A PILOT STUDY OF SHORT-TERM EFFECTS
OF NON-SEDENTARY BEHAVIOUR ON MOOD CHANGES
Aleksandar Matic, Venet Osmani, Andrei Popleteev and Oscar Mayora-Ibarra
CREATE-NET, Via alla Cascata 56/D, 38133, Trento, Italy
Keywords: Sedentary behaviour, Accelerometers, Mood changes, Mental health, Pervasive computing.
Abstract: Sedentary behaviour is considered to be a major factor impacting a number of health outcomes while use of
technology in the workplace is increasingly contributing to the amount of sedentary time. Even if guidelines
on the amount of physical activity are followed outside of workplace, sedentary work style still has a
deleterious impact on health. There have been a number of studies that investigated the impact of sedentary
behaviour on general health while no study has reported investigation of effects of sedentary behaviour on
mood changes. The aim of this pilot study, which included 14 knowledge workers, was to objectively
measure sedentary time using accelerometers and investigate correlation with the self-reported mood
changes in short term. The results show that sedentary behaviour has negative effects on the mood even
over a short period of one day. Although a pilot study, this is the first study to examine short term
psychological implications of sedentary behaviour. Randomised controlled experiments are needed to
further clarify these findings.
1 INTRODUCTION
Conveniences of modern life in developed countries
are having an adverse effect on health. This is
particularly manifested through decreased physical
activity, where people spend significant amount of
time in sedentary activities. Physical inactivity leads
to a number of health outcomes and various media
campaigns are designed to encourage increase in
physical activity levels and promote healthy
lifestyle. However, a general increase in physical
activities outside work hours is not sufficient to
compensate for deleterious effects of sedentary time
in workplaces. Sedentary work style has adverse
health risks, where negative effects on health occur
even if people follow the guidelines on physical
activity outside of the workplace (Tremblay et al.,
2010; Healy et al., 2008). As a number of studies
have shown (Sanchez-Villegas et al., 2008;
Tremblay et al., 2010; Hamilton et al., 2007;
Hamburg et al., 2007; Zwart et al., 2007; Hu et al.,
2003; Healy et al., 2008) prolonged sitting leads to
an array of health complications, including diabetes,
high blood pressure and obesity.
Office workers typically adhere to a work style
that requires sitting for prolonged periods of time.
Therefore, monitoring sedentary patterns in
workplace provides a good basis for polices that
encourage healthy work style and minimize the time
spent in sedentary activities. In addition to physical
health effects, sedentary behaviour also has a
negative effect on mental health. The closest study
to our work has linked prolonged sitting time with
mental health disorders (Sanchez-Villegas et al.,
2008), however, no study has been reported that has
investigated correlation between sedentary patterns
and mood changes.
The aim of our pilot study was to investigate
how sedentary work style affects office workers’
subjectively reported mood. Measuring mood
changes is important when considering mood shifts
are linked to serious diseases, such as depression, bi-
polar disease but also cost economies an average of
$44 billion per year in Lost Productive Time (LPT)
(Stewart et al., 2003).
In our study we used well-established
technologies to monitor sitting time, namely mobile
phones with embedded accelerometers. Using
mobile phones as a sensing device not only allows
for precise measurements of physical activities and
sedentary periods (thus overcoming recall biases and
floor effects (Tudor-Locke and Myers, 2001)), but
also provides an unobtrusive monitoring platform
that does not interfere with the typical office
299
Matic A., Osmani V., Popleteev A. and Mayora-Ibarra O..
A PILOT STUDY OF SHORT-TERM EFFECTS OF NON-SEDENTARY BEHAVIOUR ON MOOD CHANGES.
DOI: 10.5220/0003874702990304
In Proceedings of the International Conference on Health Informatics (HEALTHINF-2012), pages 299-304
ISBN: 978-989-8425-88-1
Copyright
c
2012 SCITEPRESS (Science and Technology Publications, Lda.)
workers’ routines. In this manner, we were able to
measure precisely sedentary behaviour at workplace
and correlate the findings with the reported mood
changes. The study described in this paper is the first
to have investigated this issue and we hope it will
motivate further research in the implications of
sedentary behaviour on psychological response.
2 BACKGROUND AND RELATED
WORK
Sedentary behaviour (from the Latin “sedere”
meaning “sitting”) refers to the category of
behaviours for which the energy expenditure is low,
including prolonged sitting periods at work, home,
in car, bus, and in leisure time (Tremblay et al.,
2010). Sedentary behaviour, characterized by the
SITT formula (Sedentary behaviour frequency,
Intensity, Time and Type), is often associated with
adverse health outcomes. Emerging studies suggest
that independently and qualitatively different from a
lack of moderate or intense physical activity,
sedentary behaviour has effects on human
metabolism, physical function and health outcomes
and therefore it should be considered as a separate
and unique phenomenon (Tremblay et al., 2010).
Even if one meets physical activity guidelines,
sitting for prolonged periods increases the risks for
obesity, metabolic syndrome, type 2 diabetes and
cardiovascular disease (Tremblay et al., 2010;
Hamilton et al., 2007).
Previous research initiatives have relied on self-
reporting methods to acquire information about
activities that correspond to sitting routines. Self-
reporting of physical activity have a number of
drawbacks including floor effects and recall bias
(Tudor-Locke and Myers, 2001) while continuous
monitoring over long periods of time may disturb
habitual physical activity (Baranowski, 1988). On
the other hand, accelerometers can provide
recognition of physical activities such as sitting,
running, standing (Kwapisz et al., 2010; Mathie et
al., 2004) and in addition record further details such
as duration, frequency and intensity of movements.
Accelerometers have become critical in investigation
of sedentary habits that are difficult to recall with the
questionnaire method. With the advancement of
technology, accelerometers have been used to
capture sedentary patterns in an objective manner,
thus overcoming the drawbacks of self-reporting.
Usage of accelerometers allowed significant results
in establishing the influence of sedentary behaviour
on metabolism (Hamburg et al., 2007), vascular
health (Hamburg et al., 2007) and bone mineral
content (Zwart et al., 2007). Moreover, recent
studies link sedentary lifestyle and obesity (Hu et al.,
2003), cancer (Gierach et al., 2009) and
psychosocial health (Martin et al., 2009). Sanchez-
Villegas et al. (Sanchez-Villegas et al., 2008) found
that the risk of mental disorders was 31% higher for
prolonged patterns of sitting in front of a computer
or a TV, comparing subjects that spend more than
42h/week in front of a TV with those watching less
than 10.5h/week. The extensive literature survey on
the implications of sedentary lifestyle, provided by
Tremblay et al. (Tremblay et al., 2010), concludes
that there is a need to understand the factors of
sedentary behaviour and to implement interventions
to reduce population-wide levels of sedentary
behaviour, shifting some proportion of time from
sitting to various types of physical activity. The
authors (Tremblay et al., 2010) suggest that this
should be done empirically. By conducting an
empirical study that investigates non-sitting periods
during working time, we have studied the correlation
between sedentary patterns at workplace and office
workers’ mood changes, which may form the basis
for strategies to reduce prolonged sitting periods.
Our work is the first to investigate the influence of
sedentary behaviour on mood changes, opening up a
research avenue to explore psychological effects of
increasing prevalence of sedentary habits.
3 MOTIVATION
Examining the influence of sedentary behaviour on
psychological responses of an individual stemmed
from an immense research body that was focused on
physiological implications of sedentary lifestyle
(Sanchez-Villegas et al., 2008; Tremblay et al.,
2010; Hamilton et al., 2007; Hamburg et al., 2007;
Zwart et al., 2007; Hu et al., 2003; Healy et al.,
2008). The postulates of physiological psychology
that interrelate physiological and psychological
processes (Andreassi, 2006) prompted us to
investigate whether prolonged periods of sitting may
also cause psychological responses.
The current literature provides number of
directions for the possible correlation between
sedentary behaviour and psychological reactions:
Effects of sedentary behavior on insulin
sensitivity are extensively examined (Hamburg
et al., 2007; Balkau et al., 2008; Jennings et al.,
1986) indicating that even single portions of
prolonged inactivity, such as days or weeks,
decrease insulin sensitivity in normal subjects.
HEALTHINF 2012 - International Conference on Health Informatics
300
On the other hand, recent studies have shown
that insulin affects the secretion of serotonin in
the brain. It also impacts memory and mood
(Reagan, 2007) and causes mood changes
(Gonder-Fredericl et al., 1989), while even
postpartum mood disorders were attributed to
decreased insulin levels (Balkau et al., 2008).
The lipid metabolism and prolonged periods of
inactivity are related according to the animal
studies (Bey and Hamilton, 2003) which
indicated the correlation even for shorter
periods of inactivity such as 4 hours or several
days. Wells et al. (Wellsl et al., 1998) claim
that the presence of lipids can induce
sleepiness and mood changes.
The effects of prolonged sitting on sympathetic
activity are demonstrated in (Jennings et al.,
1986). On the other hand, close relationship
between anxiety and sympathetic nervous
system activity is well-established (Wellsl et
al., 1998) while there were also attempts to
correlate sympathetic activity with mood states
(Yu et al., 2008) and diurnal variations
(Rechlin et al., 1995).
However, it would be inappropriate to claim
correlation between sedentary behaviour and the
mood based solely on the physiological parameters
as an underlying factor. Rather, this provided us the
motivation to examine the effects of sedentary
behaviour on psychological responses independently
from physiological parameters.
4 STUDY DESIGN
4.1 Mood Assessment
The mood may depend on a number of different
factors, such as circadian rhythms (Clark et al.,
1999), type of environment (Adan and Guardia,
1993), quality of sleep (Volkers et al.,, 1998), state
of health, private problems or other factors
incomprehensible not only to direct measurement
but also difficult for an individual to identify.
Therefore, it may be impossible to consider all the
factors that influence the mood and provide the
ultimate conclusion about the exact cause of one’s
state of mood. For this reason, our approach is to
follow relative changes of mood rather than focus on
an absolute mood state. This is because we assume
that the interval between two mood assessments of a
couple of hours (in our design) is not sufficient for a
significant change in “background” factors. These
factors, such as private problems for example, are
likely to be constantly present during relatively
longer periods of time while, the events within that
period have pre-dominant influence on relative
changes of mood. In our mood study we aim to
capture sedentary patterns during the day that
correlate with similar responses in individuals’
mood. Our method for assessing mood fluctuations
during the day is based on EMA (Ecologically
Momentary Assessment) approach in order to
compare retrospective and momentary mood data
(Smyth and Stone, 2003). The EMA approach,
which involves asking participants to report their
psychological state multiple times a day, reduces the
critical issue of retrospective recall of extended time
intervals. The retrospective recall issue is related to
cognitive and emotive limitations that bias the recall
of autobiographical memory (Smyth and Stone,
2003) influencing subject’s report by most salient
events during the recall interval. The questionnaire
we used was derived from a well-established scale
for mood study – the Profile of Mood States
(POMS) scale that consists of 65 items in its
standard version. However, long and repeated mood
questionnaires become a burden on subjects,
therefore a short version of the POMS scale was
used. We derived 8 adjectives from the POMS scale,
namely cheerful, sad, tensed, fatigued, energetic,
relaxed, annoyed and friendly that were rated on 5-
point scale (1-not at all, 2- a little, 3- moderately, 4
quite a bit, 5- extremely). The scores were summed
across the items related to positive and negative
expression in order to generate a single score for
each mood report. The difference of that score
(between two sequential questionnaires) was taken
as a measure of relative change of subject’s mood.
The questionnaires were administered three times a
day, scheduled to best fit with the work routines,
while also allowing the user to manually invoke the
questionnaire. Typically, the questionnaires were
answered in the morning, after lunch and at the end
of working day.
4.2 Sedentary Time
Accelerometers provide an important research tool
able to reliably measure and classify a number of
physical activities, including walking, jogging,
sitting, standing (Kwapisz et al., 2010), and more
complex activities such as estimation of metabolic
energy expenditure, sit-to stand transfers, and
assessment of balance and intensity of physical
activity (Kwapisz et al., 2010). In our study we
focused on distinguishing between sedentary and
A PILOT STUDY OF SHORT-TERM EFFECTS OF NON-SEDENTARY BEHAVIOUR ON MOOD CHANGES
301
non-sedentary time. Typical approach is recording
accelerometer data in 1-min epochs and a threshold
of <100 counts per minute (CPM) is chosen to
classify sedentary time (Tremblay et al., 2010; Healy
et al., 2008). Total sedentary time is calculated as a
sum of all sedentary minutes, while each minute
interval where number of accelerometer counts is
above 100 is considered a non-sedentary break
(Healy et al., 2008). The current studies
investigating sedentary behavior typically use
dedicated devices such as ActiGraph, TriTrac,
Caltrac, Actiwatch or Actical (Trost et al., 2005) that
directly provide the number of counts. On the other
hand, accelerometers are widely available in smart
phones due to their role in user interfaces (Mobile
Dev&Design, 2011). Since mobile phones are
providing the raw accelerometer signal, instead of
calculating counts we opted for a simpler approach
of analyzing standard deviation of accelerometer
signal (Kwapisz et al., 2010). However, we
compared the performance of the two approaches for
inference of sedentary time in our experimental
settings and they demonstrated comparable
performance.
In our study we used a mobile phone; not only
because it does not require an additional sensing
device to be carried but also it does not impact the
typical behavior routines of office workers. Clearly,
an issue with technological monitoring is that
subjects may not always wear the device and this is
an issue that we will address in the upcoming
studies. The typical approach to deal with this issue
of using a 20-minute inactivity criterion to identify a
non-wearing period was not applicable in our study
as office workers may sit continuously for more than
20 minutes.
5 EXPERIMENTS AND RESULTS
We recruited 14 participants (9 males, 4 females), all
of them office workers not connected with this
study, for 5 working days (characteristics of the
sample are shown in Table 1).
Table 1: Characteristics of the sample.
Age (years) 31.5±8.2
Work hours per week 38.5±2.0
Duration between two
questionnaires (min)
203.1±39.4
Non-sedentary time in
one interval (min)
39.3±19.2
The participants filled in the mood
questionnaires in the beginning, in the middle and at
the end of working day. There were no significant
differences between men and women either in the
relevant parameters (such as age, number of working
hours or type of the job regarding sedentary
routines) or in the measures (such as a number of
reported positive/negative mood changes and
average non sedentary time within one monitored
interval). None of the participants was a cigarette
smoker nor reported health problems. After
discarding intervals due to non-completed reports,
the data analyzed contained 151 monitored intervals,
74 and 77 intervals of positive and negative mood
changes respectively. Self-reported mood change,
measured as a difference in scores between two
consecutive questionnaires, was analyzed regarding
sedentary time, acquired using accelerometer
embedded in smart phones. Statistical analysis was
performed using SPSS while sedentary patterns were
inferred with R-2.13.
Figure 1, Figure 2 and Figure 3 show the
distributions of Spearman correlation between non-
sedentary time and reported change in mood. The
mean correlation between non-sedentary time and
positive/negative/overall mood changes was
0.18±0.36 (min=-0.34, max=0.86), -0.23±0.37
(min=-0.90, max=0.33) and 0.35±0.38 (min=-0.48,
max=0.86) respectively.
Figure 1: Distribution of Spearman correlation between
non-sedentary time and positive mood score change.
The non-sedentary time/positive mood changes
and non-sedentary-time/overall mood changes
distributions were significantly greater than 0
(t=1.753, t=3.451 respectively, P<0.1, P<0.005)
while non-sedentary time/negative mood changes
distribution was significantly less than 0 (t=-2.339
HEALTHINF 2012 - International Conference on Health Informatics
302
P<0.05). None of the distributions was significantly
skewed.
Figure 2: Distribution of Spearman correlation between
non-sedentary time and negative mood score change.
The results suggest that the time spent in non-
sedentary activities is positively correlated with
changes in reported positive/overall mood and
negatively correlated with changes in reported
negative mood. On the other hand, the reported
mood at the beginning of monitored intervals had
moderately low impact on the sedentary behaviour
in the following interval (r=0.24, r=-0.16 and r=0.23
respectively for initial positive/negative/overall
mood score and non-sedentary time across subjects).
This suggests that sedentary patterns were not
influenced by subjects’ mood
Figure 3: Distribution of Spearman correlation between
non-sedentary time and overall mood score change.
6 CONCLUSIONS
Our pilot study has shown correlation between
sedentary patterns at workplace and mood changes.
Considering the potential health implications, this
study provides a basis to establish new health
recommendations and create work place policies that
minimize sedentary work style, so that wellbeing of
office workers is further improved. Use of mobile
phones will allow workers to get feedback on their
mood change scores and correlation with non-
sedentary periods. This will form part of a
persuasive feedback application that will be
developed to encourage a healthy work style.
REFERENCES
A. Sanchez-Villegas, I. Ara, F. Guille´n-Grima, M. Bes-
Rastrollo, J.J.Varo-Cenarruzabeitia, M. A. Martı´nez-
Gonza´lez, “Physical activity, sedentary index, and
mental disorders in the SUN cohort study”. Med. Sci.
Sports Exerc. Vol 40(5): 827–834, 2008.
W. F. Stewart, J. A. Ricci, E. Chee, S. R. Hahn, and D.
Morganstein, “Cost of lost productive work time along
US workers with depression”. JAMA, Vol. 289,
Number 23 (2003), 3135-3144
M. S. Tremblay, R. C. Colley, T. J. Saunders, G. N. Healy,
and Neville Owen,”Psychological and health
implications of a sedentary lifestyle”. Applied
Physiology, Nutrition, and Metabolism, vol. 35(6), pp
725-740, December 2010.
Tudor-Locke CE, Myers AM. Challenges and
opportunities for measuring physical activity in
sedentary adults. Sports Med 2001;31:91–100
Hamilton, M. T., Hamilton, D. G., and Zderic, T. W. 2007.
Role of low energy expenditure and sitting in obesity,
metabolic syndrome, type 2 diabetes, and
cardiovascular disease. Diabetes, 56(11): 2655–2667.
doi:10.2337/db07-0882. PMID:17827399.
Baranowski T. Validity and reliability of self report
measures of physical activity: an information-
processing perspective. Res Q Exerc Sport 1988;
59:314–27
J. Kwapisz, G. M. Weiss, S. A. Moore, “Activity
recognition using cell phone accelerometers”, Human
Factors, pp 10-18, 2010. (Retrieved from
http://storm.cis.fordham.edu/~gweiss/papers/sensorK
DD-2010.pdf)
M. J. Mathie, A. C. F. Coster, B. H. Lovell, and B. G.
Celler, “Accelerometry: providing an integrated,
practical method for long-term, ambulatory monitoring
of human movement”, Journal of Physiological
Measurement, vol 25(2), 2004.
N. M. Hamburg, C. J. McMackin, A. L. Huang, S. M.
Shenouda. M. E. Widlansky, E. Schulz, et al.,
“Physical inactivity rapidly induces insulin resistance
A PILOT STUDY OF SHORT-TERM EFFECTS OF NON-SEDENTARY BEHAVIOUR ON MOOD CHANGES
303
and microvascular dysfunction in healthy volunteers”,
Arterioscler. Thromb. Vasc. Biol. 27(12):2650–2656,
2007.
S. R. Zwart, A. R. Hargens, S. M. Lee, B. R. Macias, D. E.
Watenpaugh, K. Tse, S. M. Smith, “Lower body
negative pressure treadmill exercise as a
countermeasure for bed rest-induced bone loss in
female identical twins”, Bone, vol 40(2), pp 529-537,
2007.
F. B. Hu, Y. T. Li, G. A. Golditz, W. C. Willett, and J. E.
Manson, “Television watching and other sedentary
behaviors in relation to risk of obesity and type 2
diabetes mellitus in women”, JAMA, vol 289 14), pp
1785–1791, 2003.
G. L. Gierach, S. C. Chang, L. A. Brinton, J. V. Lacey, A.
R. Hollenbeck, A. R., A. Schatzkin, M. F. Leitzmann,
“Physical activity, sedentary behavior, and
endometrial cancer risk in the NIH–AARP Diet and
Health Study”. Int. J. Cancer, vol 124(9), pp 2139–
2147, 2009.
C. K. Martin, T. S., Church, A. M Thompson, C. P.
Earnest, S. N. Blair, “Exercise dose and quality of life:
a randomized controlled trial”. Arch. Intern. Med. Vol
169(3), pp 269–278, 2009.
L. A. Clark, D. Watson, J. Leeka, “Diurnal variation in the
positive affect”, Motivation and Emotion, vol 13(3),
pp 205-234, 1999.
A. Adan, and J. Guardia, “Circadian variation of
selfreported activations: a multidimensional
approach”.Chronobiologia, Volume 20, Number 3-4,
233-24, 1993
A. C. Volkers, J. H. M. Tulen, and W. W. Van Den Broek,
Relationships between sleep quality and diurnal
variations in mood inhealthy subjects”.
http://www.nswo.nl/userfiles/files/publications/
jaarboek-1998/volkers.pdf, (1998), 1-5
J. M. Smyth and A. A. Stone, “Ecological momentary
assessment research in behavioral medicine”. Journal
of Happiness Studies, Vol. 4, pp 35-52, 2003.
Stand Up Australia: Sedentary behavior in workers”,
Brochure on the study undertaken by Baker IDI Heart
and Diabetes Institute and the Cancer Prevention
Research Centre at the University of Queensland in
Partnership with Meadibank Private. August 2009..
G. N. Healy, D. W. Dunstan, J. Salmon, and E. Cerin,
“Breaks in sedentary time: Beneficial associations
with metabolic risk”. Diabetes care, Vol. 31, number
4, 2008.
S. G. Trost, K. L. Mciver, and R. R. Pate, “Conducting
accelerometer-based activity assessments in field-
based research”, Medicine & Science in Sports &
Exercise, DOI: 10.1249/01.mss0000185657.86065.98,
2005.
Mobile Dev & Design, http://mobiledecdesign.com
[accessed: February 2011]
J. Sivula, W. B. Lake, R. A. Betzold, "Method and
apparatus for implementing activity sensing in a pulse
generator", US Patent number 5.052.388, 1989.
B. Balkau, L. Mhamdi, J. M. Oppert, J. Nolan, A. Golay,
F. Porcellati, et al. “Physical activity and insulin
sensitivity: the RISC study”, Diabetes, 57(10): pp.
2613–2618. 2008.
G. Jennings, L. Nelson, P. Nestel, M. Esler, P. Korner, D.
Burton, and J. Bazelmans, ”The effects of changes in
physical activity on major cardiovascular risk factors,
hemodynamics, symphatetic function, and glucose
utilization in man: a controlled study of four levels of
activity”, Circulation, Vol 73, pp 30-40, 1986.
L. Reagan, “Insulin signaling effects on memory and
mood”, Current Opinion in Parmacology, Col 7, pp
633-637, 2007.
L. A. Gonder-Fredericl, D. J. Cox, S. A. Bobbitt, and J.
Pennebaker, “Mood changes associated with blood
glucose fluctuations in insulin-dependend diabetes
mellitus”, Health Psychology, Vol. 8(1), pp 45-59,
1989.
A. Wells, N. W. Read, and A. Macdonald, “Effects of
Carbohydrate and Lipid on Resting Energy
Expenditure, Heart Rate, Sleepiness, and Mood”,
Physiological Behavior, Vol 63(4), pp 621-628, 1998.
L. Bey, and M. T. Hamilton, “Suppression of skeletal
muscle lipoprotein lipase activity during physical
inactivity: a molecular reason to maintain daily low-
intensity activity”. The Journal of Physiology, Vol.
551, pp: 673-682, Sep 1, 2003.
B. H. Yu, E. H. Kang, M. G. Ziegler, P. J. Mills, and J. E.
Dimsdale, “Mood states, sympathetic acriviy, and in
vivo b-adrenergic receptor function in a normal
population”. Depression and Anxiety, Vol. 25(7), pp:
5590564, 2008.
T. Rechlin, M. Weis, and W. P. Kaschka, “Is diurnal
variation of mood associated with parasympathetic
activity?”. Journal of Affective Disorders, Vol. 34(3),
pp: 249-255, June 8, 1995.
J. L. Andreassi, “Psycholophysiology: human behavior
and physiological response”. Routledge, 2006.
HEALTHINF 2012 - International Conference on Health Informatics
304