeuninger I. Dance movement therapy group intervention in stress 51. Mcneil J, Cadieux S, Finlayson G, Blundell JE, Doucet É. The effects
treatment: a randomized controlled trial (RCT). Arts In Psychother. of a single bout of aerobic or resistance exercise on food reward. Ap-
2012;39(5):443–50. petite. 2015;84:264–70.
30. Röhricht F, Papadopoulos N, Priebe S. An exploratory randomized 52. Irani F, Platek SM, Bunce S, Ruocco AC, Chute D. Functional near
controlled trial of body psychotherapy for patients with chronic de- infrared spectroscopy (fNIRS): an emerging neuroimaging technol-
pression. J Affect Disord. 2013;151(1):85–91. ogy with important applications for the study of brain disorders. Clin
31. Allet L, Müller-Pinget S, Punt I, et al. Dance therapy combined with Neuropsychol. 2007;21:9–37.
patient education improves quality of life of persons with obesity: a 53. Wang H, Chen Y, Li X, Wang J, Zhou Y, Zhou C. Moderate-intensity
pilot feasibility study for a randomised controlled trial. Obes Res Clin aerobic exercise restores appetite and prefrontal brain activity to images
Pract. 2017;11(1):79–87. of food among persons dependent on methamphetamine: a functional
32. Murphy EC, Carson L, Neal W, Baylis C, Donley D, Yeater R. Ef- near-infrared spectroscopy study. Front Hum Neurosci. 2019;13:400.
fects of an exercise intervention using Dance Dance Revolution on 54. Small DM. Changes in brain activity related to eating chocolate: from
endothelial function and other risk factors in overweight children. pleasure to aversion. Brain. 2001;124(9):1720–33.
Int J Pediatr Obes. 2009;4(4):205–14. 55. Le DS, Pannacciulli N, Chen K, et al. Less activation of the left dor-
33. Gellish RL, Goslin BR, Olson RE, McDonald A, Russi GD, Moudgil solateral prefrontal cortex in response to a meal: a feature of obesity.
VK. Longitudinal modeling of the relationship between age and max- Am J Clin Nutr. 2006;84(4):725–31.
imal heart rate. Med Sci Sports Exerc. 2007;39(5):822–9. 56. Volkow ND, Morales M. The brain on drugs: from reward to addic-
34. Sayette MA, Shiffman S, Tiffany ST, Niaura RS, Martin CS, Shadel WG. tion. Cell. 2015;162(4):712–25.
The measurement of drug craving. Addiction. 2000;(95, Suppl 2):S189–210. 57. Banks ML. Utility of preclinical drug versus food choice procedures
35. Gibbons C, Caudwell P, Finlayson G, King N, Blundell J. Validation of a to evaluate candidate medications for methamphetamine use disor-
new hand-held electronic data capture method for continuous monitoring der. Ann N Y Acad Sci. 2017;1394(1):92–105.
of subjective appetite sensations. Int J Behav Nutr Phys Act. 2011;8:57. 58. Czoty PW, Stoops WW, Rush CR. Evaluation of the “pipeline” for
36. Killgore WD, Young AD, Femia LA, Bogorodzki P, Rogowska J, development of medications for cocaine use disorder: a review of
Yurgelun-Todd DA. Cortical and limbic activation during viewing of translational preclinical, human laboratory, and clinical trial research.
high- versus low-calorie foods. Neuroimage. 2003;19(4):1381–94. Pharmacol Rev. 2016;68(3):533–62.
37. Yamashita Y, Maki A, Ito Y, Watanabe E, Mayanagi Y, Koizumi H. 59. Banks ML, Hutsell BA, Blough BE, Poklis JL, Negus SS. Preclinical
Noninvasive near-infrared topography of human brain activity using assessment of lisdexamfetamine as an agonist medication candidate
intensity modulation spectroscopy. Opt Eng. 1996;35(4):1046–9. for cocaine addiction: effects in rhesus monkeys trained to discrimi-
38. Cope M, Delpy DT, Reynolds EOR, Wray S, Wyatt J, Zee PVD. nate cocaine or to self-administer cocaine in a cocaine versus food
Methods of quantitating cerebral near infrared spectroscopy data. choice procedure. Int J Neuropsychopharmacol. 2015;18(8):pyv009.
Adv Exp Med Biol. 1988;222:183. 60. Kattenstroth J-C, Kalisch T, Holt S, Tegenthoff M, Dinse HR. Six
39. Jurcak V, Tsuzuki D, Dan I. 10/20, 10/10, and 10/5 systems revisited: months of dance intervention enhances postural, sensorimotor, and
their validity as relative head-surface-based positioning systems. cognitive performance in elderly without affecting cardio-respiratory
Neuroimage. 2007;34(4):1600–11. functions. Front Aging Neurosci. 2013;5.
40. Molavi B, Dumont GA. Wavelet-based motion artifact removal for func- 61. Koch SC, Riege RFF, Tisborn K, Biondo J, Martin L, Beelmann A.
tional near-infrared spectroscopy. Physiol Meas. 2012;33(2):259–70. Effects of dance movement therapy and dance on health-related psycho-
41. Strangman G, Culver JP, Thompson JH, Boas DA. A quantitative logical outcomes. A meta-analysis update. Front Psychol. 2019;10:1806.
comparison of simultaneous BOLD fMRI and NIRS recordings dur- 62. Lappe C, Herholz SC, Trainor LJ, Pantev C. Cortical plasticity
ing functional brain activation. Neuroimage. 2002;17(2):719–31. induced by short-term unimodal and multimodal musical train-
42. Hoshi Y. Functional near-infrared optical imaging: utility and limitations ing. J Neurosci. 2008;28(39):9632–9.
in human brain mapping. Psychophysiology. 2003;40(4):511–20. 63. Jang KE, Tak S, Jung J, Jang J, Jeong Y, Ye JC. Wavelet minimum
43. Finlayson G, King N, Blundell J. The role of implicit wanting in re- description length detrending for near-infrared spectroscopy.
lation to explicit liking and wanting for food: implications for appetite J Biomed Opt. 2009;14:034004.
control. Appetite. 2008;50(1):120–7. 64. Huppert TJ. Commentary on the statistical properties of noise and its
44. Oustric P, Thivel D, Dalton M, et al. Measuring food preference and implication on general linear models in functional near-infrared spec-
reward: application and cross-cultural adaptation of the Leeds Food troscopy. Neurophotonics. 2016;3(1):010401.

EXERCISE, FOOD REWARD, AND MA DEPENDENCE Medicine & Science in Sports & Exercise® 2253

Copyright © 2021 by the American College of Sports Medicine. Unauthorized reproduction of this article is prohibited.