Scott et al. Int J Bipolar Disord (2020) 8:37 https://doi.org/10.1186/s40345-020-00200-6 REVIEW Open Access An evidence map of actigraphy studies exploring longitudinal associations between rest‑activity rhythms and course and outcome of bipolar disorders Jan Scott1,2,3, Francesc Colom4,5, Allan Young2,8, Frank Bellivier3,6,7 and Bruno Etain3,6,7* Abstract Background: Evidence mapping is a structured approach used to synthesize the state-of-the-art in an emerging field of research when systematic reviews or meta-analyses are deemed inappropriate. We employed this strategy to summarise knowledge regarding longitudinal ecological monitoring of rest-activity rhythms (RAR) and disease modi- fiers, course of illness, treatment response or outcome in bipolar disorders (BD). Structure: We had two key aims: (1) to determine the number and type of actigraphy studies of in BD that explored data regarding: outcome over time (e.g. relapse/recurrence according to polarity, or recovery/remission), treatment response or illness trajectories and (2) to examine the range of actigraphy metrics that can be used to estimate dis- ruptions of RAR and describe which individual circadian rhythm or sleep–wake cycle parameters are most consistently associated with outcome over time in BD. The mapping process incorporated four steps: clarifying the project focus, describing boundaries and ‘coordinates’ for mapping, searching the literature and producing a brief synopsis with summary charts of the key outputs. Twenty-seven independent studies (reported in 29 publications) were eligible for inclusion in the map. Most were small-scale, with the median sample size being 15 per study and median duration of actigraphy being about 7 days (range 1–210). Interestingly, 17 studies comprised wholly or partly of inpatients (63%). The available evidence indicated that a discrete number of RAR metrics are more consistently associated with transi- tion between different phases of BD and/or may be predictive of longitudinal course of illness or treatment response. The metrics that show the most frequent associations represent markers of the amount, timing, or variability of RAR rather than the sleep quality metrics that are frequently targeted in contemporary studies of BD. Conclusions: Despite 50 years of research, use of actigraphy to assess RAR in longitudinal studies and examination of these metrics and treatment response, course and outcome of BD is under-investigated. This is in marked contrast to the extensive literature on case–control or cross-sectional studies of actigraphy, especially typical sleep analysis metrics in BD. However, given the encouraging findings on putative RAR markers, we recommend increased study of putative circadian phenotypes of BD. Keywords: Evidence map, Actigraphy, Longitudinal, Bipolar, Response, Course, Modifiers, Lithium, Phenotype, Domains Introduction *Correspondence: bruno.etain@inserm.fr In the 2017 update on the global burden of disease dem- 3 Université Paris Diderot, Sorbonne Paris Cité, UMR-S 1144, 75013 Paris, onstrates that, the four most burdensome non-communi- France cable conditions (in terms of DALYs: disability adjusted Full list of author information is available at the end of the article © The Author(s) 2020. 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 Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted 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 licence, visit http://creat​iveco​ mmons​.org/licen​ses/by/4.0/. Scott et al. Int J Bipolar Disord (2020) 8:37 Page 2 of 22 life years) are: cardiovascular disease (CVD), cancers, few can accurately estimate key markers of circadian tim- musculoskeletal conditions, and mental disorders (GBD ing (Zee et al. 2014; FDA 2019; Depner et al. 2020). The 2017; Whiteford et al. 2013). In individuals aged 15–49, lack of regulatory approval and absence of basic research mental disorders in general, and mood disorders specifi- on or rigour of testing their performance quality means cally, are regarded as the most burdensome conditions consumer devices are not currently recommended for worldwide. In the last 3 to 4 decades, considerable pro- use in this research field (Baron 2018; FDA 2019; Moar gress has been made in the management of CVD and 2018). Currently, RDOC indicates that actigraphy could cancer through precision medicine. The application of be applied as an ‘analytic unit’ for longitudinal studies precision medicine is in its infancy in psychiatry, but of circadian rhythms and sleep–wake cycles (Cuthbert funding initiatives (such as the European H2020 plat- 2014). Such studies are now underway, but there is no form) are supporting studies of personalized diagnostics consensus regarding the most useful actigraphy-derived and therapeutics across a range of mental conditions RAR parameters to estimate and/or report (Scott et al. including bipolar disorders (BD) (Schumann et al. 2014). 2019; Smagula 2016. Also, a preliminary scoping exercise Like precision approaches to other chronic disorders, of this topic suggested that relatively few actigraphy stud- these studies have largely focused on more quantifi- ies have employed a longitudinal or prospective design in able metrics (such as ‘omics’ and brain scanning) as an BD. Therefore, we decided to collate any existing publica- attempt to increase the biological validity of observed tions on these issues and synthesise the current state of symptoms (Frey et al. 2013; Hellwig and Domschke knowledge. 2019; MacDonald et al. 2019). This strategy also reflects This paper has three aims: the proposals described in the RDOC (research domain criteria) paradigm that identifies key domains and ana- (a) To determine the number and type of actigraphy lytic units for translational research (Cuthbert and Insel studies of samples comprised wholly or partly of 2013; Insel 2014). Although personalized diagnostics for individuals with BD that explore data regarding: BD are likely to remain a long-term aspiration, there is outcome over time (e.g. relapse/recurrence accord- emerging evidence that multi-platform, integrated sci- ing to polarity, or recovery/remission), treatment ence approaches could offer a pathway for stratifying response or illness trajectories. individuals with BD to improve outcome prediction, (b) To examine the range of actigraphy metrics that namely treatment response, relapse and recurrence, and can be used to estimate disruptions of RAR and individual illness trajectories (Scott et al. 2018). How- describe which individual circadian rhythm or ever, as these efforts progress, there is a need to extent sleep–wake cycle parameters are most consistently the search for phenotypes beyond laboratory settings and associated with outcome over time in BD. to test potential biomarkers in real-world settings, such (c) To demonstrate the use of evidence mapping as a as field and cohort studies, comparative effectiveness process for synthesising the state-of-the-art in the research, and large-scale pragmatic trials (Brietzke et al. field of ecological monitoring of RAR and disease 2019; Scott et al. 2019). modifiers, course of illness, treatment response or One of the domains identified by RDOC is the arousal outcome in BD. and regulatory system, and RDOC suggests that sleep– wake cycle and circadian rhythms (which we will refer to as rest-activity rhythms or RAR) represent a core con- Methods struct for investigation (Insel 2014; Smagula 2016). This As the use of evidence-mapping dictates the structure of is of relevance to BD, as converging evidence indicates the paper, we provide a brief overview of this approach in that RAR disruptions can differentiate BD cases from Table 1. Below, we outline the application to the project healthy controls (HC) and/or other comparator groups described in this article. (Ritter et al. 2011; Geoffroy et al. 2015; Ng et al. 2015; De Crescenzo et al. 2017; Scott et al. 2017). Further, real-time Rationale for use of mapping monitoring of most of these phenomena may be possible As noted in the introduction, the goal of this evidence with wrist-worn devices. Of course, there has been an map is to present an overview of the extent, range, nature, exponential increase in the development and use of con- and findings of clinical research on RAR metrics meas- sumer grade wearables (and smart phone apps) in clinical ured by actigraphy and a range of potential outcomes of populations (Faurolt-Jepson et al. 2012; Krane-Gartiser BD. Like many evidence maps, the rationale for using this et al. 2019). However, many demonstrate low inter- process was determined by prior knowledge of the field. device reliability, none of the apps or programmes are Many research groups, including our own, have synthe- currently validated against ‘gold standard’ measures and sized research findings or undertaken meta-analyses of Scott et al. Int J Bipolar Disord (2020) 8:37 Page 3 of 22 Table 1 Brief overview of evidence mapping Evidence Mapping Evidence mapping is defined as- “a form of knowledge synthesis that addresses an exploratory research question aimed at mapping key concepts, types of evidence, and gaps in research related to a defined area or field by systematically searching, selecting, and synthesizing existing understanding” The gold standard methods used in evidence-based research are systematic review and meta-analysis. These classic approaches are both rigor- ous and provide readers with detailed information about narrow questions (e.g. the efficacy of drug X for disorder Y). However, developments from this archetypal model are now being developed to meet more diverse needs regarding evidence synthesis (Arksey 2003; Miake-Lye et al. 2016). The obvious example is the use of rapid reviews (which address urgent topics and do not always fully adhere to systematic methodolo- gies) or scoping reviews (that identify extensive bodies of literature but do not usually provide detailed synthesis) (Colquhoun et al. 2014). In this context, evidence mapping has emerged as another process that allows researchers and their audience to develop an understanding of the extent and distribution of evidence in on a broad topic, highlighting what is known and also where gaps exist (Katz et al. 2003; Hetrick et al. 2010). Although it applies systematic and replicable methodology, the process is iterative, as expert consensus and preliminary literature searches inform the next steps in the review, so the final product is broader and less detailed than the traditional approaches (Vallarino et al. 2015). Furthermore, the goal is not to provide detailed statistical comparisons, but rather to offer an essential snapshot of what is or is not known at this moment in time. It is presumed that an evidence map for any given field will be followed up in the future when research devel- ops to a point where it becomes justifiable to apply more formal reviews and pooled analyses The depth of evidence synthesized will differ depending on its purpose of an evidence map e.g. an overview of the extent, range and nature of research activity might entirely exclude any details of study findings (Miake-Lye et al. 2016). However, if the latter are considered, they can be represented by graphs and figures, instead of or as well as a detailed table such as the summary tables provided in systematic reviews (Hetrick et al. 2010) Given the proposed scope of this map, we determined that a simple written summary with basic tables and figures would be offered in the results section of the paper. A detailed Table summarizing sample characteristics and research findings (in an Appendix) and citations in the reference list only would ensure readers focused on key findings we summarize. Individual readers could then choose to examine supporting evidence and relevant references for themes that they wished to investigate further pooled data regarding BD cases compared with HC or consultation among the co-authors, the questions and other (comparator) groups (Ritter et al. 2011; Geoffroy scope of the mapping project were defined as follows- et al. 2015). Some of those projects identified small-scale studies that used real-time monitoring of BD cases and/ Step 1: Clarification of the focus of the map. or that examined associations between baseline actigra- It was agreed that the focus would be: phy recordings and longitudinal course and outcome of BD (Scott et al. 2017). Nearly all of those projects were (a) What evidence exists on this topic and based on a undertaken several decades ago and differences in sam- review of study characteristics and methodologies: pling, diagnosis or research methods meant the publi- what conclusions can be drawn about the quality of cations failed to meet eligibility criteria for systematic evidence? reviews focused on a specific research question (De Cres- (b) What areas are, or are not, well-researched? cenzo et al. 2017). Further, it was known that most of (c) Based on the available evidence, what advice can be the reported data were insufficient (in terms of amount offered to clinicians and investigators in the field? or quality) to merit inclusion in meta-analyses (Ng et al. 2015). As such, an evidence map offered a realistic option to gain an understanding of existing research in this area, Step 2: Description of the eligibility criteria (boundaries especially capturing insights from multiple small-scale for inclusion and exclusion), specification of co‑ordinates studies and mapping provides a means of summarizing to be mapped (study characteristics associated the range of publications and findings in a transparent with outcome/response, etc.) and definition of key variables way (Vallarino et al. 2015). The potential advantage over a (RAR metrics): basic narrative or selective review is that the search pro- (a) Studies were eligible if: cess is reproducible, and the map specifically addresses Inclusion criteria- gaps in the knowledgebase (Hetrick et al. 2010). However, findings are described in brief, so a map does not aim to i) The sample or cohort was wholly or partly comprised offer the detailed lists of citations and extensive tables of BD cases that met clinical or research diagnostic included in systematic reviews or meta-analyses. criteria ii) It reported associations between any RAR actigraphy Mapping process metric recorded at baseline and a clinical or research- The process usually incorporates four steps (clarifying defined outcome assessed at follow-up in BD cases focus; describing boundaries and ‘coordinates’ for map- (or vice versa). Note: For the purposes of the map, ping; literature search; charting the output maps). After the focus was on BD cases, so we extracted informa- Scott et al. Int J Bipolar Disord (2020) 8:37 Page 4 of 22 tion on inter-individual or intra-individual changes in of UP and BD; phase of illness at recruitment); commu- actigraphy or outcomes in the clinical cases (i.e. we nity or clinical setting (e.g. in- or outpatient); duration of aimed to avoid focusing on findings of case–control study; duration of actigraphy recording; and outcomes studies or cross-sectional comparisons of RAR met- reported. The latter could include: outcome prediction; rics between BD and HC, which have been reviewed response to introduction of treatment or treatment with- many times before). drawal; acute or long-term treatment effects; longitudi- iii) The publication reported any association between nal course of BD (continuous or repeated cross-sectional RAR metrics recorded by actigraphy and the out- assessment) in terms of illness progression (naturalist come or response of BD in the short or long term follow-up or associated with treatment introduction or (this could include e.g. naturalistic observational withdrawal), symptom exacerbation, relapse/recurrence/ studies of treatment interventions and/or rand- recovery/remission and/or putative disease course modi- omized controlled trials). fiers (as defined in a study). Also eligible were: naturalistic iv) Presence of comorbidities was not an exclusion crite- observational studies in clinician settings (unspecified/ rion, but details were noted. uncontrolled treatment), naturalistic monitoring of ill- v) Findings from self-report questionnaires, consumer ness with unspecified/uncontrolled treatment and/or grade wearables or smartphone apps were eligible if randomized controlled trials that incorporated actigra- the study also reported data for actigraphy record- phy (if not identified via previous criteria). ings and/or the RAR findings extracted from ques- (c) Definitions of Rest-Activity Rhythms. tionnaires or wearables were reported alongside the It is clear from actigraphy studies in BD that the met- equivalent metric derived from actigraphy. rics reported have differed by location (i.e. geography) and decade of study. However, researchers have repeat- Exclusion criteria- edly acknowledged the close connection between sleep The map aimed to capture the extent of the existing and circadian structure and that actigraphy data can be literature, so there were minimal exclusion criteria. The used for ‘rhythmometric analysis’ of RAR (Calogiuri et al. most significant criterion was that studies that failed to 2013; Smagula, 2016; Wirz-Justice, 2007). As such, all the report any measures (quantity and/or or timing) of day- metrics listed in Table 2 are considered as potentially rel- time activity were ineligible. Also excluded were single evant RAR markers in this project. case reports or studies where the recruitment of par- Table 2 gives an overview of the key RAR metrics that ticipants was not based on a current diagnosis of BD or have been derived from actigraphy recordings in psy- depression (UP and BD) e.g. studies of individuals at risk chiatry (definitions and descriptions of each metric are of BD or of individuals attending medical clinics (actigra- provided in Table 3 in Appendix) (Calogiuri et al. 2013; phy has been used to assess treatment outcome in CVD Ancoli-Israel et al. 2003). As shown, investigators have and cancer). focused on different rhythmometric procedures, which (b) Study Coordinates. need to be considered when interpreting the map. For It was agreed that the following would be mapped: instance, early studies were likely to report parametric year and geographic location of study (we also noted if statistics (especially in the USA) (Nelson et al. 1979). the study specified seasonality, etc.); sample characteris- Three variables (mesor, acrophase and amplitude) were tics (demographics; diagnoses e.g. BD subtypes, inclusion estimated using cosinor methods (with the p value Table 2 Potential measures of rest-activity rhythms (further details are given in Appendix) Parametric (Cosinor) Model Non-parametric Sleep Quantity Analysis Model Sleep Analysis- Circadian Model (Mean or Variability) Phase Model Midline estimating statistic of Inter-daily Stability (IS) Mean Activity (MA) Sleep Onset rhythm (MESOR) Amplitude Intra-daily Variability (IV) Time in Bed (TIB) (± Lights Off ) Sleep Offset Acrophase Least active period: L5/ L5 Onset Total Sleep Time (TST) Sleep Midpoint Most active period:M10/M10 Onset Sleep Onset Latency (SOL) Sleep Regularity Index (SRI) Amplitude Sleep Efficiency (SE %) Relative Amplitude Fragmentation Index (FI %) Wake after sleep onset (WASO) (± N awakenings > 5 min) Scott et al. Int J Bipolar Disord (2020) 8:37 Page 5 of 22 signifying the probability that the data really show circa- change in RAR (if combined with baseline or follow- dian periodicity), whilst more recently these metrics have up clinical measures). been derived using regression techniques (which report similar variables but assume more complex patterns and rhythms and are more robust for larger study popula- Step 4. Charting: screening and positioning the relevant tions) (Fernandez and Hermida 1998). Non-parametric evidence within the map methods report a wider range of variables to describe the Based on the specified aims of the study, we agreed three quantity and timing of activity and rest, and especially core outputs: provide insights into variability/stability of rhythms and any RAR disruptions (Calogiuri et al. 2013; Natale et al. a) Nature & quality of research: a Table of study coor- 2009). They are often preferred to parametric models and dinates was planned. Co-authors would review the it is argued that non-parametric models better represent Table independently and produce a brief written the complexity of RAR than cosinor models (van Some- synopsis of the key characteristics of the available ren et al. 1997). Variables derived from basic sleep analy- research. This information was used to formulate a sis (such as total sleep time: TST) are probably the most consensus on the nature and quality of the extant lit- widely reported measures in research in BD. One reason erature. for this is that these sleep quantity are much easier to b) Extent of research: two maps would be generated. estimate from raw data (and do not rely on more com- The first would describe year and country of study to plex algorithms). Sleep variables are useful for estimating provide a snapshot of research in this field by geog- duration and fragmentation of sleep patterns but are less raphy and over time. The second map would give reflective of circadian rhythmicity. Estimation of variabil- an overview of the most common methodologies ity in values for each sleep parameter is encouraged in and research themes. These would be identified by contemporary literature on RAR and greater reporting of extracting information regarding study coordinates sleep onset/offset/midpoint or sleep regularity index has and coding each of these separately, e.g. setting (e.g. been employed to give a greater insight into rhythmo- inpatient); type (e.g. naturalistic observational; treat- metrics (Bei et al. 2016). ment outcome; response criteria; etc.); reported interventions (e.g. proportion of individuals taking lithium; numbers allocated to psychological inter- Step 3: Literature Search vention); whether interventions were compared (e.g. The search approach was the same as utilized in sys- lithium versus quetiapine). These data could be com- tematic reviews. Search terms were applied to PubMed, bined to show the overall pattern of research. MEDLINE, PsycINFO, EMBASE, CINAHL and Web of c) A map would be generated to identify the range of Science databases No limits were set for language, type RAR parameters reported across all studies and of study or date of publication (the original telemetric which individual metrics appeared to be consistently motion sensors and actometers were used in BD in the found to be associated with course and outcome of early 1970s). The search terms included the full range BD. Such a map can offer a heuristic framework and of RAR metrics and potential outcomes listed in Step 2, be the starting point for a systematic review or meta- supplemented by searches using names of medications analysis at a future date. (e.g. lithium, carbamazepine) or for research groups known to have contributed publications when this topic initially emerged in the literature. This approach was Results used to ensure the broadest possible range of publica- Twenty-seven independent studies (reported in 29 pub- tions were identified. Furthermore, hand-searches were lications) were included in the map. Table 4 in Appendix undertaken of all reference lists of studies noted in nar- gives an overview of the studies and the core character- rative reviews, systematic reviews, and meta- analyses on istics. To maintain the simplicity expounded in evidence this or similar topics. The final list of eligible studies was mapping we have not included all the citations after every generated based on broad relevance to the scope of this statement (the citations are in the reference list and the map. Key drivers of selection of publications were that: key studies can be easily identified from Table 3). In the studies in the evidence map, the median age of a) The study reported the use of an actigraphy device to all participants was 45 years and 60% were female. The undertake daytime or 24-h monitoring of RAR in BD median sample size was 15 per study (range: 2–75); 484 b) The study endpoint was described in terms of a rec- individuals with BD (out of a total of 560) participated ognizable clinical outcome or measure of response or in actigraphy and the median duration of recordings was Scott et al. Int J Bipolar Disord (2020) 8:37 Page 6 of 22 7 days (range 1–210). Studies before 1990 tended to be most relevant to specific psychiatric disorders. Review- smaller, but sample size did not appear to be determined ing publications after 2006 suggests this this trend started by decade of publication, with several small-scale pro- to reverse when BD researchers increased their focus on jects or BD subsamples (included in larger samples or sleep–wake cycles and also with the rise in interest in cohorts) being followed-up longitudinally in recent stud- precision diagnostics and therapeutics. ies. Overall, 17 studies comprised wholly or partly of As shown in Fig. 2, only 22% studies reported follow-up inpatients (63%). Two studies reported daytime activity assessments at > 6 months (N = 6). Naturalistic observa- monitoring only, one study used actigraphy primarily to tional studies and treatment outcome studies are equally confirm reported patterns of hypersomnia whilst a RCT common (N = 10; 37%). In nine studies, > 50% of the sam- examined changes in objective levels of morning activity ple were taking lithium or the study specifically explored following a psychological intervention targeted at sleep RAR and lithium treatment. inertia. Although our preliminary work for the evidence-map There was evidence that statistical analyses of RAR identified > 30 possible RAR parameters (parametric have become more sophisticated over time, but earlier and non-parametric measures; mean or variable meas- studies were more likely to address directly the asso- ures of sleep analysis; putative circadian phase markers ciation between RAR and phase of BD (e.g. transitions extracted from sleep quantity analysis), the eligible stud- between mania, and depression), and change in RAR fol- ies only consistently reported estimates for seven metrics lowing initiation or withdrawal of treatment (especially (see Fig. 3). Early studies mainly reported cosinor metrics lithium) (e.g. Kripke et al. 1979; Wolff et al. 1985). The (amplitude, acrophase and mesor), whilst the more recent sampling, design, and analytic strategies of most studies studies increasingly report all raw data (either in the main demonstrate several methodological flaws (e.g. repeated text of the study or Appendices). However, in the middle testing of small samples with correction, absence of decades, there was a tendency to only report statistically reporting of non-significant outcomes). Overall, the con- significant findings, without any indication as to whether sensus judged the quality of the studies to be modest. other parameters were measured and/or which were Figure 1 shows that > 50% studies (14 of 27) were found to be non-significant. Interestingly, when number undertaken in the USA and it is only in the last decade of studies measuring a parameter is compared with the that the number of studies undertaken across Europe likelihood that a statistically significant association would has matched the USA. The earliest studies reported the be found, five variables were associated with outcome seminal work undertaken by Kupfer and associates and in > 50% of the studies. Namely: Variability/Rigidity of pursued by researchers at NIMH; more recent studies RAR (4 out of 4), 24 h rest activity cycle (22 out of 25), represent projects exploring RAR as disease course mod- amplitude (5 out of 6), acrophase/phase advance (5 out of ifiers or examining RAR as potential predictors of treat- 7), and sleep efficiency (SE: 5 out of 8). In contrast, most ment response and/or to monitor treatment outcomes. measures of sleep quantity show weaker links: wake after Table 3 also demonstrates that there were no studies (meeting our eligibility criteria) that reported RAR in BD for more than a decade (1994–2006). This is most likely a consequence of the fact that psychiatry research shifted towards the reporting of metrics favoured by experts in sleep/sleep disorders, rather than focusing on metrics 6 5 4 3 2 1 0 Fig. 2 Key coordinates of actigraphy studies – the graph shows the =<1980' >1981' >1991' >2001' >2010' main characteristics or reported purpose of the publications. Note: USA EUROPE NEW ZEALAND SOUTH KOREA Only descriptors reported in >  = 5 studies are included in the chart. Fig. 1 Location of study & decade of publication (Total = 27). (Note: However, the Total N reported in the graph exceeds the number if > 1 publication was identified relating to a dataset from a particular of citations (n = 28) as most studies are counted in > 1 descriptor location, we included the date of the first publication only) category Scott et al. Int J Bipolar Disord (2020) 8:37 Page 7 of 22 25 100% 20 80% 15 60% 10 40% 5 20% 25 9 8 8 7 6 4 0 0% N Studies Reporng Parameter % Studies Reporng Associaon with Outcome Key- SE: Sleep efficiency; SOL: Sleep onset latency; TST: Total sleep time; WASO: Wake after sleep onset Fig. 3 Comparison of number of studies reporting each actigraphy metric & proportion of those studies (%) in which the parameter shows a significant association with an outcome of interest sleep onset and sleep onset latency (WASO/SOL: 3 out treatments may have different effects on phase in BD-II, of 8) and TST (1 out of 9) being the only metrics showing and phase shifts may can be associated with outcome at any significant associations. follow-up. The variable 24 h rest-activity cycle represents It was not possible to provide specific evidence of the a global estimate of RAR and its primary value to the map nature of the metrics that were most useful in different is that it confirms that RAR can be used to assess a range types of studies, although there was some trend of inter- of clinical outcomes in BD. Amplitude appeared to be a est. Using information précised in Table 3, available useful marker of illness course and treatment response evidence largely supports the view that mania is associ- or outcome. It was noted that variability/rigidity of RAR ated with circadian phase advance and BD-depression was only estimated in four small samples (Ns = 3, 8, 10, with phase delay. The latter was not obvious in studies of 12). These studies reported extended continuous dura- mixed samples of UP and BD depression. Interestingly, tions of actigraphy and found the metrics to be significant treatment may be associated with phase change, and markers of longitudinal course (including admissions). Scott et al. Int J Bipolar Disord (2020) 8:37 Page 8 of 22 However, the studies did not report sufficient raw data or findings are of interest, encourage pursuit of research in basic analyses to allow further investigation or interpre- this field, but clearly expose that the quality of evidence tation of the findings. and available published studies is modest. Next, the second question addressed what areas are, Discussion or are not, well-researched? Our findings about the This article explores associations between RAR metrics extent, nature and quality of the research, indicate that measured via actigraphy and course, outcome and/or the available publications demonstrate the diverse ways treatment response in BD. The project arose from three in which RAR can be studied in the context of BD, but converging strands of work. First, a background scop- that no single area has been targeted for a prolonged ing exercise undertaken for the H2020-funded consor- period of research. Notably, the early studies showed a tium exploring a range of putative markers of lithium keen interest in treatment response (or change in illness response phenotypes (including actigraphy) (Scott et al. with treatment withdrawal), but these themes were not 2019); second, scrutiny of RDOC publications that sug- pursued to a definitive conclusion. However, this theme gest undertaking longitudinal actigraphy to explore RAR is now drawing interest again and two recent studies phenotypes in BD (Insel 2014); and third, a study of cir- used actigraphy to explore RAR and acute treatment out- cadian and sleep–wake cycles in lithium responders and come, namely ketamine infusions for depression (albeit non-responders (Scott et al. 2020). These projects pro- in sample with many more UP than BD cases) and inpa- vided us with sufficient insights into this research field tient treatment of mania (such as Blue Blocking Glasses) and, alongside our prior knowledge, led us to conclude (Duncan et al. 2017; Henriksen et al. 2020). The map that the existing literature on RAR and course/outcome/ also highlights evidence that clinical depressive symp- response was unlikely to be suitable for systematic review toms are associated with robustness of circadian rhythm or meta-analysis. Furthermore, it appeared it would be (e.g. Hwang et al. 2017), whilst current RAR patterns counter-productive to use these classic evidence-based may predict future mental state (Salvatore et al. 2008). techniques as they would prematurely restrict the focus Also, in a small inpatient sample, it was shown that RAR of our investigation to a specific or narrow question of recorded by actigraphy were associated with clinical pro- what presently seems to be an imprecise research area. gress, whilst estimates made with a consumer wearable As such, it was agreed that utilizing evidence mapping were not (Averill et al. 2019). This is a useful reminder might enable some synthesis of the diverse range of (pre- that research grade devices are still more accurate than dominantly small-scale) studies currently available. consumer grade device and apps (and so actigraphy is Regarding our first question (what evidence exists and likely to continue to have a role in the immediate future). what conclusions can be drawn about the quality of evi- Importantly, individuals with BD are prepared to wear dence?). We conclude that limited evidence exists, but actiwatches and/or consumer devices for extended peri- what is available indicates that a discrete number of RAR ods (> 6 months) and this, allied with new conceptual- metrics are more consistently associated with transition izations of and analytic approaches to of RAR (such as between different phases of BD and/or may be predictive fractal activity as an indicator system adaptability and of longitudinal course of illness or treatment response. lagged data analyses, etc.), may lead to a greater under- The metrics that show the most frequent associations standing of the chronological sequence of change of represent markers of the amount, timing, or variability RAR, mood, cognition and other cardinal features of BD of RAR rather than the sleep quality metrics that are fre- (Knapen et al. 2020; Walker et al. 2020). quently targeted in contemporary studies of BD. How- The final question asked, based on the available evi- ever, these putative ‘circadian’ signals should be viewed in dence, what are the potential ways forward for the field? context, as there is evidence of selective reporting of both We suggest that one important option would be to metrics and outcomes, thus increasing the likelihood develop a consensus within research consortia or wider of publication bias. Notably, the statistical associations groups of clinicians and investigators regarding the most reported derive from multiple small-scale studies that appropriate RAR metrics to include in future actigra- together included < 500 individuals with BD. In sum, the phy studies of BD (Scott et al. 2017). It is particularly Scott et al. Int J Bipolar Disord (2020) 8:37 Page 9 of 22 important that this dialogue focuses on metrics that offer scrutinized not only to identify mean sleep duration and insights into RAR in BD that will have the optimal clini- timing but also variability between weekdays and week- cal as well as research utility for this field. This should be ends or large differences in duration of sleep onset/offset/ prioritized over simply generating a list of RAR measures duration over 2–3 weeks. recommended for estimation or reporting of sleep or Whilst consumer grade wearables and smart phone sleep–wake research in general (Sack et al. 2007). These apps are attractive to many, it is important to remember discussions will necessarily consider any RAR metrics that they all await regulatory approval for use as clinical that may represent trait or state variables (e.g. there is devices (and commercial developments of longer bat- emerging evidence that relative amplitude, amplitude, tery life to avoid repeated need for charging). However, interdaily stability and intradaily variability may be their use alongside actigraphy could be helpful, both to markers of disease onset as well as disease progression) validate different measures of RAR that might be derived (Merikangas et al. 2019; Scott et al. 2020). Developing a from consumer devices but also to facilitate ecological consensus is perhaps the most important recommenda- momentary assessment (EMA) research (Merikangas tion that arises from this mapping exercise, as without et al. 2019; Depner et al. 2020). Research in this area is such a course of action there is a clear possibility that we ongoing and has the advantage of combining self-report will continue to lack the means to make detailed cross- ratings of events, cognition, energy, and mood states with study comparisons of the metrics that would offer most objective RAR data (de Wild-Hartmann et al. 2013; Meri- promise as diagnostic or response biomarkers. kangas et al. 2019). However, these advances necessarily An obvious gap in the available literature is the lack require researchers to work more closely alongside stat- of youth studies of actigraphy that take a longitudinal isticians and computational data scientists as the optimal perspective. Given the interest in staging models and methods for analyzing and understanding links between temperament, research on the transition between sub- RAR, symptoms and behaviours are more sophisticated threshold conditions or at-risk mental states and the (including non-linear dynamics, entropy) than the basic onset of full-threshold illness episodes would be wel- approaches that are most commonly used and/or the come. In individuals with established BD, consideration algorithms that are currently available. Assuming this should be given to RAR metrics that might be com- dialogue can occur, combining these approaches would bined in multi-platform research of precision diagnos- allow to comprehensive exploration of the course of ill- tics and therapeutics (Scott et al. 2019). For instance, ness, prediction of treatment response and a range of precision studies that focus on prediction of response to patient reported outcomes (Moskowitz and Young, 2006; mood stabilizers such as lithium, and the identification Marino et al. 2013). For clinicians, actigraphy offers of biosignatures (rather than single biomarkers) might access to objective ecological monitoring that provides a consider giving priority to reporting RAR markers such far more reliable and valid measure of RAR than can be as amplitude and variablility (given our mapping findings obtained by self-rating scales (Depner et al. 2020; Mul- regarding emerging evidence for their importance). The ligan 2016). However, not all clinicians feel confident in mapping exercise demonstrated that actigraphy stud- extracting data and metrics from the devices and this is ies are viable in inpatient as well as outpatient settings, likely to mitigate against wider dissemination beyond as such, wider consideration should be given to monitor- specialist settings. Another advantage of actigraphy is ing inpatient progress using objective measures such as that it is a more feasible option for routine clinical prac- actigraphy, with the option of continuation to post-dis- tice than putative ‘omic’ and brain scanning markers of charge settings. Of course, longitudinal monitoring that treatment response or illness outcome as these are less allows early warning signs of relapse or recurrent is also available, more expensive and often concentrated in feasible with this ecological measure (Bellivier et al. 2015; research settings (Gooley and Chua 2014). Also, if the Ritter et al. 2011). The above steps would address some focus is on RAR, then actigraphic monitoring would be of the explicit aims outlined by RDOC publications but easier to integrate into a clinical management and moni- could be valuable for clinicians e.g. they might consider toring package than evaluation of other objective circa- using electronic monitoring or including sleep diaries as dian markers e.g. measurement of dim light melatonin part of the clinical assessment of BD. The latter could be onset (DLMO) (Depner et al. 2020). Scott et al. Int J Bipolar Disord (2020) 8:37 Page 10 of 22 Conclusions and charts. Scott, Etain and Colom created the first full draft of the manuscript. All authors read and approved the final manuscript. Despite 50 years of research, and several papers iden- tified by this evidence map being highly cited, use of Funding actigraphy to assess RAR in longitudinal research and FB is the Principal Investigator and all the authors are investigators on the R-LiNK (Response to Lithium Network) project. This project has received fund- examination of these metrics and treatment response, ing from the European Union Horizon 2020 research and innovation program course and outcome of BD is under-investigated. This is ((EU.3.1.1. Understanding health, wellbeing and disease: Grant No 754907). in marked contrast to the extensive literature on case– BE received funding from: Institut National de la Santé et de la Recherche Médicale (INSERM – Research Protocol C0829), Assistance Publique des Hôpi- control or cross-sectional studies of actigraphy, espe- taux de Paris (APHP – Research Protocol GAN12). FC received funding from: cially typical sleep analysis metrics in BD, which are Spanish Ministry of Economy and Competitiveness, through a FIS (PI15/00588; reported in independent studies, systematic reviews, PI19/00009), and the Secretaria d’Universitats i Recerca del Departament d’Economia i Coneixement de la Generalitat de Catalunya, Government meta-analyses and meta-regressions. The most obvi- of Catalonia (2017_SGR_134). AY received funding from: National Institute ous reason for this marked disparity in the amount of for Health Research (NIHR) Biomedical Research Centre at South London research undertaken using these different methodolo- and Maudsley NHS Foundation Trust and King’s College London. The views expressed are those of the authors and not necessarily those of the NHS, the gies, is that case–control and cross-sectional studies NIHR, or the Department of Health. are easier to undertake, usually involve much briefer time frames, and have lower resource requirements Data availability statement Data sharing is not applicable to this article as no datasets were generated or than longitudinal research. Also, it was notable that analysed during the current study. All the data referred to in this manuscript prospective studies identified in this evidence map were are publicly available in the manuscripts describing the original studies. still predominantly shorter-term projects, often report- Ethics approval and consent to participate ing outcomes over weeks rather than months. The authors assert that all procedures contributing to this work comply with This evidence map suggests that the existence of sev- the ethical standards of the relevant national and institutional committees on eral often-quoted publications has probably led to an human experimentation and with the Helsinki Declaration of 1975, as revised in 2008. overestimation of the extent and robustness of existing evidence of longitudinal associations between RAR and Consent for publication treatment response, course, and outcome in BD. How- Not applicable. ever, we are aware of promising emerging findings in Competing interests this area (McCarthy et al. 2019; Scott et al. 2020) and The authors report no financial affiliation or other relationship relevant to the there is tremendous potential for expanding this field of subject matter of this article. JS is a Visiting Professor at Diderot University, Paris and an Honorary Professor at Kings College, London. AY declares paid research to inform precision medicine projects in psy- lectures and advisory boards for the following companies with drugs used chiatry. Further, there is a strong argument that, rather in affective and related disorders: AstraZeneca, Eli Lilly, Lundbeck, Sunovion, than actigraphy being replaced by consumer grade Servier, Livanova, and Janssen. He is a paid consultant to Johnson & Johnson, is lead Investigator for Embolden Study (AstraZeneca), BCI Neuroplasticity devices and apps, it may be more fruitful to establish study and a pharma sponsored Aripiprazole in Mania study. His institution has frameworks for clinical and research projects that com- received funding for investigator-initiated studies from AstraZeneca, Eli Lilly, bine these approaches. Linova, Lundbeck, Wyeth, and Janssen. Author details 1 Institute of Neuroscience, Newcastle University, Newcastle upon Tyne, UK. Abbreviations 2 Centre for Affective Disorders, Department of Psychological Medicine, Insti- BD: Bipolar disorders; CVD: Cardiovascular disease; d: Days; DALYs: Disability tute of Psychiatry, Psychology & Neuroscience, King’s College London, London, adjusted life years; DLMO: Dim light melatonin onset; EMA: Ecological momen- UK. 3 Université Paris Diderot, Sorbonne Paris Cité, UMR-S 1144, 75013 Paris, tary assessment; HC: Healthy controls; IS: Inter-daily stability; IV: Intra-daily France. 4 IMIM-Hospital del Mar-CIBERSAM, Barcelona, Catalonia, Spain. variability; Li: Lithium; N: Number; NIMH: National Institute of Mental Health; 5 Universitat Autònoma de Barcelona Barcelona-Catalonia, Barcelona, Spain. RDOC: Research domain criteria; RAR​: Rest activity rhythms; SD: Standard 6 Département de Psychiatrie Et de Médecine Addictologique, AP-HP, GH deviation; SE: Sleep efficiency; SOL: Sleep onset latency; TST: Total sleep time; Saint-Louis – Lariboisière – F. Widal, 75475 Paris, France. 7 Inserm U114475006, UP: Unipolar; WASO: Wake after sleep onset; wks: Weeks; yrs: Years. Paris, France. 8 South London and Maudsley NHS Foundation Trust, Bethlem Royal Hospital, Kent, UK. Authors’ contributions All authors identified the key questions and coordinates for the map. Scott collated papers for the evidence map, Etain and Scott created summary tables Scott et al. Int J Bipolar Disord (2020) 8:37 Page 11 of 22 Appendix See Tables 3 and 4. Table 3 Parameters reported in studies of Rest–Activity Rhythms: additional details Cosinor model (parametric) Midline estimating statistic of rhythm (MESOR) Estimates the arithmetic mean of the values throughout the 24-h period Amplitude Distance from the rhythmic mean to the peak or trough of a mathematical model Acrophase Time interval during which the highest values are expected Nonparametric approach Interdaily stability Quantifies the degree of resemblance between the activity patterns on individual days; ranges from 0 to 1 and may typically be about 0.6 for healthy adults Intradaily Variability Quantifies the fragmentation of periods of rest and activity; ranges from 0 to 2 and typically is < 1 for healthy adults, with higher values indicating a more fragmented rhythm L5 Average of the activity values for the 5 least active consecutive hours in the 24-h cycle L5 Onset Onset time of the five most restful consecutive hours M10 Average of the activity values for the 10 most active consecutive hours in the 24-h cycle M10 Onset Onset time of the 10 most active consecutive hours Amplitude Difference between M10 and L5 Relative Amplitude Calculated by dividing AMP by the sum of L5 and M10; ranges from 0 to 1, with higher values indicating higher amplitude of the rhythm Sleep quantity analysis (reported as Individual Mean values or Intra-Individual Variability e.g. SD, RMSSD) Time in bed (TIB) Difference between the get-up time and bedtime (Lights Off (LO)) (Lights Off: time bedroom lights are turned off is reported in some therapy studies) Total sleep time (TST) Duration of the sleep episode, calculated on the actual time of sleep start and sleep end Sleep latency (SOL) Difference between time in bed and the actual time the monitored subject starts sleeping Mean activity (MA) Number of movements within specified epoch (usually 1 min) Number of awakenings longer than 5 min (NA > 5) Number of awakenings longer than 5 min Wake after sleep onset (WASO) Wake time (min) after sleep onset Sleep efficiency (SE %) Percentage of effective sleep during time in bed Fragmentation Index % of sleep continuity/discontinuity. It is calculated as the amount of time associated with move- (FI %) ment (restlessness) during the sleep period expressed as a percentage Putative Circadian Phase Measures (derived from Sleep quantity analysis data) Sleep Onset Time is the clock time (HH:MM) of the first epoch scored as sleep in each main rest interval Sleep Offset Mean sleep offset time is the clock time (HH:MM) of the last epoch scored as sleep in each main rest interval Sleep midpoint Calculated as the midpoint between sleep onset and sleep offset Sleep Regularity Index (SRI) This is defined as the percentage probability of a person being asleep (or awake) at any two time points 24 h apart Cosinor model descriptions from Nelson et al (1979); nonparametric approach descriptions from Van Someren et al (1997). Sleep quantity analysis variables and descriptions as from Natale et al (2009), Krane-Gartiser et al. (2019) and Bei et al (2016) IS quantifies regularity in activity (overall variability in a time-signal after averaging across days) & IV quantifies fragmentation; some argue that SRI best quantifies how rapidly sleep patterns change between consecutive days Table 4 Studies of actigraphic recordings of treatment outcome or course of illness in samples that include individuals diagnosed with bipolar disorders (BD) Study: Sample characteristics Actigraphy: number of days (d)a Findings and comments 1st author, location, year N (BD) Total N*, Mean Age (SD) yrs; Diagnoses % Female Kupfer, 4 N = 11; ~ 52 b; 3d at baseline & 3d at wk 3 Baseline (pre-treatment) mean levels of day- USA time motor activity in un-medicated unipolar Scott et al. Int J Bipolar Disord Bipolar Depression (N = 4) & Unipolar Depres- 80% (1974) sion (n = 7) depression > BD depression, with significant differences between 19:00 & mid-day (but not mid-day until 19:00) BD cases treated with lithium; UP cases treated with amitriptyline Post-treatment activity differences were (2020) 8:37 no longer evident; (response in Unipolar Depression showed the greatest improve- ment in activity) Weiss, 6 Mania (N = 3) & Bipolar Agitated Depression ~ 38 b; 7-9d; additional recording 3 manic inpatients were assessed pre-treat- USA (N = 3) 66% at 20 wks ment; then all 6 inpatients were assessed (1974) before & after initiation of Lithium. Some individuals repeated actigraphy recordings at 5 months Baseline (pre-treatment) levels of motor activ- ity (counts per minute) in un-medicated mania > agitated depressed, with the most significant differences reported between 22:00 and 02:00 h Heninger, 13 N = 15; 55b 2d pre-lithium & 2d at 2 wks All participants were inpatients who initiated USA BD-I = 11; 84% lithium treatment (1977) BD-II = 2; Unipolar Disorder = 2 Duration of sleep increased by > 60 min after starting lithium (~ 6.3 h to 7.5 h). When data were reviewed for mania and depression separately, individuals with mania showed reduction in motor activity in parallel with symptom reduction, whilst individuals with depression showed increased motor activity with reduction in symptoms Kripke, 7 BD: 28% b 3d Adults assessed through complete cycles USA across episodes (repeated) of mania and depression. In 5 individuals (1978) there was evidence that a circadian rhythm ‘free-ran’ fast & in 5 individuals, evidence suggested that lithium slowed a circadian rhythm. The authors concluded that the pro- phylactic benefit of lithium may derive from slowing or delaying an overfast circadian clock to prevent desynchronization Page 12 of 22 Table 4 (continued) Study: Sample characteristics Actigraphy: number of days (d)a Findings and comments 1st author, location, year N (BD) Total N*, Mean Age (SD) yrs; Diagnoses % Female Wehr, 10 Mania & ~ 40b 2d Motor activity levels were significantly different USA Bipolar Depression across 24 h and indicated ‘phase advance’ Scott et al. Int J Bipolar Disord (1980)# of 1–3 h in mania or depression (compared with 14 HC) Acrophase for motor activity was significantly earlier in Mania (13:37 h) & Depression (14:10 h) compared to HC (15:57 h) Acrophase did not differ in 8 BD cases that (2020) 8:37 were assessed in both the Manic & Depres- sive phases of illness Godfrey, 4 N = 9; 77%b Median Inpatients wore an actiwatch during the New Zealand Unipolar Disorder = 5; ~ 12d daytime only for 7-27d of the admission. (1984) BD-II, Hypomania = 4 Data were also collected on 9 HC. Daytime activity level was lowest in Depression and highest in HC. During actigraphy, mean day- time activity level increased in Depression & reduced in Hypomania. Change in activity was not correlated with change in depressed mood (no measures reported for hypoma- nia). However, patients continued to show much lower levels of activity compared with HC. (The study abandoned some speech recordings due to low cooperation by some hypomanic patients) Wolff, 30 Mania, 42 ± 13; 66% 3d Intra-individual recordings demonstrated that: USA (1984)# Bipolar Depression or Euthymia Mean amount of motor activity in Mania > Depression with the most significant differences being early & later in the daytime Activity levels did not differ significantly in Mania versus Euthymia Total activity (count/24 h) in 18 HC > Euthymic BD Page 13 of 22 Table 4 (continued) Study: Sample characteristics Actigraphy: number of days (d)a Findings and comments 1st author, location, year N (BD) Total N*, Mean Age (SD) yrs; Diagnoses % Female Joffe, 15 N = 19, 43 ± 3; 58% 3d pre- & 3d post-initiation of -carbamaz- Inpatients were included who began carba- USA epine mazepine treatment after a placebo washout Scott et al. Int J Bipolar Disord BD Depression = 15; (1987) Unipolar Disorder = 2; of > 2 wks; actigraphy was undertaken during Schizoaffective = 1 washout and ~ 35 days later. 7 individuals (37%) met response criteria. Baseline patterns of motor activity were not significantly differ- ent in Responders vs. Non-responders Mean 24 h activity increased over time (& (2020) 8:37 mean depression score decreased), with a trend toward greater change in Responder. When time of day was considered, Respond- ers showed higher activity from 13:00 h onwards, with significantly greater activity during evening & night-time Klein, 4 BD, Not ­reportedb 3d with lithium; Long-term lithium was discontinued, and USA in remission 3d at ~ 3 wks post-lithium withdrawal outpatients were followed for 3 months of (1991) placebo treatment and up to 12 months. Two individuals experienced manic relapse with symptoms commencing within 2 wks of discontinuation, 2 remained well. Relapse was characterized by fragmentation of sleep–wake patterns, disorganization of 24 h rest-activity cycles & increased motor activity, especially at night. The latter was accompa- nied by a marked reduction in SE Klein, 10 BD, 49 ± 12b 3d with lithium; As with the previous publication, long-term USA in remission 3d at ~ 3 wks post- lithium withdrawal lithium was discontinued, and outpatients (1992) were followed for 3 months of placebo treatment. Seven experienced manic relapse: 4 within 3 weeks and all 7 within 3 months of lithium discontinuation. These individuals had higher baseline levels of motor activity (daytime) even though asymptomatic and showed higher daytime & night-time activity (during follow-up) than individuals who remained well (N = 3) Page 14 of 22 Table 4 (continued) Study: Sample characteristics Actigraphy: number of days (d)a Findings and comments 1st author, location, year N (BD) Total N*, Mean Age (SD) yrs; Diagnoses % Female Raoux, 5 N = 26; 44 ± 21; 77% 3d treatment-onset & Inpatients were assessed at beginning & end France, 3d pre-discharge of admission for melancholic depression Scott et al. Int J Bipolar Disord Unipolar Depression = 21; (1994) BD Depression = 5 (median stay = 26d). All patients had current or past exposure to benzodiazepines. Diurnal hypoactivity, reduced amplitude & longer TST were found at admission. Overall activity level and amplitude increased by discharge. There was an inverse correlation between (2020) 8:37 baseline activity level & clinical improve- ment (daytime hypoactivity), with a trend for Responders (N = 21) to a lower baseline activity levels than Non-responders (N = 5). Amplitude was a robust marker of response, but acrophase was difficult to interpret because of large intragroup phase variability (~ 25% showed a phase shift of ~ 1 hr) Todder (2006) & Baune, 6 N = 26; 49 ± 13; 58% 2 × 11d Todder reported 4-week follow-up of patients Germany Treatment-resistant depression: with 6d gap & 27 HC. Sleep parameters in patients (2007) Unipolar = 21; showed significant change in the 1st week BD-II = 6 of follow-up (increased TST & reduced SOL). Even with depression improvement, many sleep parameters differed from HC Baune reported the 4-week study of anti- depressant & quetiapine. At final follow-up, 16 individuals were classified as Responders. Amount of activity had increased signifi- cantly at follow-up (& depression score had decreased) Benedetti, 39 BD-I 44 ± 10; 54% 7d Study examined antidepressant effects of sleep Italy Depression deprivation & light therapy; 14 individuals (2007) were receiving lithium (36%) Responders (N = 26) & Non-responders (N = 13) did not differ on actigraphy param- eters at baseline (although mesor was higher in Responders). Responders showed increase in daytime activity especially in the morning compared with Non-responders. Total sleep was shorter in Responders. Acrophase was advanced in Responders (of ~ 57 min) but delayed by 5 min in Non-responders Page 15 of 22 Table 4 (continued) Study: Sample characteristics Actigraphy: number of days (d)a Findings and comments 1st author, location, year N (BD) Total N*, Mean Age (SD) yrs; Diagnoses % Female Salvatore, 36 BD-I: Mania/ Mixed States & Remission 44 ± 10; 81% 3d in episode & 3d in remission BD-I cases were in and outpatients; 72% were USA (2008)# prescribed lithium Scott et al. Int J Bipolar Disord 36 acutely ill BD-I cases were compared with 32 HC at baseline. Then recordings from the BD-I cases when ill & in remission were compared Manic cases showed phase advance (~ 2 h) compared with HC & lower amplitude (2020) 8:37 (mesor & TST did not differ from HC) Mean daytime activity (counts/minute were higher in HC (67 ± 10) than Mania/Mixed States (59 ± 32; but Mania ~ Remission (60 ± 11) Total sleep was significantly longer in remission (Remission > HC; HC = Mania/Mixed States) Acrophase was significantly earlier in Mania/ Mixed States (14:01hrs) < Remission (14:26 h) < HC (16:05 h) Scott, 10 N = 18 43 ± 12; 30-200d Naturalistic longitudinal follow-up of indi- UK Unipolar Disorder = 8; 60% viduals attending a lithium clinic who were (2011)$ BD-I = 10 invited to wear actiwatch for > 1 month (many continued for > 6 months) Lower mean daytime activity (counts/min- ute) and variability in sleep pattern/dura- tion (alongside psychological measures such as self-esteem) was associated with shorter time to hospital admission dur- ing ~ 15-month follow-up (but admissions according to polarity were not reported) Page 16 of 22 Table 4 (continued) Study: Sample characteristics Actigraphy: number of days (d)a Findings and comments 1st author, location, year N (BD) Total N*, Mean Age (SD) yrs; Diagnoses % Female Kim, 29 BD-II = 26; 36 ± 10; 55% 56d Open label trial of lithium v. quetiapine (actual (2014) & Hwang (2017) BD Depression N for each group included in analyses varied Scott et al. Int J Bipolar Disord South Korea slightly between the two publications). Individuals were drug free at baseline Quetiapine group (N = 10–12): SE increased (~ 6%) & WASO decreased at 6 weeks follow- up & SE improvement maintained at final follow-up (8 weeks) (2020) 8:37 Lithium group (N = 15–17): Modest improve- ment in SE (~ 1.5%), other parameters showed some fluctuation without obvious improvement Both quetiapine & lithium affected several circadian parameters, including peak activity time and robustness of circadian rhythm, but exerted different effects on acrophase; clini- cal depressive symptoms were associated with robustness of circadian rhythm during the 8-week treatment phase Novak, 6 BD-I ~ 37 b; Median 8 individuals were recruited, 7 agreed to par- Czech Republic 66% ~ 210d ticipate and 6 of them provided long-term (2014) recordings (5–80 months, but about 30% loss of data). 4 individuals reported no relapses, one reported 3 episodes (over 37 months) & one reported 14 (over 80 months). No manic relapses occurred The best signal for identifying relapse was for IS (inter-daily stability)- a marker of regularity of rest-activity patterns (overall classifica- tion ~ 66%) Page 17 of 22 Table 4 (continued) Study: Sample characteristics Actigraphy: number of days (d)a Findings and comments 1st author, location, year N (BD) Total N*, Mean Age (SD) yrs; Diagnoses % Female Kaplan, 75 N = 159, 36 ± 11; 7d 159 individuals with BD spectrum disorders USA 66% were assessed for hypersomnia & followed- Scott et al. Int J Bipolar Disord BD- I = 143, (2015) BD-II = 13; up at 7 months later (222 ± 73d). Baseline BD NOS = 3 actigraphy recordings in 75 individuals identified 2 forms of hypersomnia: ‘long sleepers’ (although actigraphy showed these individuals spent longer in bed but did not have longer TST) and those with excessive (2020) 8:37 sleepiness (lower daytime activity) During follow-up, 21 individuals (17.5%) had a hypomanic/manic & 37 (30.1%) a major depressive episode. Self-reported ‘Excessive daytime sleepiness’ predicted relapse into hypomania/mania (p < 0.01); actigraphy was not used to directly assess objective markers of relapse Gershon, 37 BD-I 35 ± 10; 42d Lower amount of daytime activity & later daily USA 62% onset of activity were both associated with (2016) greater likelihood of onset of a depressive episode &/or increase in depressive symp- tom severity. Weaker evidence suggested a slightly elevated mid-day peak & less even- ing activity were also noted Moon, 26 BD-I = 23 30 ± 11; At admission; 2 wks, 4 wks & recovery All BD cases were recruited during a hospital South Korea Manic = 21; Depressed = 4. (Manic & Depres- 62% admission. Analyses included 26 manic and (2016) sive Episodes = 1) 5 depressive episodes. Final wake time was delayed by ~ 1 h and SE improved by ~ 4% during admission Advanced phase was recorded in mania but returned to normal after treatment; likewise, delayed phase in depression returned to normal after treatment. These shifts led to BD cases more closely resembling HC (N = 18) Phase advance or delay observed in biochemi- cal rhythms (such as cortisol) were much greater than sleep–wake cycle changes identified by actigraphy Page 18 of 22 Table 4 (continued) Study: Sample characteristics Actigraphy: number of days (d)a Findings and comments 1st author, location, year N (BD) Total N*, Mean Age (SD) yrs; Diagnoses % Female Bewernick, 2 N = 16; 49 ± 9; 7d 12 individuals (10 inpatients) completed the Germany 66% study. Actigraphy was not undertaken during Scott et al. Int J Bipolar Disord Unipolar Depression = 14; (2017) BD Depression = 2 sleep. Individuals with higher activity levels at baseline showed a larger reduction in symptoms of depression after one week. Change in severity of depression correlated with increased in amount of daytime activity (with significance for some but not all rat- (2020) 8:37 ings). Activity also showed a lagged correla- tion with mood rating in the following hours (but not vice versa) Duncan, 21 N = 51; ~ 42; 9d: Data from 51 trial participants receiving keta- USA Bipolar Depression = 21; 59% (4-days pre- & 5 days post-ketamine mine (21 responders; 30 non-responders) (2017) Unipolar Depression = 30 were compared with data from 38 who received placebo Pre-ketamine infusion: evidence of phase advanced activity pattern (acrophase) & lower mesor (144 vs. 167) was associated with good response. Persistent response (day 1 & 3 post-infusion) was characterized by higher amplitude & increased mesor Krane-Gartiser, 3 Series of BD-I cases with >  = 2 admissions ~ 50b 1d per admission Intra-individual 24-h activity patterns for Norway (2018) (over 1–6 months) recurrences of the same polarity show similar patterns; but episodes with different polarity show intra-individual differences in activity patterns. Acute phases were characterized by greater variation between rest-activity and periods with distinct irregularity in activity (measured by sample entropy) in all cases Averill, 2 N = 24; 38 ± 14; 17d Inpatients were recruited to a study of activity, New Zealand (2018) 22 Unipolar Depression; 37% psychomotor speed & treatment response; 2 BD Depression usable actigraphy data available from 16 individuals (2 receiving lithium) Change in 24-h activity count was correlated with treatment response Data from consumer grade activity monitor did not correlate with outcomes Page 19 of 22 Table 4 (continued) Study: Sample characteristics Actigraphy: number of days (d)a Findings and comments 1st author, location, year N (BD) Total N*, Mean Age (SD) yrs; Scott et al. Int J Bipolar Disord Diagnoses % Female Kaplan, 40 Inter-episode 39 ± 14; 14d 58 patients were randomized to a two-arm USA BD I 70% (7d pre- & 7d post-intervention) trial; 40 participated in a one session inter- (2018) (with insomnia) vention experiment actigraphy data were available on 29 (2020) 8:37 A behavioural intervention (Rise-Up: aimed at overcoming sleep inertia) was compared with a control condition (education: PE) Individuals in the Rise-Up group were sig- nificantly more active in the ­1st hour upon waking compared to PE Henriksen, 20 BD-I, 45 ± 13; 2d (baseline & day 5) The trail recruited 32 individuals: this study Norway (2020) Mania 70% reports data for 20; 10 inpatients were allo- cated to the blue blocking glasses (BBG) & 10 to placebo glasses 60% of all cases were receiving lithium At day 5: SE improved more in the BBG compared with the control group (4% v 0%). The BBG group showed lower activity count during the sleep period (L5) & lower WASO & a trend for lower FI Knapen, 14 BD-I 45 ± 12; ~ 180d Continuous monitoring of temporal relation- Netherlands 79% ship between sleep, daily rhythm (measured (2020) by actigraphy) and mood changes. Time- lagged association showed significant nega- tive correlation between depression and fractal patterns of motor activity (a measure of rigidity or adaptability of physiological systems) with higher level of depression associated with greater rigidity (with 5–7-day delay) %, Mean and SD: percentages, means and standard deviations are reported to the nearest whole number; BD Bipolar Disorder, BD-I/II/NOS Bipolar I/II Disorder, BD Not Otherwise Specified, d day, h hours, HC Healthy Controls, N number, NA not available, SD Standard Deviation, SE Sleep Efficiency, SOL Sleep Onset Latency, TST Total Sleep Time, yrs years. *Total N is reported only if sample contains cases with other diagnoses (as well as BD); a Estimated for some studies that did not specify exact time frames; bThis either indicates that data were reported as medians and/or data were missing for some variables; #The same sample of BD cases was studied in different mental states; $ Analysis reported by Scott (2011) is described in Scott et al. (2017) Page 20 of 22 Scott et al. Int J Bipolar Disord (2020) 8:37 Page 21 of 22 Received: 2 July 2020 Accepted: 25 August 2020 digit​al-healt​h/digit​al-healt​h-softw​are-prece​rtifi​catio​n-pre-cert-progr​ am). Accessed 20 May 2020. Frey B, Andreazza AC, Houenou J, Jamain S, Goldstein B, Frye M, et al. Biomark- ers in bipolar disorder: a positional paper from the International Society for Bipolar Disorders Biomarkers Task Force. Aust N Z J Psychiatry. Apr References ;47(4):321–32. Ancoli-Israel S, Cole R, Alessi C, Chambers M, Moorcroft W, Pollak CP. The GBD 2017 Disease and Injury Incidence and Prevalence Collaborators.Global, role of actigraphy in the study of sleep and circadian rhythms. Sleep. regional, and national incidence, prevalence, and years lived with 2003;26(3):342–92. disability for 354 diseases and injuries for 195 countries and territories, Arksey H. Scoping the field: services for carers of people with mental health 1990–2017: a systematic analysis for the Global Burden of Disease Study problems. Health Soc Care Community. 2003;11(4):335–44. 2017. The Lancet. 8 Nov 2018; 392:1789–858. Averill I, Beaglehole B, Douglas K, Crowe M, Inder M, Lacey L, et al. Activation Geoffroy P, Scott J, Boudebesse C, Lajnef M, Henry C, Leboyer M, Bellivier F, therapy for the treatment of inpatients with depression - protocol for a Etain B. Sleep in patients with remitted bipolar disorders: a meta-analysis randomised control trial compared to treatment as usual. BMC Psychiatry. of actigraphy studies. Acta Psychiatr Scand. 2015;131(2):89–99. 2019;19(1):52. Gershon A, Ram N, Johnson SL, Harvey AG, Zeitzer JM. Daily actigraphy profiles Baron K. Feeling validated yet? A scoping review of the use of consumer-tar- distinguish depressive and inter-episode states in bipolar disorder. Clin geted wearable and mobile technology to measure and improve sleep. Psychol Sci. 2016;4(4):641–50. Sleep Med Rev. 2018;40:151–9. Godfrey H, Knight R. The validity of actometer and speech activity measures in Baune BT, Caliskan S, Todder D. Effects of adjunctive antidepressant therapy the assessment of depressed patients. Br J Psychiatry. 1984;145(2):159–63. with quetiapine on clinical outcome, quality of sleep and daytime motor Gooley J, Chua E. Diurnal regulation of lipid metabolism and applications of activity in patients with treatment-resistant depression. Hum Psychop- circadian lipidomics. J Genet Genomics. 2014;41:231–50. harmacol. 2007;22(1):1–9. Hellwig S, Domschke K. Update on PET imaging biomarkers in the diagnosis of Bei B, Wiley JF, Trinder J, Manber R. Beyond the mean: a systematic review on neuropsychiatric disorders. Curr Opin Neurol. 2019;32(4):539–47. the correlates of daily intraindividual variability of sleep/wake patterns. Heninger GR, Kirstein L. Effects of lithium carbonate on motor activity in mania Sleep Med Rev. 2016;2(8):108–24. and depression. J Nerv Ment Dis. 1977;164(3):168–75. Bellivier F, Geoffroy P, Etain B, Scott J. Sleep-and circadian rhythm–associated Henriksen T, Gronli J, Assmus J, Fasmer O, Schoeyen H, Leskauskaite I, et al. pathways as therapeutic targets in bipolar disorder. Expert Opin Ther Blue-blocking glasses as additive treatment for mania: effects on Targets. 2015;19:747–63. actigraphy-derived sleep parameters. J Sleep Res. 2020;2020:e12984. Benedetti F, Dallaspezia S, Fulgosi MC, Barbini B, Colombo C, Smeraldi E. Phase Hetrick S, Parker A, Callahan P, Purcell R. Evidence mapping: illustrating an advance is an actimetric correlate of antidepressant response to sleep emerging methodology to improve evidence-based practice in youth deprivation and light therapy in bipolar depression. Chronobiol Int. mental health. J Eval Clin Pract. 2010;16:1025–30. 2007;24(5):921–37. Hwang JY, Choi JW, Kang SG, Hwang SH, Kim SJ, Lee YJ. Comparison of the Bewernick BH, Urbach AS, Bröder A, Kayser S, Schlaepfer TE. Walking away from effects of Quetiapine XR and lithium monotherapy on actigraphy-meas- depression-motor activity increases ratings of mood and incentive drive ured circadian parameters in patients with bipolar ii depression. J Clin in patients with major depression. Psychiatry Res. 2017;247:68–72. Psychopharmacol. 2017;37(3):351–4. Brietzke E, Hawken ER, Idzikowski M, Pong J, Kennedy SH, Soares CN. Integrat- Insel T. The NIMH Research Domain Criteria (RDoC) Project: precision medicine ing digital phenotyping in clinical characterization of individuals with for psychiatry. Am J Psychiatry. 2014;171:395–7. mood disorders. Neurosci Biobehav Rev. 2019;104:223–30. Joffe RT, Uhde TW, Post RM, Minichiello MD. Motor activity in depressed Calogiuri G, Weydahl A, Carandente F. Methodological issues for studying the patients treated with carbamazepine. Biol Psychiatry. 1987;22(8):941–6. rest-activity cycle and sleep disturbances: a chronobiological approach Kaplan KA, McGlinchey EL, Soehner A, Gershon A, Talbot LS, Eidelman P, using actigraphy data. Biol Res Nurs. 2013;15(1):5–12. Gruber J, Harvey AG. Hypersomnia subtypes, sleep, and relapse in bipolar Colquhoun H, Levac D, O’Brien K, Straus S, Tricco A, Perrier L, et al. Scoping disorder. Psychol Med. 2015;45(8):1751–63. reviews: time for clarity in definition, methods, and reporting. J Clin Kaplan KA, Talavera DC, Harvey AG. Rise and shine: A treatment experiment Epidemiol. 2014;67(12):1291–4. testing a morning routine to decrease subjective sleep inertia in insom- Cuthbert BN. Translating intermediate phenotypes to psychopathology: the nia and bipolar disorder. Behav Res Ther. 2018;111:106–12. NIMH Research Domain Criteria. Psychophysiology. 2014;51(12):1205–6. Katz DL, Williams A, Girard C, Goodman J, Comerford B, Behrman A, Bracken M. Cuthbert B, Insel T. Toward the future of psychiatric diagnosis: the seven pillars The evidence base for complementary and alternative medicine: meth- of RDoC. BMC Medicine. 2013;11:126. ods of evidence mapping with application to CAM. Altern Ther Health De Crescenzo F, Economou A, Sharpley A, Gormez A, Quested D. Actigraphic Med. 2003;9(4):22–30. features of bipolar disorder: a systematic review and meta-analysis. Sleep Kim SJ, Lee YJ, Lee YJ, Cho SJ. Effect of quetiapine XR on depressive symptoms Med Rev. 2017;33:58–69. and sleep quality compared with lithium in patients with bipolar depres- Depner C, Cheng P, Devine J, Khosla S, de Zambotti M, Robillard R, Vakulin A, sion. J Affect Disord. 2014;157:33–40. Drummond S. Wearable technologies for developing sleep and circadian Klein E, Lavie P, Meiraz R, Sadeh A, Lenox RH. Increased motor activity and biomarkers: a summary of workshop discussions. Sleep. 2020;43(2):254. recurrent manic episodes: predictors of rapid relapse in remitted de Wild-Hartmann J, Wichers M, van Bemmel A, Derom C, Thiery E, Jacobs N, bipolar disorder patients after lithium discontinuation. Biol Psychiatry. van Os J, Simons C. Day-to-day associations between subjective sleep 1992;31(3):279–84. and affect in regard to future depression in a female population-based Klein E, Mairaz R, Pascal M, Hefez A, Lavie P. Discontinuation of lithium treat- sample. Br J Psychiatry. 2013;202:407–12. ment in remitted bipolar patients: relationship between clinical outcome Duncan W, Slonena E, Hejazi N, Brutsche N, Yu K, Park L, et al. Motor-activity and changes in sleep-wake cycles. J Nerv Ment Dis. 1991;179(8):499–501. markers of circadian timekeeping are related to ketamine’s rapid antide- Knapen S, Li P, Riemersma-van der Lek R, Verkooijen S, Boks M, Schoevers pressant properties. Biol Psychiat. 2017;82(5):361–9. R, Scheer F, Hu K. Fractal biomarker of activity in patients with bipolar Faurholt-Jepsen M, Brage S, Vinberg M, Christensen EM, Knorr U, Jensen HM, disorder. Psychol Med. 2020: 1, 1–8. Kessing LV. Differences in psychomotor activity in patients suffering from Krane-Gartiser K, Scott J, Nevoret C, Benard V, Benizri C, Brochard H, et al. unipolar and bipolar affective disorder in the remitted or mild/moderate Which actigraphic variables optimally characterize the sleep-wake depressive state. J Affect Disord. 2012;141(2–3):457–63. cycle of individuals with bipolar disorders? Acta Psychiatr Scand. Fernandez J, Hermida R. Inferential statistical method for analysis of non-sinu- 2019;139(3):269–79. soidal hybrid time series with unequidistant observations. Chronobiol Int. Kripke DF, Mullaney DJ, Atkinson M, Wolf S. Circadian rhythm disorders in 1998;15:191–204. manic-depressives. Biol Psychiat. 1978;13(3):335–51. Food and Drug Administration (FDA). 2019. Digital Health Software Precer- Kripke DF, Judd LL, Hubbard B, Janowsky DS, Huey LY. The effect of lithium tification (Pre-Cert) Program. https​://www.fda.gov/medic​al-devic​es/ carbonate on the circadian rhythm of sleep in normal human subjects. Biol Psychiatry. 1979;14(3):545–8. Scott et al. Int J Bipolar Disord (2020) 8:37 Page 22 of 22 Kupfer D, Weiss B, Foster G, Detre T, McPartland R. Psychomotor activity in Scott J, Hidalgo-Mazzei D, Strawbridge R, Young A, Resche-Rigon M, Etain affective states. Arch Gen Psychiatry. 1974;30(6):765–8. B, et al. Prospective cohort study of early biosignatures of response to MacDonald K, Krishnan A, Cervenka E, Hu G, Guadagno E, Trakadis Y. lithium in bipolar-I-disorders: overview of the H2020-funded R-LiNK initia- Biomarkers for major depressive and bipolar disorders using metabo- tive. Int J Bipolar Disord. 2019;7(1):20. lomics: a systematic review. Am J Med Genet B Neuropsychiatr Genet. Scott J, Murray G, Henry C, Morken G, Scott E, Angst J, Merikangas KR, Hickie 2019;180(2):122–37. IB. Activation in bipolar disorders: a systematic review. JAMA Psychiatry. Marino M, Li Y, Rueschman M. Measuring sleep: accuracy, sensitivity, and 2017;74(2):189–96. specificity of wrist actigraphy compared to polysomnography. Sleep. Scott J, Hennion V, Meyrel M, Bellivier F, Etain B. (2020) An ecological study of 2013;36:1747–55. objective rest-activity markers of lithium response in bipolar I disorder. McCarthy M, Wei H, Nievergelt C, Stautland A, Maihofer A, Welsh DK. Paper submitted. Chronotype and cellular circadian rhythms predict the clinical response Smagula SF. Opportunities for clinical applications of rest-activity rhythms to lithium maintenance treatment in patients with bipolar disorder. in detecting and preventing mood disorders. Curr Opin Psychiatry. Neuropsychopharmacology. 2019;44:620–8. 2016;29(6):389–96. Merikangas KR, Swendsen J, Hickie IB, Cui L, Shou H, Merikangas A, et al. Todder D, Caliskan S, Baune B. Night locomotor activity and quality of sleep Real-time mobile monitoring of the dynamic associations among motor in quetiapine-treated patients with depression. J Clin Psychopharmacol. activity, energy, mood, and sleep in adults with bipolar disorder. JAMA 2006;26(6):638–42. Psychiatry. 2019;76(2):190–8. Vallarino M, Henry C, Etain B, Gehue L, MacNeill C, Scott E, et al. An evidence Miake-Lye I, Hempel S, Shanman R. What is an evidence map? A systematic map of psychosocial interventions for the earliest stages of bipolar disor- review of published evidence maps and their definitions, methods, and der. Lancet Psychiatry. 2015;2(6):548–63. products. Syst Rev. 2016;5:28. van Someren E, Kessler A, Mirmiran M, Swaab D. Indirect bright light improves Moar J. Where Now for Wearables. 2018; https​://www.junip​erres​earch​.com/ circadian rest-activity rhythm disturbances in demented patients. Biol docum​ent-libra​ry/white​-paper​s/where​-now-for-weara​bles. Accessed 20 Psychiat. 1997;41:955–63. May 2020. Walker W, Walton J, DeVries A, Nelson R. Circadian rhythm disruption and Moon J, Cho C, Son G, Geum D, Chung S, Kim H, et al. Advanced circadian mental health. Transl Psychiatry. 2020;10(1):28. phase in mania and delayed circadian phase in mixed mania and depres- Wirz-Justice A. How to measure circadian rhythms in humans. Medicographia. sion returned to normal after treatment of bipolar disorder. EBioMedicine. 2007;29:84–90. 2016;11:285–95. Wehr T, Muscettola G, Goodwin F. Urinary 3-methoxy-4-hydroxyphenylglycol Moskowitz D, Young S. Ecological momentary assessment: what it is and why circadian rhythm. Early timing (phase-advance) in manic-depres- it is a method of the future in clinical psychopharmacology. J Psychiatry sives compared with normal subjects. Arch Gen Psychiatry. 1980 Neurosci. 2006;31(1):13–20. Mar;37(3):257–63. Mullington J. Developing biomarker arrays predicting sleep and circadian- Weiss B, Foster F, Reynolds C, Kupfer D. Psychomotor activity in mania. Arch coupled risks to health. Sleep. 2016;39(4):727–36. Gen Psychiatry. 1974;31(3):379–83. Natale V, Plazzi G, Martoni M. Actigraphy in the assessment of insomnia: a Whiteford HA, Degenhardt L, Rehm J, Baxter AJ, Ferrari AJ, Erskine HE, et al. quantitative approach. Sleep. 2009;32:767–71. Global burden of disease attributable to mental and substance use Nelson W, Tong Y, Lee J, Halberg F. Methods for cosinor rhythmometry. disorders: findings from the Global Burden of Disease Study 2010. Lancet. Chronobiologia. 1979;6:305–23. 2013;382:1575–86. Novak D, Albert F, Spaniel F. Analysis of actigraph parameters for relapse Wolff E, Putnam F, Post R. Motor activity and affective illness. The relationship prediction in bipolar disorder: a feasibility study. Conf Proc IEEE Eng Med of amplitude and temporal distribution to changes in affective state. Arch Biol Soc. 2014;2014:4972–5. Gen Psychiatry. 1985;42(3):288–94. Ng T, Chung K, Ho F, Yeung W, Yung K, Lam T. Sleep-wake disturbance in inter- Zee P, Badr M, Kushida C, Mullington J, Pack A, Parthasarathy S, et al. Strategic episode bipolar disorder and high-risk individuals: a systematic review opportunities in sleep and circadian research: report of the Joint Task and meta-analysis. Sleep Med Rev. 2015;20:46–58. Force of the Sleep Research Society and American Academy of Sleep Raoux N, Benoit O, Dantchev N. Circadian pattern of motor activity in major Medicine. Sleep. 2014;37(2):219–27. depressed patients undergoing antidepressant therapy: Relationship between actigraphic measures and clinical course. Psychiatry Res. 1994;52:85–98. Publisher’s Note Ritter PS, Marx C, Bauer M, Leopold K, Pfennig A. The role of disturbed sleep Springer Nature remains neutral with regard to jurisdictional claims in pub- in the early recognition of bipolar disorder: a systematic review. Bipolar lished maps and institutional affiliations. Disord. 2011;13(4):437. Sack R, Auckley D, Auger R, Carskadon M, Wright K, Vitiello M, Zhdanov I. Circadian rhythm sleep disorders: Part II, advanced sleep phase disorder, delayed sleep phase disorder, free-running disorder, and irregular sleep- wake rhythm. An American Academy of Sleep Medicine review. Sleep. 2007;30:1484–501. Salvatore P, Ghidini S, Zita G, De Panfilis C, Lambertino S, Maggini C, Baldessarini R. Circadian activity rhythm abnormalities in ill and recovered bipolar I disorder patients. Bipolar Disord. 2008;10:256–65. Schumann G, Binder E, Holte A, de Kloet E, Oedegaard K, Robbins T, et al. Stratified medicine for mental disorders. Eur Neuropsychopharmacol. 2014;24(1):5–50. Scott J. Clinical parameters of circadian rhythms in affective disorders. Eur Neuropsychopharmacol. 2011;21(suppl 4):S671–5. Scott J, Etain B, Bellivier F. Can an integrated science approach to precision medicine research improve lithium treatment in bipolar disorders? Front Psychiatry. 2018;9:360.