Target Population

The BMNC dataset is a CITE-seq dataset from Stuart et al. (2019) (link), consisting of 30,672 cells with a panel of 25 antibodies. We use the expression matrices as quantified in the original experiment. For gene expression, in order to facilitate comparisons with other methods, we use standard log-normalization with default parameters in Seurat. We apply a CLR transformation to normalize protein data within each cell. We use PCA to reduce the dimensionality of both datasets, taking 30 RNA and 18 protein dimensions to construct the WNN graph. When performing a targeted re-clustering of T cell populations (Figure S2I), we repeated all preprocessing steps and performed the same procedure on 14,901 cells identified as T cells.

A series of sequential processes were employed in an effort to derive consensus definitions for key terms in sedentary behavior research (Fig. 1). These processes included i) a literature search, ii) establishment of a Steering Committee of SBRN members, iii) invitation to all SBRN members to participate in the consensus project, iv) selecting a list of key terms, v) developing a conceptual model, vi) drafting definitions for key terms, vii) collecting input and feedback on the conceptual model and draft definitions from participating SBRN members, viii) compiling input and finalizing (reaching consensus on) the conceptual model and definitions, ix) preparing the manuscript with sign-off by all participants, x) and finally disseminating project results through publication and presentation.Fig. 1

Terminology Consensus Project timeline

A literature search was performed in March 2016 to i) review the current use of the SBRN [25 (link)] definition, and potential deviations from this definition, and ii) examine current operational definitions of sedentary behavior and related terms (e.g., screen time, sedentary behavior patterns, bouts and breaks) and any evidence of inconsistencies, differences, conflicts, or concerns over variations in definitions employed. To identify current relevant articles a search with filters capturing papers published in the past 5 years (May 2011 to May 2016) was conducted in PubMed (see Table 3 for search terms). The articles were selected if there was mention of sedentary behavior and/or related term definitions, cut-points, measurement challenges/recommendations or standard processing techniques for accelerometer data in the title or abstract, with no regard to the population (e.g., child or adult; healthy or unhealthy), type of study, intervention, or comparison being explored. After reviewing the included full-text articles, data were extracted from relevant articles including the aim of the study, defined/discussed terms, targeted population, and definitions/relevant information.Table 3

Literature search strategy

((definition*[Title/Abstract] OR consensus[Title/Abstract] OR standard*[Title/Abstract]) AND

(sedentary[Title/Abstract] OR sitting[Title/Abstract] OR reclining[Title/Abstract] OR

stationary[Title/Abstract])) AND (time[Title/Abstract] OR duration*[Title/Abstract] OR

bout*[Title/Abstract] OR pattern*[Title/Abstract] OR interruption*[Title/Abstract] OR

break*[Title/Abstract] OR type*[Title/Abstract] OR characteristic*[Title/Abstract] OR intermittence*[Title/Abstract])

In addition to gathering background information for the project, the literature search allowed for the identification of authors of key papers who were invited to form a Steering Committee for the SBRN Terminology Consensus Project (MST (Chair), TJS, VC, AEL, SFMC, TA, MJMC and project management support from SA and JDB from SBRN). Key terms from the literature search were collated and the Steering Committee members were asked to add or remove key terms from the list. A final draft list of terms was agreed upon by this Committee. An email was sent to the SBRN membership, consisting of researchers, scholars, practitioners, trainees and students interested in sedentary behavior (1094 members worldwide in April, 2016), soliciting interest in participating in the project and asking for suggestions for key terms to be included in the survey.
The Steering Committee identified the most common key terms reported and deliberated through a short survey and email communication to arrive at draft definitions for each term, important caveats for certain age and ability groups, examples to assist with interpretation and, when available, references for the proposed definition. The final list of terms included stationary behavior, sedentary behavior, standing, screen time, non-screen-based sedentary time, sitting, reclining, lying, and sedentary behavior pattern. This process also led to the development of a conceptual model to help position the key terms in relation to one another. The draft definitions, caveats, examples, references, and conceptual model were included in a survey developed for distribution to participating SBRN members. The survey asked participants to assess the clarity of, and agreement with, the conceptual model and proposed definitions (using a five point scale from strongly agree to strongly disagree), while also providing an opportunity for general comments. Consensus was determined a priori to have been achieved if ≥75% of respondents strongly agreed or somewhat agreed with a particular question (see Additional file 1: for complete survey). Note that the term “physical inactivity” was not included in the survey because there were no suggested changes to the existing SBRN [25 (link)] definition.
Input from all participants, including the Steering Committee, was consolidated (by MST, SA, JDB) and revisions to the conceptual model, definitions, caveats, examples, and references were made by the Steering Committee. The draft manuscript, including the revised conceptual model and definitions, was sent to all participants for additional review and comments. After additional revisions, a revised draft of the manuscript was resent to all participants for comments, group consensus, and assessment of likelihood of use. Finally, the further revised manuscript (third review) was recirculated to the project participant group for final comments and sign-off for submission.

This study was conducted in Panji, Kota Bharu district, Kelantan, Malaysia (Fig. 3), located at the east cost of Peninsular Malaysia and has the highest population among the 15 sub-districts of Kota Bharu, the capital state of Kelantan. A total of 338 respondents were recruited in this study. The population of interest in this study involved residents in Kota Bharu district and considered only residents who have attained 18 years old and above. Sample unit is residents living in Kota Bharu district of more than a year and aged more than 18 years. The target population comprised all the households in Kota Bharu District (491,237); however, it is impossible to conduct a study with such a large number within a limited time period and inadequate financial budget. Therefore, a multi- stage random sampling technique was used in selecting the appropriate sample in order to evaluate the objectives of this study and to ensure that households in the districts had the same possibility of being included in the study (Dlamini et al., 2017). Initially, one district of Kelantan state (Kota Bharu) was selected out of 10 total districts. In the second stage, one sub-district of Kota Bharu District (Panji) was selected out of 15 total sub-districts. Eventually, 338 households were randomly selected as sample size. Convenient sampling was also used to select respondents due to time constraint and response obtained from target population. The localities involved were Kampung Tapang, Kampung Chempaka, Kampung Belukar, Kampung Panji, Taman Sri Iman, Taman Desa Kujid and Taman Bendahara.Fig. 3

Location of the study area in Panji, Kota Bharu district, Kelantan, Malaysia (Source:ArcGis Software version 10.2; source of shape file: Department of Drainage and Irrigation, obtained with consent)

Patients between 18 and 65 years of age referred to a treatment package for single-episode depression will be recruited (Table 1) with minimal exclusion criteria to recruit representative adult outpatients who would typically receive treatment in routine practice (Table 1), of which the majority are women (71%) and aged 18–35 (68%) (S. Figure 1). Patients over 65 (approximately 0.7% of the target population) are excluded because of potential age-related cognitive decline, concomitant medical conditions, or medications that could interact with assessments or treatment (S. Figure 1). Allocation into the subcohorts is based on eligibility, e.g., MRI compatibility, scheduling, and patient willingness to participate.Table 1

Inclusion and exclusion criteria for patients

Patient inclusion criteria:
• Fulfilment of ICD-10 diagnostic criteria for a primary depressive episode (i.e., not secondary to known organic or other psychiatric disorder) • Referral to a treatment package for single-episode depression • Age between 18 and 65 years
Exclusion criteria:
• Psychosis or psychotic symptoms • History of severe head trauma involving hospitalization or unconsciousness for more than 5 min • Known, substantial structural brain abnormalities • Insufficient Danish language skills to complete questionnaires and cognitive testing
Additional exclusion criteria for subcohort I:
• Severe somatic disease • Contraindications for MRI (e.g., metal implants, claustrophobia, or back problems)
Additional exclusion criteria for subcohort I:
• Use of psychotropic drugs • Exposure to radioactivity > 10 mSv within the last year • Pregnancy or breastfeeding

The primary depressive episode, consistent with the International Statistical Classification of Diseases and Related Health Problems version 10 (ICD-10) criteria for MDD without psychotic features (F32.1, F32.2, F32.8 and F32.9), is confirmed by a specialist in psychiatry at the central diagnostic and referral centre.

We used data from the China Health and Retirement Longitudinal Study (CHARLS) in wave 4 of 2018 for cross-sectional analysis. CHARLS aims to collect high-quality data on households and individuals aged 45 and older in China to analyze aging and promote research on healthy aging. CHARLS surveyed participants for basic information, health status, health insurance and health care, and retirement. CHARLS was approved by the Ethics Review Committee of Peking University, and all participants signed an informed consent form before the investigation and voluntarily participated in the survey [35 ].
The target population selected was people aged 45 and above. The missing rate of the dependent variable values was 18.18%, and 15,636 participants were included in the study after excluding those with missing key variable values. The total missing rate for the remaining values of all variables was 1.33%, so we directly excluded participants with missing variable values, and the final number of samples included in the study was 15,428. Data inclusion process are shown in Fig. 1.Fig. 1

Data inclusion process

A process evaluation employing qualitative and quantitative methods prospectively tracked the implementation to describe how the social innovation was initiated, carried out and how participants responded to the innovation. The process evaluation was performed according to the UK Medical Research Council guidelines: implementation, mechanism of impact and context [41 (link)]. We developed a logic model, underpinning the assumptions on which the intervention was thought to function (see Additional file 1). For each of the three components, key questions needing an answer to understand the process were formulated, followed by identifying the target population, data sources, procedures, and tools (see Additional file 1). The implementation of the innovation: What was delivered and how it was delivered, including the procedures used to approach and attract facilitators, mentors, and group stakeholders (recruitment), the participation (reach), and the efforts of the facilitators (dose). Mechanism of impact: The participants’ responses to and interactions with the innovation. In this component, we explored why specific reactions to social innovation resulted in particular outcomes. Furthermore, we also explored the problems the groups addressed, the type and relevance of prioritised issues, actions taken, the interaction between group and facilitator and methods used. Context: What contextual aspects that influenced the innovation, the implementation and the mechanism of impact, leading to different outcomes. The outcomes of the social innovation included the relevance of identified problems and completion of PDSA cycles, knowledge of perinatal care, perspectives of gaining knowledge and performance of antenatal care. The following data collection modes and tools were used to monitor data of the three process evaluation components and the outcomes of the social innovation: