Functional linear modeling of activity data shows analgesic-mediated improved sleep in dogs with spontaneous osteoarthritis pain

Abstract

In humans, pain due to osteoarthritis has been demonstrated to be associated with insomnia and sleep disturbances that affect perception of pain, productivity, and quality of life. Dogs, which develop spontaneous osteoarthritis and represent an increasingly used model for human osteoarthritis, would be expected to show similar sleep disturbances. Further, these sleep disturbances should be mitigated by analgesic therapy. Previous efforts to quantify sleep in osteoarthritic dogs using accelerometry have not demonstrated a beneficial effect of analgesic therapy; this is despite owner-reported improvements in dogs’ sleep quality. However, analytic techniques for time-series accelerometry data have advanced with the development of functional linear modeling. Our aim was to apply functional linear modeling to accelerometry data from osteoarthritic dogs participating in a cross-over non-steroidal anti-inflammatory (meloxicam) drug trial. Significant differences in activity patterns were seen dogs receiving drug (meloxicam) vs. placebo, suggestive of improved nighttime resting (sleep) and increased daytime activity. These results align with owner-reported outcome assessments of sleep quality and further support dogs as an important translational model with benefits for both veterinary and human health.

Full text

1
SCIENTIFIC REPORTS | (2019) 9:14192 | https://doi.org/10.1038/s41598-019-50623-0
www.nature.com/scientificreports
Functional linear modeling of
activity data shows analgesic-
mediated improved sleep in dogs
with spontaneous osteoarthritis
pain
M. E. Gruen
1,2, D. R. Samson3 & B. D. X. Lascelles1,2,4,5,6
In humans, pain due to osteoarthritis has been demonstrated to be associated with insomnia and
sleep disturbances that aect perception of pain, productivity, and quality of life. Dogs, which develop
spontaneous osteoarthritis and represent an increasingly used model for human osteoarthritis, would
be expected to show similar sleep disturbances. Further, these sleep disturbances should be mitigated
by analgesic therapy. Previous eorts to quantify sleep in osteoarthritic dogs using accelerometry have
not demonstrated a benecial eect of analgesic therapy; this is despite owner-reported improvements
in dogs’ sleep quality. However, analytic techniques for time-series accelerometry data have advanced
with the development of functional linear modeling. Our aim was to apply functional linear modeling
to accelerometry data from osteoarthritic dogs participating in a cross-over non-steroidal anti-
inammatory (meloxicam) drug trial. Signicant dierences in activity patterns were seen dogs receiving
drug (meloxicam) vs. placebo, suggestive of improved nighttime resting (sleep) and increased daytime
activity. These results align with owner-reported outcome assessments of sleep quality and further
support dogs as an important translational model with benets for both veterinary and human health.
In humans, clear evidence exists that chronic pain interferes with sleep1. Sleep disturbances decrease quality of
life, are associated with higher anxiety and depression2, and worsen chronic pain symptoms3. A common cause
of chronic pain is osteoarthritis (OA). Several studies have reported insomnia4,5, decreased sleep quality6, and
increased self-reporting of pain7 in people with OA. Dogs also suer from OA that is pathologically and symp-
tomatically similar to humans. ese similarities have led to the dog’s emergence as a good naturally-occurring
model for understanding human arthritis pain8,9. An improved understanding of the association between OA and
sleep in dogs will enhance their use as a model for human OA.
In humans, sleep quality is oen measured objectively using actigraphy7; lower activity counts indicative
of less movement are presumed to reect higher quality sleep. Disturbances of sleep occur due to pain states,
but interestingly there are little data on the use of actigraphy to monitorsleep quality in relation to pain relief.
Analgesia-associated modication of sleep in dogs with OA has been previously evaluated by our laboratory
using accelerometry and an owner-completed sleep quality questionnaire, the Sleep and Night Time Restlessness
Evaluation (SNoRE)10. e SNoRE is a six-item instrument which asks owners to rate comfort and quality fea-
tures of their dog’s sleep. In this study, dogs with osteoarthritis wore accelerometers over a ve-week period;
they received meloxicam (Metacam®, Boehringer-Ingelheim) and placebo, each for two weeks, in a randomized
crossover design. Using the SNoRE questionnaire, the study found that dogs receiving meloxicam had improved
1Department of Clinical Sciences, North Carolina State University College of Veterinary Medicine, Raleigh, NC, USA.
2Comparative Pain Research and Education Center, North Carolina State University, Raleigh, NC, USA. 3Department
of Anthropology, University of Toronto Mississauga, Mississauga, ON, Canada. 4Translational Research in Pain
Program, North Carolina State University, College of Veterinary Medicine, Raleigh, NC, USA. 5Thurston Arthritis
Center, UNC School of Medicine, Chapel Hill, NC, USA. 6Center for Translational Pain Research, Department of
Anesthesiology, Duke University, Durham, NC, USA. Correspondence and requests for materials should be addressed
to M.E.G. (email: Margaret_gruen@ncsu.edu)
Received: 24 June 2019
Accepted: 12 September 2019
Published: xx xx xxxx
OPEN
Content courtesy of Springer Nature, terms of use apply. Rights reserved

2
SCIENTIFIC REPORTS | (2019) 9:14192 | https://doi.org/10.1038/s41598-019-50623-0
www.nature.com/scientificreports
www.nature.com/scientificreports/
sleep quality when compared to placebo; however, no dierence in mean accelerometry counts was observed.
Questionnaire items that demonstrated the highest responsiveness to meloxicam treatment regarded twitching,
dreaming, shiing position, and vocalizing. We concluded that while the SNoRE questionnaire demonstrated
responsiveness validity, criterion validity of the instrument could not be inferred due to the lack of a change in
activity as measured with accelerometry [10]. is previous study has an important limitation; however, accel-
erometry data were collected every minute and then averaged across multiple hours in the statistical analysis.
is resulted in an inability to detect short-term changes in behavior reported by owners (e.g., shiing position).
Functional data analysis, particularly functional linear modeling (FLM), has been developed specically to
evaluate actigraphy time-series data. is analytical approach has led to important discoveries about chronotype
variation and sleep-wake regulation across human groups living in natural, non-industrial environments11–14. In
animals, functional data analysis has been used to demonstrate subtle activity changes in cats with degenerative
joint disease15. is latter study in cats highlights the potential applications for functional data analysis in studies
of pain and activity, but to date, no studies in dogs or cats have used FLM to assess the impact of pain relief on
activity or sleep in a chronic pain state.
Pain-related sleep disturbance in humans is considered to be due to spontaneous pain as opposed to move-
ment or activity-related pain; self-reported spontaneous pain in humans has been very dicult to measure in ani-
mal models. Assessment of sleep quality in the canine osteoarthritis model may provide a highly relevant method
to assess the eectiveness of analgesics on spontaneous pain. In the present study, we use FLM to reanalyze our
data from the SNoRE study to evaluate the utility of using FLM to measure improved sleep quality due to the
alleviation of spontaneous pain, and to assess the validity of our questionnaire.
Results
Subjects. Fieen dogs (10 spayed females and 5 neutered males) had complete activity data sets and were
included in the analysis. Dogs had a mean (±SD) age of 10.29 ± 2.48 years, weight of 31.53 ± 5.39 kgs, Canine
Brief Pain Inventory (CBPI) pain score of 3.85 ± 1.5.
SNoRE questionnaire data. As previously reported10, the overall score on the SNoRE instrument detected
a positive improvement due to the NSAID (p = 0.001) and detected a dierence between the NSAID and placebo
(p = 0.041). Table1 details the change from baseline score for each questionnaire item and for the overall score.
Linear mixed-eects model. Our results replicated the previously reported nding: using this method we
found no eect of treatment on nighttime activity. Weekend and weekday were included as covariates as previous
work has shown a dierence in activity on weekdays versus weekends in dogs29. is model found that only week-
end versus weekday was signicantly associated with nighttime activity; regardless of treatment, dogs were more
active on weekend nights than weekdays. Table2 shows the results of the model for nighttime activity.
Results for the linear mixed-eects model for daytime activity are shown in Table3. As with the nighttime
activity, weekend vs. weekday was signicantly associated with daytime activity; dogs were signicantly more
active on weekend days than weekdays, regardless of treatment. In addition, dogs with higher baseline CBPI
scores were more active during the day (p = 0.041) and male dogs were more active than female dogs (p = 0.032).
e correlations of daytime activity with these two variables were modest (CBPI r = 0.21; Sex r = 0.35).
Functional linear modeling. In contrast to the linear mixed-eects model, the functional linear model
found signicant dierences in activity when the dogs were receiving meloxicam versus placebo. Figure1 shows
the group level circadian activity pattern, with red representing the drug group and black representing the pla-
cebo group. is plot shows a clear separation of the mean circadian activity and identies the time periods when
the curves dier between groups: (23:00–06:00, 07:00–12:00, 21:00–22:00). In general, dierences in circadian
activity were apparent, with the drug group characterized by decreased nighttime activity between 23:00–06:00,
and increased daytime activity between the periods of both 07:00–12:00 and 21:00–22:00. In other words, we
found that when receiving meloxicam, relative to placebo, the dogs had greater nighttime resting and more pro-
nounced activity during the day.
NSAID – Baseline Placebo – Baseline NSAID - Placebo
Mean dierence p-value Mean dierence p-value Mean dierence p-value
Q1:Movement −1.00 0.038 −0.33 0.554 −0.67 0.313
Q2:Twitching −1.20 0.028 −0.20 0.638 −1.40 0.012
Q3: Dreaming −1.80 <0.001 −0.40 0.32 −1.40 <0.001
Q4:Shiing position −1.13 0.059 −0.13 0.860 −1.00 0.165
Q5: Vocalizing −1.07 0.010 −0.60 0.082 −0.47 0.131
Q6: Pacing 0.27 0.452 0.33 0.371 −0.07 0.915
Total (Q 1–5) −6.20 0.001 −1.27 0.456 −4.93 0.029
Total (6Q) −6.47 0.001 −0.93 0.616 −5.53 0.041
Table 1. Results of the SNoRE questionnaire for each individual question as well as total score for all six
questions (6Q) and for a modication using only ve questions (5Q) with the removal of Question 6. Results are
shown for each treatment period compared to baseline, and comparing treatment with an NSAID (meloxicam)
to placebo.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

3
SCIENTIFIC REPORTS | (2019) 9:14192 | https://doi.org/10.1038/s41598-019-50623-0
www.nature.com/scientificreports
www.nature.com/scientificreports/
Discussion
In this study, when we used functional linear modeling, we found a robust and signicant dierence (previously
masked when applying a more traditional technique) in the pattern of nighttime and daytime activity in dogs
receiving meloxicam, compared to placebo. e traditional technique of analyzing high-frequency longitudinal
activity data uses summary values, averaging activity counts over large periods of time; this sacrices the detail
available in the data. Importantly, summary statistical approaches do not allow for understanding patterns of
activity. Functional linear modeling (FLM) is designed specically to address these limitations; the granularity
of the data is harnessed to allow analysis of the pattern of activity across the day16. Using this approach, we found
improved sleep, as dened by decreased nighttime activity, and this was supported by the owners’ assessment
Estimate Standard Error z-value P(>|z|)
Age 0.228 0.154 1.480 0.139
Weekend/Weekday 0.108 0.039 2.732 0.006
CBPI Score −0.196 0.160 1.220 0.223
Sex −0.169 0.165 1.021 0.307
Weight −0.155 0.180 0.861 0.389
Treatment −0.022 0.040 0.554 0.580
Table 2. Model-averaged coecients from linear mixed-eects model evaluating covariate eects on canine
nighttime activity. Signicant eects were found only for day of the week (weekend activity was higher than
weekday). CBPI = Canine Brief Pain Inventory.
Estimate Standard Error z-value P(>|z|)
Age −0.082 0.194 0.420 0.675
Weekend/Weekday 0.101 0.028 3.521 <0.001
CBPI Score 0.388 0.189 2.040 0.041
Sex 0.414 0.192 2.148 0.032
Weight 0.218 0.212 1.026 0.305
Treatment 0.016 0.028 0.550 0.582
Table 3. Model-averaged coecients from linear eects model evaluating covariate eects on canine daytime
activity. Signicant eects were found for day of the week (weekend activity was higher than weekday), CBPI
score (dogs with higher baseline CBPI scores were more active), and sex (males were more active than females).
Figure 1. A functional linear modeling comparison between the 24-hour sleep-wake pattern of placebo and
drug trials. e bottom panel illustrates the point-wise critical value (dotted line). is is the proportion of all
permutation F values at each time point at the signicance level of 0.05. When the observed F-statistic (solid
line) is above the dotted line, it is concluded the two groups have signicantly dierent mean circadian activity
patterns at those time points.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

4
SCIENTIFIC REPORTS | (2019) 9:14192 | https://doi.org/10.1038/s41598-019-50623-0
www.nature.com/scientificreports
www.nature.com/scientificreports/
of quality of sleep of their pet. To our knowledge, FLM has not been applied to activity data from any species in
relation to pain and pain relief previously.
We first re-evaluated the traditional technique by replicating the previous study’s findings using linear
mixed-eects modeling10. Of the tested covariates, only activity on the weekend, relative to weekday, was signi-
cantly associated with nighttime activity: dogs were more active on weekends regardless of treatment. While these
null results with regard to treatment are consistent with the previous study, they are in contrast to the reports of
owners; using the SNoRE questionnaire, owners were able to detect a dierence in sleep quality when their dogs
were on meloxicam. In the previous study, average activity per hour over the nighttime period showed a small, but
non-signicant, decrease in activity with meloxicam treatment10. By applying an FLM approach, we were able to
detect this dierence in dogs’ overnight activity and found activity data matched owners’ assessments. e most
signicant dierences were found during the overnight period from approximately 11 pm-6 am; dogs were less
active (and possibly sleeping) when receiving meloxicam rather than placebo. e 11 pm–6 am time interval cor-
responds to the time that the majority of owners designated as “night-time;” it also fully encompasses the period
from 12 am-5 am that was designated as “night-time” for all owners in the previous study10. A smaller, though sig-
nicant, dierence was seen between approximately 7 pm and 8 pm, where dogs receiving meloxicam were more
active than dogs receiving placebo; for the majority of the evening, dogs were not signicantly dierent. e lack
of dierence in the evening may be due to the mediating eect of owner interaction during the evening hours.
Another possibility is that the owners were instructed to give the medication in the evening, but the exact timing
was not recorded. It is possible that the lack of dierence between placebo and NSAID in the evening was due to
decreased eectiveness towards the end of the 24-hour dosing interval. In contrast, dogs receiving placebo were
signicantly less active during the morning hours between 7 and 11 am. is nding ts with work from other
species, including humans, showing increased daytime sleepiness following poor quality sleep17, and an increase
in the pain-alleviating eect of sleep during the rst half of the day7.
e ndings from this study support the use of the SNoRE questionnaire as an owner-completed outcome for
sleep quality in dogs with chronic pain. Our results support criterion validity of the SNoRE. e items from the
SNoRE questionnaire that were most able to distinguish a positive eect of meloxicam on sleep were specically
regarding low-level movements such as twitching and dreaming. Indeed, dreaming was rated as signicantly
dierent from baseline while taking meloxicam, and change on this measure was signicantly dierent between
meloxicam and placebo. Unfortunately, we do not currently know what algorithms to apply to our activity data to
be able to distinguish the types of movements being captured; future work is necessary to distinguish movements
associated with dreaming or twitching from other types of movement. Non-steroidal anti-inammatory medi-
cations, like meloxicam, are considered generally sleep-neutral in humans, with no eect on REM sleep18. Future
work is needed to understand what owners are classifying as “dreaming” and how this is associated with pain.
This study represents the first application of FLM to sleep in dogs, and the first application of FLM to
activity data used to measure the eects of analgesics in chronic pain. Recognition of the impact of dogs as
naturally-occurring models of disease is increasing8,19,20; as such, interest in dog sleep is growing. An area of grow-
ing research is the connection between chronic pain, sleep disturbances, and cognitive impairment in people21,22.
As with people, sleep has demonstrated importance in learning and memory consolidation in dogs23. It is likely
that sleep impairment due to chronic pain would have similarly disruptive eects on canine cognition; this is an
area of future research. Other chronic diseases in dogs, such as cognitive dysfunction, are associated with changes
in sleep-wake cycle24; characterization of the normal sleep-wake cycle in pet dogs, and further, the relationship
between pet and caregiver sleep patterns, would be a valuable contribution to the eld of veterinary and compar-
ative medicine. A comparison of overnight activity between dogs with osteoarthritis and age matched controls
remains a gap in the eld; such a comparison would allow us to determine the extent of baseline sleep distur-
bances in dogs with chronic pain compared to those without. However, the current study supports the hypothesis
that dogs with chronic pain have disturbances in sleep that are relieved when they receive adequate pain man-
agement. Each dog served as their own control, increasing our condence that, rather than age or group-level
dierences between dogs, these dierences were due to the treatment. is study also provides additional support
for the applicability of dogs as a model of OA-associated induced and spontaneous pain in humans25, especially
with an objective measure used commonly in studying human sleep26,27 and aprotocol for analysis of the data.
In summary, this study has demonstrated improved sleep quality due to pain relief in a spontaneous canine
model of osteoarthritis pain. It has also supported two important ndings: rst, the SNoRE questionnaire has cri-
terion validity and appears to be useful in detecting the alleviation of sleep disturbances associated with chronic
pain in dogs; second, FLM is a more sensitive analytic technique for activity data in dogs—this technique was able
to detect a dierence in activity patterns when traditional summary techniques were not. ese results further
expand the translational potential of client-owned dogs with osteoarthritis pain; they are a model of spontaneous
pain-associated sleep disturbance and could be involved in the evaluation of putative analgesics.
Methods
SNoRE study. Experimental details for the SNoRE study have been previously described10. is was a dou-
ble-masked, placebo-controlled crossover study design. Client-owned dogs (n = 15) that met rigorous inclusion
and exclusion criteria received meloxicam and placebo in random order: meloxicam or placebo for two weeks,
followed by a one-week washout, followed by two more weeks of meloxicam or placebo. Meloxicam was dosed at
0.2 mg/kg by mouth on the rst day, followed by 0.1 mg/kg once daily; placebo was volume matched and visually
identical. Owners were instructed to administer the meloxicam/placebo treatments in the evening (at approxi-
mately 6 pm) each day of the study. Dogs wore accelerometers (Actical, Philips Respironics) to collect data each
minute on spontaneous activity. Owners kept a sleep-diary and noted the times when they went to bed in the
evening and rose in the morning. is allowed activity data to be classied as night-time activity (NTA) or day-
time activity (DTA) for use in the model. Owners completed a series of clinical metrology instruments at various
Content courtesy of Springer Nature, terms of use apply. Rights reserved

5
SCIENTIFIC REPORTS | (2019) 9:14192 | https://doi.org/10.1038/s41598-019-50623-0
www.nature.com/scientificreports
www.nature.com/scientificreports/
timepoints in the study. For the purposes of the current study, data are included from the Canine Brief Pain
Inventory (CBPI28) completed on Day 0, and the Sleep and Night Time Restlessness Evaluation (SNoRE10) com-
pleted on Days 0, 14, 21, 28, and 35. is 6-item questionnaire evaluates sleep quality in dogs over a 7-day period.
is study was approved by the North Carolina State University Institutional Animal Care and Use Committee
(approval #07-188-O); all experiments were performed in accordance with relevant guidelines and regulations.
All dog owners were over 18 years old and provided written informed consent.
Data analysis. To replicate previous ndings, we rst analyzed activity data using a linear mixed-eects
model, with the covariates of age, weight, sex, CBPI pain sub-score, and treatment (meloxicam or placebo). Based
on previous work showing a dierence in weekend and weekday activity patterns in dogs29, we added this as a
covariate for our linear model. As functional data analysis is performed across a 24-hour period, we repeated our
linear model for daytime activity (not previously reported).
Functional linear modeling (FLM) is specically designed for actigraphy time-series data analysis. We used
it to statistically characterize and illustrate 24-hour sleep-wake patterns of the same dogs that were either given a
placebo or drug. e strength of this approach is that it measures raw activity counts within and between samples
and avoids summary statistics that can mask dierences across groups16. erefore, we applied a non-parametric
permutation test method in the R package “actigraphy”30, as it does not rely on distributional assumptions.
Signicance was calculated by counting the proportion of permutation F values that are larger than the F statis-
tics for the observed pairing. Here, we used the point-wise test (with 500 permutations; bspline method) which
provides a curve that represents the proportion of permutation F values that are larger than the F statistic for the
observed pairing at each point in the time series16.
Individual items and total scores on the SNoRE questionnaire were analyzed using matched pairs t-tests to
compare the meloxicam and placebo treatment (Days 14 and 35) to baseline (Day 0), and to each other. To adjust
for these multiple comparisons, a critical p-value of 0.016 was considered signicant.
References
1. Mathias, J. L., Cant, M. L. & Bure, A. L. J. Sleep disturbances and sleep disorders in adults living with chronic pain: a meta-analysis.
Sleep Med. 52, 198–210, https://doi.org/10.1016/j.sleep.2018.05.023 (2018).
2. Alvaro, P. ., oberts, . M. & Harris, J. . A. Systematic eview Assessing Bidirectionality between Sleep Disturbances, Anxiety,
and Depression. Sleep 36, 1059–1068, https://doi.org/10.5665/sleep.2810 (2013).
3. Uchmanowicz, I., oltuniu, A., Stepien, A., Uchmanowicz, B. & osinczu, J. e inuence of sleep disorders on the quality of life
in patients with chronic low bac pain. Scand. J. Caring Sci. 33, 119–127, https://doi.org/10.1111/scs.12610 (2019).
4. Campbell, C. M. et al. Sleep, Pain Catastrophizing, and Central Sensitization in nee Osteoarthritis Patients With and Without
Insomnia. Arthritis Care Res. 67, 1387–1396, https://doi.org/10.1002/acr.22609 (2015).
5. wiatowsa, B., la, A., aciborsi, F. & Maslinsa, M. e prevalence of depression and insomnia symptoms among patients
with rheumatoid arthritis and osteoarthritis in Poland: a case control study. Psychol. Health Med. 24, 333–343, https://doi.org/10.10
80/13548506.2018.1529325 (2019).
6. Wilcox, S. et al. Factors related to sleep disturbance in older adults experiencing nee pain or nee pain with radiographic evidence
of nee osteoarthritis. J. Am. Geriatr. Soc. 48, 1241–1251, https://doi.org/10.1111/j.1532-5415.2000.tb02597.x (2000).
7. Tang, N. ., Goodchild, C. E., Sanborn, A. N., Howard, J. & Salovsis, P. M. Deciphering the temporal lin between pain and sleep
in a heterogeneous chronic pain patient sample: a multilevel daily process study. Sleep 35, 675–687A, https://doi.org/10.5665/
sleep.1830 (2012).
8. linc, M. P. et al. Translational pain assessment: could natural animal models be the missing lin? Pain 158, 1633–1646, https://doi.
org/10.1097/j.pain.0000000000000978 (2017).
9. Lascelles, B. D. X., Brown, D. C., Maixner, W. & Mogil, J. S. Spontaneous painful disease in companion animals can facilitate the
development of chronic pain therapies for humans. Osteoarthritis C artilage 26, 175–183, https://doi.org/10.1016/j.joca.2017.11.011
(2018).
10. nazovicy, D., Tomas, A., Motsinger-eif, A. & Lascelles, B. D. Initial evaluation of nighttime restlessness in a naturally occurring
canine model of osteoarthritis pain. PeerJ 3, e772, https://doi.org/10.7717/peerj.772 (2015).
11. Samson, D. ., Crittenden, A. N., Mabulla, I. A., Mabulla, A. Z. P. & Nunn, C. L. Chronotype variation drives night-time sentinel-
lie behaviour in hunter-gatherers. Proc. Biol. Sci. 284, https://doi.org/10.1098/rspb.2017.0967 (2017).
12. Samson, D. . et al. What is segmented sleep? Actigraphy eld validation for daytime sleep and nighttime wae. J. Sleep Res. https://
doi.org/10.1016/j.sleh.2016.09.006 (2016).
13. Samson, D. ., Crittenden, A. N., Mabulla, A. I., Mabulla, A. Z. P. & Nunn, C. L. Hadza sleep biology: evidence for exible sleep-
wae patterns in hunter-gatherers. Am. J. Phys. Anthropol. 162, 573–582, https://doi.org/10.1002/ajpa.23160 (2017).
14. Samson, D. . et al. Segmented sleep in a nonelectric, smallscale agricultural society in Madagascar. Am. J. Hum. Biol. 29, 1–13,
https://doi.org/10.1002/ajhb.22979 (2017).
15. Gruen, M. E. et al. e Use of Functional Data Analysis to Evaluate Activity in a Spontaneous Model of Degenerative Joint Disease
Associated Pain in Cats. PLoS One 12, e0169576, https://doi.org/10.1371/journal.pone.0169576 (2017).
16. Wang, J. et al. Measuring the impact of apnea and obesity on circadian activity patterns using functional linear modeling of
actigraphy data. J. Circadian Rhythms 9, 11, https://doi.org/10.1186/1740-3391-9-11 (2011).
17. Alexandre, C. et al. Decreased alertness due to sleep loss increases pain sensitivity in mice. Nat. Med. 23, 768–774, https://doi.
org/10.1038/nm.4329 (2017).
18. Bohra, M. H., aushi, C., Temple, D., Chung, S. A. & Shapiro, C. M. Weighing the balance: how analgesics used in chronic pain
inuence sleep? Brit. J. Pain 8, 107–118, https://doi.org/10.1177/2049463714525355 (2014).
19. Schutt, T. et al. Dogs with Cognitive Dysfunction as a Spontaneous Model for Early Alzheimer’s Disease: A Translational Study of
Neuropathological and Inammatory Marers. J. Alzheimers Dis. 52, 433–449, https://doi.org/10.3233/Jad-151085 (2016).
20. Head, E. A canine model of human aging and Alzheimer’s disease. Biochim. Biophys. Acta. 1832, 1384–1389, https://doi.
org/10.1016/j.bbadis.2013.03.016 (2013).
21. Innes, . E. & Sambamoorthi, U. e Association of Perceived Memory Loss with Osteoarthritis and elated Joint Pain in a Large
Appalachian Population. Pain Med. 19, 1340–1356, https://doi.org/10.1093/pm/pnx107 (2018).
22. Waler, M. P. & Sticgold, . Sleep-dependent learning and memory consolidation. Neuron 44, 121–133, https://doi.org/10.1016/j.
neuron.2004.08.031 (2004).
23. is, A. et al. e interrelated eect of sleep and learning in dogs (Canis familiaris); an EEG and behavioural study. Sci. Rep. 7, 41873,
https://doi.org/10.1038/srep41873 (2017).
Content courtesy of Springer Nature, terms of use apply. Rights reserved

6
SCIENTIFIC REPORTS | (2019) 9:14192 | https://doi.org/10.1038/s41598-019-50623-0
www.nature.com/scientificreports
www.nature.com/scientificreports/
24. Landsberg, G. M., DePorter, T. & Araujo, J. A. Clinical Signs and Management of Anxiety, Sleeplessness, and Cognitive Dysfunction
in the Senior Pet. Vet. Clin. N. Am. 41, 565–590, https://doi.org/10.1016/j.cvsm.2011.03.017 (2011).
25. Meeson, . L., Todhunter, . J., Blunn, G., Nui, G. & Pitsillides, A. A. Spontaneous dog osteoarthritis - a One Medicine vision. Nat.
Rev. Rheumatol. 15, 273–287, https://doi.org/10.1038/s41584-019-0202-1 (2019).
26. Sadeh, A. e role and validity of actigraphy in sleep medicine: an update. Sleep Med. Rev. 15, 259–267, https://doi.org/10.1016/j.
smrv.2010.10.001 (2011).
27. Martoni, M., Bayon, V., Elbaz, M. & Leger, D. Using actigraphy versus polysomnography in the clinical assessment of chronic
insomnia (retrospective analysis of 27 patients). Presse Med. 41, e95–e100, https://doi.org/10.1016/j.lpm.2011.07.019 (2012).
28. Brown, D. C., Boston, . C., Coyne, J. C. & Farrar, J. T. Ability of the canine brief pain inventory to detect response to treatment in
dogs with osteoarthritis. J. Am. Vet. Med. Assoc. 233, 1278–1283 (2008).
29. Lascel les, B. D. et al. Evaluation of a digitally integrated accelerometer-based activity monitor for the measurement of activity in cats.
Vet. Anaesth. Analg. 35, 173–183, https://doi.org/10.1111/j.1467-2995.2007.00367.x (2008).
30. Shannon, B. et al. How apnea and obesity eect circadian activity patterns using functional linear modeling of actigraphy data. Sleep
35, A437-A437, Proceedings of the 26th Annual Meeting of the Association of Professional Sleep Societies (2012).
Author Contributions
B.D.X.L. designed the original study; D.R.S., M.E.G. and B.D.X.L. designed the current study. D.R.S. performed
the analyses and prepared the gure; M.E.G. draed the manuscript; all authors reviewed the manuscript.
Additional Information
Competing Interests: D.R.S. declares no nancial or non-nancial competing interests. M.E.G. has performed
consulting work and B.D.X.L. has participated in sponsored CE for Boehringer-Ingelheim (manufacturers of
Metacam® used in the original study).
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and
institutional aliations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International
License, which permits use, sharing, adaptation, distribution and reproduction in any medium or
format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Cre-
ative Commons license, and indicate if changes were made. e images or other third party material in this
article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the
material. If material is not included in the article’s Creative Commons license and your intended use is not per-
mitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the
copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
© e Author(s) 2019
Content courtesy of Springer Nature, terms of use apply. Rights reserved

1.
2.
3.
4.
5.
6.
Terms and Conditions
Springer Nature journal content, brought to you courtesy of Springer Nature Customer Service Center GmbH (“Springer Nature”).
Springer Nature supports a reasonable amount of sharing of research papers by authors, subscribers and authorised users (“Users”), for small-
scale personal, non-commercial use provided that all copyright, trade and service marks and other proprietary notices are maintained. By
accessing, sharing, receiving or otherwise using the Springer Nature journal content you agree to these terms of use (“Terms”). For these
purposes, Springer Nature considers academic use (by researchers and students) to be non-commercial.
These Terms are supplementary and will apply in addition to any applicable website terms and conditions, a relevant site licence or a personal
subscription. These Terms will prevail over any conflict or ambiguity with regards to the relevant terms, a site licence or a personal subscription
(to the extent of the conflict or ambiguity only). For Creative Commons-licensed articles, the terms of the Creative Commons license used will
apply.
We collect and use personal data to provide access to the Springer Nature journal content. We may also use these personal data internally within
ResearchGate and Springer Nature and as agreed share it, in an anonymised way, for purposes of tracking, analysis and reporting. We will not
otherwise disclose your personal data outside the ResearchGate or the Springer Nature group of companies unless we have your permission as
detailed in the Privacy Policy.
While Users may use the Springer Nature journal content for small scale, personal non-commercial use, it is important to note that Users may
not:
use such content for the purpose of providing other users with access on a regular or large scale basis or as a means to circumvent access
control;
use such content where to do so would be considered a criminal or statutory offence in any jurisdiction, or gives rise to civil liability, or is
otherwise unlawful;
falsely or misleadingly imply or suggest endorsement, approval , sponsorship, or association unless explicitly agreed to by Springer Nature in
writing;
use bots or other automated methods to access the content or redirect messages
override any security feature or exclusionary protocol; or
share the content in order to create substitute for Springer Nature products or services or a systematic database of Springer Nature journal
content.
In line with the restriction against commercial use, Springer Nature does not permit the creation of a product or service that creates revenue,
royalties, rent or income from our content or its inclusion as part of a paid for service or for other commercial gain. Springer Nature journal
content cannot be used for inter-library loans and librarians may not upload Springer Nature journal content on a large scale into their, or any
other, institutional repository.
These terms of use are reviewed regularly and may be amended at any time. Springer Nature is not obligated to publish any information or
content on this website and may remove it or features or functionality at our sole discretion, at any time with or without notice. Springer Nature
may revoke this licence to you at any time and remove access to any copies of the Springer Nature journal content which have been saved.
To the fullest extent permitted by law, Springer Nature makes no warranties, representations or guarantees to Users, either express or implied
with respect to the Springer nature journal content and all parties disclaim and waive any implied warranties or warranties imposed by law,
including merchantability or fitness for any particular purpose.
Please note that these rights do not automatically extend to content, data or other material published by Springer Nature that may be licensed
from third parties.
If you would like to use or distribute our Springer Nature journal content to a wider audience or on a regular basis or in any other manner not
expressly permitted by these Terms, please contact Springer Nature at
onlineservice@springernature.com