The breakthrough discoveries for thriving with bipolar disorder (BD2) integrated network longitudinal cohort protocol Authors: Emily G. Baxi, Megan Shanahan, Daniel L. Pham, Veronica C. Beck, Kelsey M. Barcomb, Eric J. Nestler, Mark A. Frye, Cara M. Altimus, Katherine E. Burdick Published: 2025 ABSTRACT Bipolar Disorder (BD) is a highly complex and heterogeneous disorder. As such, accurate diagnosis is often delayed, and effective treatment options are limited. The Integrated Network, a program of Breakthrough Discoveries for thriving with Bipolar Disorder (BD2), was designed as a platform for longitudinal deep phenotyping of a diverse group of people with bipolar disorder to define disease trajectories and to gain insight into the biological drivers of the illness. Biosamples, including whole blood, serum, plasma, and peripheral blood mononuclear cells (PBMCs), will be collected longitudinally and will also be made available. The Integrated Network is designed to be the largest and most comprehensive prospective longitudinal study conducted in bipolar disorder, allowing for the development of precision-based treatment strategies that optimize quality of life for people living with bipolar disorder. Introduction Bipolar Disorder (BD) is a chronic, recurrent psychiatric illness characterized by alternating episodes of mania or hypomania and depression (Vieta et al. 2018). It affects an estimated 1–3% of the global population and remains a leading cause of disability, with high rates of cognitive and functional impairment, medical comorbidities, and increased mortality. Despite decades of research, significant challenges persist in the diagnosis, treatment, and long-term management of BD, largely due to its marked clinical heterogeneity, complex etiology, and variable course. As a result, treatment remains largely trial-and-error, and many individuals do not receive care that is timely, personalized, or effective in the long term. On average, life expectancy for people living with BD is reduced by 10–20 years - largely due to cardiovascular comorbidities and death by suicide (Hoang et al. 2011; Hayes et al. 2015, 2017; John et al. 2018; Staudt Hansen et al. 2018). The successful development of diagnostic tools and therapeutics has, in part, been stalled by the complexity and heterogeneity of BD. While it is true that some BD patients function very well throughout their lives, others struggle with significant disability. The vast range of functioning is paralleled by a varied clinical course (e.g., differential recurrence rates, variability in treatment response) and substantial cognitive heterogeneity. There are recognized clinical subtypes (bipolar I disorder vs. bipolar II disorder, psychotic vs. non-psychotic) and course specifiers (i.e., rapid cycling) that reflect historical attempts to address diagnostic heterogeneity; however, these subtypes are neither based on neurobiological evidence nor do they provide a roadmap for treatment decisions or predicting outcomes. New diagnostic and therapeutic methods and comprehensive approaches are needed to identify the modifiable clinical and biological predictors of thriving with BD to provide the field with potential therapeutic mechanistic targets for future intervention and prevention efforts. Over the past two decades, a number of major research efforts have sought to characterize the biological, clinical, and functional diversity within BD and to identify predictors of outcomes. For example, the Systemic Treatment Enhancement Program for Bipolar Disorder (STEP-BD) followed over 4,000 individuals with BD to examine treatment effectiveness, clinical course, and predictors of relapse and recovery in real-world settings (Sachs et al. 2003). STEP-BD provided valuable insights into the importance of comorbidities, early symptom trajectories, and functional outcomes as predictors of long-term prognosis. Similarly, the Bipolar-Schizophrenia Network on Intermediate Phenotypes (B-SNIP) employed multimodal assessments – including neuroimaging, cognitive testing, and electrophysiology – across a transdiagnostic sample of individuals with BD and schizophrenia spectrum disorders (Tamminga et al. 2013 Nov). B-SNIP demonstrated the utility of intermediate phenotypes that occur across traditional diagnostic categories, highlighting the potential for biological subtyping in psychiatry (Koen et al. 2023; Parker et al. 2025). More recently, the Global Bipolar Collaborative, which pooled data from 13 cohorts across seven countries, has helped identify predictors of functional impairment and disease progression in BD, including the role of early adversity, cognitive deficits, comorbidities, and social determinants of health (Burdick et al. 2022). Other large-scale initiatives – such as the ENIGMA consortium (Schmaal et al. 2016; Hibar et al. 2018) and the Psychiatric Genomics Consortium (Stahl et al. 2019) – have advanced our understanding of BD by leveraging shared neuroimaging and genetic data across thousands of individuals. Advances in digital phenotyping and wearables technologies (Torous et al. 2017) are also beginning to offer new avenues for high-frequency, real-world measurement of mood and behavior. Nevertheless, few studies to date have combined deep multimodal phenotyping with a structure that enables real-time feedback between research findings and care delivery. Progress in the field of BD has been notably slower than in other areas of psychiatry and neuroscience. One major additional contributor has been a chronic lack of sustained research funding dedicated to BD. Compared to conditions like schizophrenia, major depressive disorder, or Alzheimer’s disease, bipolar disorder has historically received disproportionately fewer public and philanthropic resources, limiting the scope, scale, and continuity of research efforts (Torrey et al. 2021). This funding gap has hampered the development of integrated datasets, longitudinal cohorts, and translational infrastructure-key components needed to make meaningful strides in understanding and treating the disorder. Breakthrough Discoveries for thriving with Bipolar Disorder (BD2) was launched to address this gap by fostering a new research ecosystem capable of both advancing foundational science and improving care (Altimus et al. 2023). BD2 was initiated with a $150M commitment from three philanthropies. Its goal is to accelerate progress by supporting interconnected programs spanning early-stage biological research, centralized biobanking, genetics data platforms, and a longitudinal clinical study embedded within a learning health network (LHN). The initiative is anchored in the belief that transitional progress in BD requires both high-resolution scientific discovery and a real-world implementation infrastructure to bring findings to the point of care. The Integrated Network of BD2 (IN-BD) incorporates a longitudinal, observational cohort study designed to address key gaps in our understanding of BD and – ultimately - to inform the development of precision-based treatment approaches. Data collected through the study will provide a greater understanding of the clinical and functional heterogeneity, complexities introduced by common comorbid conditions, and temporal dynamics of the disorder. This will be achieved by incorporating core measures of clinical, cognitive, circuit-based, and biological processes that are believed to be relevant to bipolar disorder pathophysiology. Data will be collected from up to 4,000 people with a bipolar I diagnosis, each followed for at least five years, and will include clinical, imaging, wearable and app-based modalities. One of the unique design features of the IN-BD is that the longitudinal cohort research protocol is embedded within an LHN. The LHN is designed as a collaborative, multi-institutional framework that actively brings together clinicians, researchers, individuals with lived experience of BD, and data scientists across participating sites. This structure supports continuous learning and improvement by integrating standardized data from the longitudinal cohort research protocol – including clinical, biological, neuroimaging, cognitive, wearable, and patient-reported data – with real-world clinical information collected at the point of care and derived from electronic health records (EHRs) (Savitz et al. 2025). The LHN operates through structured governance and stakeholder engagement, applying principles of improvement science to systematically identify care gaps, prioritize targets for intervention, and implement and evaluate change strategies using methodologies. Through this iterative process, the LHN enables rapid testing and dissemination of evidence-informed practices across the IN-BD, ensuring that emerging insights are translated into clinical care in near-real time. Importantly, individuals with lived experience are central to all phases of the IN-BD’s development and improvement cycles, ensuring that care innovations reflect patient-centered goals and definitions of wellness. This unique integration of research and care infrastructure allows the IN-BD not only to generate high-quality, multidimensional data for scientific discovery but also to serve as a real-world testbed for precision medicine approaches in BD. By aligning research findings with clinical implementation, the IN-BD offers a scalable model for accelerating improvements in outcomes and quality of life for people living with BD. In line with BD2’s open science policy, the data, knowledge, and resources developed during the course of the program will be made available to qualified researchers and clinicians. Participant enrollment commenced in March 2024 and is ongoing. Integrated network study design The IN-BD was conceptualized and implemented with input from multiple stakeholders, including academic leaders, members of the Integrated Network Scientific Steering Committee (SSC) and people with lived experience. The planning stage began with the formation of an Executive Advisory Board (EAB) made of eight experts across multiple domains, such as bipolar disorder, learning health systems, longitudinal cohort studies, lived experience, and statistical methods. The EAB shaped the overall vision and goals of the program:To describe the heterogeneity and trajectory of bipolar I disorder in a large participant cohort using longitudinal measures of clinical, imaging, and physiological markers.To identify key clinical (e.g., recurrence rate, comorbid diagnoses, early life adversity, sleep disruption, cardiovascular risk), neural (e.g., structural changes in gray matter, white matter disease, cognitive deficits), and biological (e.g., inflammation, elevated stress hormones, genomics) processes that drive differential outcomes in bipolar disorder.To inform and implement data-driven improvements in care. To describe the heterogeneity and trajectory of bipolar I disorder in a large participant cohort using longitudinal measures of clinical, imaging, and physiological markers. To describe the heterogeneity and trajectory of bipolar I disorder in a large participant cohort using longitudinal measures of clinical, imaging, and physiological markers. To identify key clinical (e.g., recurrence rate, comorbid diagnoses, early life adversity, sleep disruption, cardiovascular risk), neural (e.g., structural changes in gray matter, white matter disease, cognitive deficits), and biological (e.g., inflammation, elevated stress hormones, genomics) processes that drive differential outcomes in bipolar disorder. To identify key clinical (e.g., recurrence rate, comorbid diagnoses, early life adversity, sleep disruption, cardiovascular risk), neural (e.g., structural changes in gray matter, white matter disease, cognitive deficits), and biological (e.g., inflammation, elevated stress hormones, genomics) processes that drive differential outcomes in bipolar disorder. To inform and implement data-driven improvements in care. To inform and implement data-driven improvements in care. Subcommittees of over 20 experts from different institutions across domains of clinical care for BD, biosample procurement and analysis, neuroimaging, and actigraphy were then convened to develop the first draft of the longitudinal cohort protocol. This Scientific Planning Committee prioritized numerous assessments and measures to be included in the centralized protocol. This document was used to inform the scientific requirements needed to recruit BD2 clinical sites. The longitudinal cohort study plans to enroll participants across 15 clinical sites. In-person visits occur at baseline and annually for at least four years. Annual visits are supplemented by more frequent remote tracking of mood and other related outcomes, as well as standard-of-care EHRs where available. A complete listing of all assessments and measures is provided in Table 1 and is overviewed by domain in Figure 1.Table 1Schedule of AssessmentsKey: Automatic Balloon Analog Risk Task (ABART)(Pleskac et al. 2008); Adverse Childhood Experience Questionnaire (ACEs)(Felitti et al. 1998); Alcohol Use Disorders Identification Test (AUDIT)(Allen et al. 1997 Jun);Barratt Impulsiveness Scale (BIS-11)(Patton JH et al.. 1995); Breakthrough Discoveries for thriving with Bipolar Disorder Bipolar Treatment Response and Switch Questionnaire (BD2 BDTR-SQ); Brief Assessment of Cognition in Schizophrenia (BACS): Symbol Coding(Keefe et al. 2004 Jun 1); Brief Psychotic Rating Scale (BPRS)(Raskin 1988); Brief Visuospatial Memory Test-Revised (BVMT-R)(Benedict et al. 1996); California Verbal Learning Test (CVLT)(Delis et al. 2017); Clinical Global Impression scale for Bipolar Disorder (CGI-BP)(Spearing et al. 1997); Cognitive Complaints in Bipolar Disorder Rating Assessment (COBRA)(Rosa et al. 2013 Aug 15); Columbia-Suicide Severity Rating Scale (C-SSRS)(Posner et al. 2011); Connor-Davidson Resilience Scale 10-item (CD-RISC)(Connor and Davidson 2003); Continuous Performance Test- Identical Pairs (CPT-I/P)(Cornblatt et al. 1988); Drug Use Screening Inventory (DUSI)(Tarter 1990); Everyday Discrimination Scale (EDS)(Williams et al. 1997); Functioning Assessment Short Test (FAST)(Rosa et al. 2007 Jun); Generalized Anxiety Disorder 7-item (GAD-7)(Spitzer et al. 2006); Letter Number Span (LNS)(Wechsler xxxx); Montgomery Asberg Depression Rating Scale (MADRS)(Montgomery and Asberg 1979); Morningness-Eveningness Questionnaire-Revised (MEQr)(Horne and Ostberg 1976); NEO Five-Factor Inventory-3 (NEO-FFF-3)(Costa and McCrae 1989); Patient Health Questionnaire-9 (PHQ-9)(Kroenke et al. 2001 Sep); Patient Mania Questionnaire (PMQ)(Cerimele et al. 2022); Penn Emotion Recognition Test (ERT)(Gur et al. 2010); Perceived Criticisms Scale (PCS)(Hooley and Teasdale 1989 Aug); Perceived Stress Scale (PSS)(Cohen et al. 1983 Dec); Pittsburgh Sleep Quality Index (PSQI)(Buysse et al. 1989); Premorbid Adjustment Scale (PAS)(Cannon-Spoor et al. 1982); Quick Inventory of Depressive Symptomatology-Clinician Version (QIDS-C)(Rush et al. 2003); Retrospective Assessment of the Lithium Response Phenotype Scale (ALDA)(Scott et al. 2017); Structured Clinical Interview for DSM-5 Disorders Research Version (SCID-5-RV)(First et al. xxxx); Wechsler Abbreviated Scale Intelligence 2nd Edition (WASI-II): Vocab and Reasoning(Wechsler 2011); Wechsler Memory Scale 3rd Edition (WMS-3): Spatial Span; Letter Number Span (LNS)(Wechsler xxxx); WHO Five Wellbeing Index (WHO-5)(Bech et al. 1996); Wide Range Achievement Test (WRAT)(Wilkinson and Robertson 2017): Reading; Young Mania Rating Scale (YMRS)(Young et al. 1978)Fig. 1Overview of longitudinal cohort study protocol. A screening visit is conducted to determine basic eligibility and interest followed by informed consent Table 1Schedule of AssessmentsKey: Automatic Balloon Analog Risk Task (ABART)(Pleskac et al. 2008); Adverse Childhood Experience Questionnaire (ACEs)(Felitti et al. 1998); Alcohol Use Disorders Identification Test (AUDIT)(Allen et al. 1997 Jun);Barratt Impulsiveness Scale (BIS-11)(Patton JH et al.. 1995); Breakthrough Discoveries for thriving with Bipolar Disorder Bipolar Treatment Response and Switch Questionnaire (BD2 BDTR-SQ); Brief Assessment of Cognition in Schizophrenia (BACS): Symbol Coding(Keefe et al. 2004 Jun 1); Brief Psychotic Rating Scale (BPRS)(Raskin 1988); Brief Visuospatial Memory Test-Revised (BVMT-R)(Benedict et al. 1996); California Verbal Learning Test (CVLT)(Delis et al. 2017); Clinical Global Impression scale for Bipolar Disorder (CGI-BP)(Spearing et al. 1997); Cognitive Complaints in Bipolar Disorder Rating Assessment (COBRA)(Rosa et al. 2013 Aug 15); Columbia-Suicide Severity Rating Scale (C-SSRS)(Posner et al. 2011); Connor-Davidson Resilience Scale 10-item (CD-RISC)(Connor and Davidson 2003); Continuous Performance Test- Identical Pairs (CPT-I/P)(Cornblatt et al. 1988); Drug Use Screening Inventory (DUSI)(Tarter 1990); Everyday Discrimination Scale (EDS)(Williams et al. 1997); Functioning Assessment Short Test (FAST)(Rosa et al. 2007 Jun); Generalized Anxiety Disorder 7-item (GAD-7)(Spitzer et al. 2006); Letter Number Span (LNS)(Wechsler xxxx); Montgomery Asberg Depression Rating Scale (MADRS)(Montgomery and Asberg 1979); Morningness-Eveningness Questionnaire-Revised (MEQr)(Horne and Ostberg 1976); NEO Five-Factor Inventory-3 (NEO-FFF-3)(Costa and McCrae 1989); Patient Health Questionnaire-9 (PHQ-9)(Kroenke et al. 2001 Sep); Patient Mania Questionnaire (PMQ)(Cerimele et al. 2022); Penn Emotion Recognition Test (ERT)(Gur et al. 2010); Perceived Criticisms Scale (PCS)(Hooley and Teasdale 1989 Aug); Perceived Stress Scale (PSS)(Cohen et al. 1983 Dec); Pittsburgh Sleep Quality Index (PSQI)(Buysse et al. 1989); Premorbid Adjustment Scale (PAS)(Cannon-Spoor et al. 1982); Quick Inventory of Depressive Symptomatology-Clinician Version (QIDS-C)(Rush et al. 2003); Retrospective Assessment of the Lithium Response Phenotype Scale (ALDA)(Scott et al. 2017); Structured Clinical Interview for DSM-5 Disorders Research Version (SCID-5-RV)(First et al. xxxx); Wechsler Abbreviated Scale Intelligence 2nd Edition (WASI-II): Vocab and Reasoning(Wechsler 2011); Wechsler Memory Scale 3rd Edition (WMS-3): Spatial Span; Letter Number Span (LNS)(Wechsler xxxx); WHO Five Wellbeing Index (WHO-5)(Bech et al. 1996); Wide Range Achievement Test (WRAT)(Wilkinson and Robertson 2017): Reading; Young Mania Rating Scale (YMRS)(Young et al. 1978) Schedule of Assessments Key: Automatic Balloon Analog Risk Task (ABART)(Pleskac et al. 2008); Adverse Childhood Experience Questionnaire (ACEs)(Felitti et al. 1998); Alcohol Use Disorders Identification Test (AUDIT)(Allen et al. 1997 Jun);Barratt Impulsiveness Scale (BIS-11)(Patton JH et al.. 1995); Breakthrough Discoveries for thriving with Bipolar Disorder Bipolar Treatment Response and Switch Questionnaire (BD2 BDTR-SQ); Brief Assessment of Cognition in Schizophrenia (BACS): Symbol Coding(Keefe et al. 2004 Jun 1); Brief Psychotic Rating Scale (BPRS)(Raskin 1988); Brief Visuospatial Memory Test-Revised (BVMT-R)(Benedict et al. 1996); California Verbal Learning Test (CVLT)(Delis et al. 2017); Clinical Global Impression scale for Bipolar Disorder (CGI-BP)(Spearing et al. 1997); Cognitive Complaints in Bipolar Disorder Rating Assessment (COBRA)(Rosa et al. 2013 Aug 15); Columbia-Suicide Severity Rating Scale (C-SSRS)(Posner et al. 2011); Connor-Davidson Resilience Scale 10-item (CD-RISC)(Connor and Davidson 2003); Continuous Performance Test- Identical Pairs (CPT-I/P)(Cornblatt et al. 1988); Drug Use Screening Inventory (DUSI)(Tarter 1990); Everyday Discrimination Scale (EDS)(Williams et al. 1997); Functioning Assessment Short Test (FAST)(Rosa et al. 2007 Jun); Generalized Anxiety Disorder 7-item (GAD-7)(Spitzer et al. 2006); Letter Number Span (LNS)(Wechsler xxxx); Montgomery Asberg Depression Rating Scale (MADRS)(Montgomery and Asberg 1979); Morningness-Eveningness Questionnaire-Revised (MEQr)(Horne and Ostberg 1976); NEO Five-Factor Inventory-3 (NEO-FFF-3)(Costa and McCrae 1989); Patient Health Questionnaire-9 (PHQ-9)(Kroenke et al. 2001 Sep); Patient Mania Questionnaire (PMQ)(Cerimele et al. 2022); Penn Emotion Recognition Test (ERT)(Gur et al. 2010); Perceived Criticisms Scale (PCS)(Hooley and Teasdale 1989 Aug); Perceived Stress Scale (PSS)(Cohen et al. 1983 Dec); Pittsburgh Sleep Quality Index (PSQI)(Buysse et al. 1989); Premorbid Adjustment Scale (PAS)(Cannon-Spoor et al. 1982); Quick Inventory of Depressive Symptomatology-Clinician Version (QIDS-C)(Rush et al. 2003); Retrospective Assessment of the Lithium Response Phenotype Scale (ALDA)(Scott et al. 2017); Structured Clinical Interview for DSM-5 Disorders Research Version (SCID-5-RV)(First et al. xxxx); Wechsler Abbreviated Scale Intelligence 2nd Edition (WASI-II): Vocab and Reasoning(Wechsler 2011); Wechsler Memory Scale 3rd Edition (WMS-3): Spatial Span; Letter Number Span (LNS)(Wechsler xxxx); WHO Five Wellbeing Index (WHO-5)(Bech et al. 1996); Wide Range Achievement Test (WRAT)(Wilkinson and Robertson 2017): Reading; Young Mania Rating Scale (YMRS)(Young et al. 1978) Overview of longitudinal cohort study protocol. A screening visit is conducted to determine basic eligibility and interest followed by informed consent Design Integrated network study implementation Participant selection criteria The IN-BD aims to produce insights that are relevant and useful to all those living with bipolar I disorder. Recruitment is focused on a diverse study cohort that is representative of the BD population, with differences in age, sex, genetic and socio-economic background, and disease severity, amongst other attributes. Inclusion criteria for the study are intentionally broad to ensure that BD is comprehensively understood. Participants must be between 18–75 years of age, have a diagnosis of bipolar I disorder (as per the Structured Clinical Interview for DSM-5 Disorders Research Version (SCID-5-RV)), and have mood and related symptoms that are stable enough to participate in the protocol. Exclusion criteria are limited to known neurodegenerative and dementing conditions and general medical conditions that are considered unstable based on clinician judgement. As the modal age of onset of bipolar disorder is 25 years old, 30% of the cohort will be 30 years or younger at the time of study entry to allow for a sufficient investigation into the early stages of the disease. Most studies report an equal sex ratio in bipolar I disorder prevalence (~1%); therefore, this study will enroll approximately an equal number of male and female participants. Bipolar I disorder prevalence does not differ significantly by race or ethnicity, so this study will enroll a racially and ethnically diverse cohort that is representative of the geographic distribution of the IN-BD sites (e.g., currently the USA and Canada). Note that the selection criteria do not exclude participants with comorbid psychiatric diagnoses or substance use. Recruitment Each site team establishes a recruitment strategy that is appropriate to their environment and site characteristics. Recruitment strategies include notifications sent through EHRs, health system patient and provider portals and messaging systems, institution websites, research databases/registries, outpatient clinical care settings, advertising, and community-based programs aimed towards the recruitment of participants. All material and messages sent or provided to potential participants are IRB-approved. As the research protocol is embedded within an LHN across all participating sites, preference is given to participants who receive their psychiatric and medical care within the same site/institution. The identification of potential participants at each site is followed by a standardized prescreening process prior to the informed consent, where no data are recorded but information is provided to the potential participant and general eligibility criteria are reviewed. Most frequently, the prescreening process is completed remotely via phone or video call but may be done in person at the local site, if preferred. Study assessments The structure of study assessments in IN-BD is domain-based with an effort to adopt standardized measures that are widely used in research and clinical care and to incorporate measures that provide a holistic view of health and wellness. Table 1 details the measures and the timeline for administration. In addition to the numerous standardized measures described, the IN-BD protocol includes the careful assessment of treatment utilization, efficacy and reported side effects. The neurocognitive battery is based upon the International Society for Bipolar Disorders–Battery for Assessment of Neurocognition (ISBD-BANC) (Yatham et al. 2010), and the neuroimaging protocol is based upon the ABCD design and focuses on structural measures of grey and white matter alongside resting-state fMRI. In between annual study visits, participants will wear a Fitbit device to passively capture activity and sleep measures on a continuous basis. App-based measurements of mood and wellness allow for higher cadence data collection on course trajectory. A qualitative interview is administered at the baseline visit to gather data on experiences with symptoms, treatments, and resilience that are difficult to ascertain in standardized research assessments (e.g., experience of mixed mood states, positive/negative aspects of living with and treatment of BD, personal resilience factors). Study bioassays The planned biosample assays (Table 2) are designed to support both clinical insight and discovery research. This collection of study assessments was iterated on and further shaped by our IN-BD SSC and several workgroups led by participating IN-BD site teams.Table 2Biosample assays included in the IN-BD longitudinal cohort protocolBlood samples are collected at baseline and annually during in-person visits. The assays encompass measures of general wellness (e.g., lipid profile, thyroid function, sex hormones), metabolic and inflammatory status (e.g., CRP, cytokines), and neuroendocrine function (e.g., cortisol levels), providing a comprehensive view of biological systems relevant to BD. In addition, blood and peripheral blood mononuclear cells (PBMCs) are stored to enable future molecular profiling. Standardized collection and centralized processing protocols ensure sample quality and enable harmonized analyses across sites, supporting the BD2 initiative’s goals of advancing precision medicine and biomarker discovery in BD Table 2Biosample assays included in the IN-BD longitudinal cohort protocolBlood samples are collected at baseline and annually during in-person visits. The assays encompass measures of general wellness (e.g., lipid profile, thyroid function, sex hormones), metabolic and inflammatory status (e.g., CRP, cytokines), and neuroendocrine function (e.g., cortisol levels), providing a comprehensive view of biological systems relevant to BD. In addition, blood and peripheral blood mononuclear cells (PBMCs) are stored to enable future molecular profiling. Standardized collection and centralized processing protocols ensure sample quality and enable harmonized analyses across sites, supporting the BD2 initiative’s goals of advancing precision medicine and biomarker discovery in BD Biosample assays included in the IN-BD longitudinal cohort protocol Blood samples are collected at baseline and annually during in-person visits. The assays encompass measures of general wellness (e.g., lipid profile, thyroid function, sex hormones), metabolic and inflammatory status (e.g., CRP, cytokines), and neuroendocrine function (e.g., cortisol levels), providing a comprehensive view of biological systems relevant to BD. In addition, blood and peripheral blood mononuclear cells (PBMCs) are stored to enable future molecular profiling. Standardized collection and centralized processing protocols ensure sample quality and enable harmonized analyses across sites, supporting the BD2 initiative’s goals of advancing precision medicine and biomarker discovery in BD Clinical assays are performed to assess general medical health and provide actionable data on key physiological domains relevant to BD. These include measurements of thyroid function (TSH, free T4 and T3), lipid profiles (HDL, LDL, total cholesterol, triglycerides), sex hormones (e.g., estradiol, testosterone), inflammatory markers (e.g., C-reactive protein), and metabolic indicators (e.g., glucose, hemoglobin A1c). These markers may help identify comorbidities and physiological states that influence disease trajectory or treatment response. Clinicians within IN-BD receive reports on clinical laboratories, which can be used in treatment planning to optimize overall health and well-being for IN-BD participants. In addition to the clinical assays, research-grade assays, including whole genome sequencing, support current deep phenotyping efforts as well as future hypothesis-driven analyses and biomarker discovery. Statistical analyses Given the complexity and multimodal nature of the data collected in IN-BD, dedicated analytic workgroups have been established for each core data modality – including clinical, cognitive, neuroimaging, biosample, qualitative, wearable and digital health domains. These workgroups, which are composed of domain experts across IN-BD clinical sites, are responsible for selecting, standardizing, and recommending statistical methods most appropriate for their respective data types. Each workgroup will develop analytic plans in alignment with the scientific objectives of the study and with the aim of harmonizing methodologies across sites. A formal power analysis will be conducted following the first major data freeze (end of October 2025) and internal release, planned for February 2026. These analyses will incorporate data from across all IN-BD modalities and key variable distributions to inform the study’s statistical power for detecting clinically meaningful effects and associations across domains. Importantly, the initial sample size target of 4,000 and a five-year duration of follow-up reflect budgetary realities, balancing the scientific ambition of the program with the funding available through the BD2 initiative. If the forthcoming power analyses suggest that different enrollment targets or follow-up durations are needed to achieve scientific goals, BD2 will make every effort to support these revised targets and the additional insights they may yield. Together, the scale and timeline are intended to support both the hypothesis-driven and exploratory analyses that can inform precision medicine approaches to care in BD. Building the integrated network infrastructure Site selection The Integrated Network was designed to bring up to 15 clinic sites together to deeply phenotype a large cohort of people living with BD. Site selection is based on a competitive application process in response to requests for application (RFAs; available at www.bipolardiscoveries.org/funding/rfa-opportunities/). The review of applications is conducted by a selection committee, consisting of members of the IN-BD SSC in conjunction with BD2 leadership. Prior to the release of the RFAs, the selection committee established the following basic criteria for site evaluation and selection:i)BD expertiseii)Enrollment capacity and diversityiii)Availability of resources and infrastructure (e.g., 3 T MRI scanner) required to complete the protocoliv)Team leadership, expertise and experience with multi-center, longitudinal studiesv)Institutional support for multi-center research and open sciencevi)Willingness to engage in a learning health network and implement data-driven changes in care practice. i)BD expertise BD expertise ii)Enrollment capacity and diversity Enrollment capacity and diversity iii)Availability of resources and infrastructure (e.g., 3 T MRI scanner) required to complete the protocol Availability of resources and infrastructure (e.g., 3 T MRI scanner) required to complete the protocol iv)Team leadership, expertise and experience with multi-center, longitudinal studies Team leadership, expertise and experience with multi-center, longitudinal studies v)Institutional support for multi-center research and open science Institutional support for multi-center research and open science vi)Willingness to engage in a learning health network and implement data-driven changes in care practice. Willingness to engage in a learning health network and implement data-driven changes in care practice. Sites are being added in three waves, each set approximately two years apart (Table 3). The first wave of site recruitment (“Wave 1”) was initiated in October 2022, with the selection of six founding clinic sites. “Wave 2” was initiated in July 2024, with the addition of five sites early in 2025. A third wave of site expansion is expected in early 2026.Table 3IN-BD site selection includes three waves of recruitmentThe phased expansion of the IN-BD is planned to occur in three waves. Site selection prioritized clinical and research expertise in BD, enrollment capacity, and readiness to participate in an LHN. The approach supports diverse representation and ensures high-quality, standardized data collection across all sites Table 3IN-BD site selection includes three waves of recruitmentThe phased expansion of the IN-BD is planned to occur in three waves. Site selection prioritized clinical and research expertise in BD, enrollment capacity, and readiness to participate in an LHN. The approach supports diverse representation and ensures high-quality, standardized data collection across all sites IN-BD site selection includes three waves of recruitment The phased expansion of the IN-BD is planned to occur in three waves. Site selection prioritized clinical and research expertise in BD, enrollment capacity, and readiness to participate in an LHN. The approach supports diverse representation and ensures high-quality, standardized data collection across all sites As of October 2025, Wave 1 and Wave 2 sites have collectively enrolled 930 participants, more than 600 of whom have completed the in-person Baseline 2 visit. Centralized study support and infrastructure The collection of standardized, reliable data was deemed paramount to the program’s success. To that end, the IN-BD was designed to incorporate three key centralized support services: a Clinical Coordinating Center (CCC), Data Coordinating Center (DCC) and Biorepository. The clinical coordinating center (CCC) The CCC, led by the Feinstein Institutes for Medical Research, performs several key functions, including: pre-screening to establish participant eligibility, centralized rating to support clinical diagnostic and symptom severity assessments for all sites, coordination of all central IRB activities, and ongoing monitoring of study activity and progress. To limit site-to-site variability, assessments are administered by a team of highly trained, centralized raters as described below. Following consent, each participant is administered the research version of the Structured Clinical Interview-DSM-5 (SCID-5-RV) (First et al. xxxx) to confirm and document primary and comorbid Axis I diagnoses – capturing both current and lifetime diagnoses. A brief version of the SCID interview is conducted during each subsequent annual visit to document any changes to diagnoses and capture detailed information on illness-associated events (e.g., new episodes, hospitalizations) that have occurred in the prior 12-month interval. Early life social and academic development across childhood, adolescence, and early adulthood is captured retrospectively using the Premorbid Adjustment Scale (PAS) (Cannon-Spoor et al. 1982) interview at baseline as well. In addition, current symptom severity is assessed annually utilizing the Young Mania Rating Scale (manic symptoms; YMRS) (Young et al. 1978), the Montgomery Asberg Depression Rating Scale and Clinician-rated Quick Inventory of Depressive Symptomatology [depressive symptoms; MADRS/QIDS-C) (Montgomery and Asberg 1979; Rush et al. 2003)], and the Brief Psychotic Rating Scale (psychotic symptoms; BPRS) (Raskin 1988). Suicidality is assessed utilizing the Columbia-Suicide Severity Rating Scale (C-SSRS) (Posner et al. 2011). Overall BD symptom severity is captured through the Clinical Global Impressions Scale-Bipolar Version (CGI-BP) (Spearing et al. 1997). Centralized raters are trained on protocols to evaluate and appropriately address safety concerns, including communication with the local site. The data coordinating center (DCC) The IN-BD is designed to collect standardized data across multiple sites and modalities. To meet these goals, the DCC, led by Indoc Research, is charged with aggregating, harmonizing, and storing all study data. The DCC-operated BD2 Data Platform (www.BD2data.indocpilot.io) houses clinical data (including data derived from EHRs), biosample data, neuroimaging data, wearable data and mobile health data collected from study participants. Sites send EHR extracts, processed wearable data, and neuroimaging data directly to the BD2 Data Platform. DCC data scientists work with site data teams to ensure that incoming data are correctly standardized, packaged, and meet the initiative’s criteria. The biosample processing core and biorepository The functions of the Biosample Processing and Biorepository Core, led by the Mayo Clinic BioPharma Diagnostics and Bioservices teams, are critical components of the IN-BD, ensuring that high-quality biological samples are collected, processed, and stored for current and future research. Blood samples are collected at both baseline and at each annual follow-up visit, and they are processed centrally to ensure standardization and high sample quality across all sites. Following shipping, centralized processing at the Mayo Clinic ensures that all biosamples are handled uniformly and variability is limited. Biospecimens include whole blood, serum, plasma, and peripheral blood mononuclear cells (PBMCs). The latter will enable future immunological profiling and the creation of lines of induced pluripotent stem cells. Combined with the in-depth phenotyping data generated by the longitudinal protocol, the biorepository will enable in-depth characterization of the biological underpinnings of BD and support the development of future precision medicine approaches with the BD2 framework. Governance The BD2 Integrated Network is guided by a multi-tiered governance structure that ensures scientific integrity, operational consistency, and cross-site collaboration. Program leadership is provided by two scientific co-chairs with complementary expertise in clinical care and research of BD, who work closely with the program team to coordinate strategy and ensure alignment with BD2’s mission. The program team manages day-to-day operations and facilitates implementation of the study protocol, infrastructure, and communications across participating sites. Oversight of the scientific direction and data is provided by the IN-BD Scientific Steering Committee (SSC), a multidisciplinary body composed of investigators, domain experts, individuals with lived experience and data science leaders. The SSC provides ongoing guidance on protocol refinement, data sharing policies, analytical priorities and publication governance. Domain-specific workgroups and subcommittees - focused on areas such as neuroimaging, biospecimens, women’s health and implementation of the LHN - support the SSC by offering targeted scientific and technical input. In addition, a Data Access Committee reviews requests for data access to ensure compliance with open science principles, privacy protections, and responsible stewardship. This coordinated structure enables BD2 to operate a complex, multi-site initiative with transparency, rigor, and responsiveness to both scientific opportunity and community needs. Discussion The BD2 Integrated Network Longitudinal Cohort Protocol represents a significant advancement in the study of BD, addressing several longstanding challenges in the field. The protocol is designed to generate a deep and comprehensive understanding of BD through the longitudinal study of up to 4,000 BD I individuals, focusing on clinical, biological, and neuroimaging data. Addressing heterogeneity in bipolar disorder One of the central challenges in BD research is the disorder’s marked heterogeneity, which has impeded the understanding of the pathophysiology of the disease and the development of more effective treatments. The Integrated Network aims to address this by capturing the full spectrum of BD manifestations through extensive phenotyping across a diverse cohort. The study’s inclusion criteria are intentionally broad, allowing for the enrollment of individuals with varying disease severity, demographic backgrounds, and comorbidities. This diversity is crucial for identifying the key clinical and biological processes that influence BD outcomes. By incorporating a wide range of data types, the study is well-positioned to explore the multifaceted nature of BD, define its trajectories, and uncover potential subtypes and/or biomarkers that could inform more personalized treatment strategies. However, while the cohort is well-suited to investigating mid- and long-term illness dynamics, a relative limitation of the current design is its exclusion of individuals at clinical risk for BD or those in the early prodromal phase. The study does not include pediatric participants or those without a confirmed diagnosis of bipolar I disorder, and therefore is not specifically designed to prospectively address questions related to illness onset, early detection, or primary prevention. These remain critical areas for future investigation and complementary study design. Nonetheless, within its defined scope, the BD2 Integrated Network represents a robust and scalable platform to advance our understanding of how people live-and can thrive-with BD across time. Integration of a learning health network A unique aspect of the BD2 Integrated Network is its incorporation of an LHN. This approach facilitates a continuous feedback loop between research findings and clinical practice, enabling the rapid translation of new insights into improved patient care. The LHN involves collaboration between clinicians, researchers, and individuals with lived experience of BD, ensuring that the study’s findings are directly relevant to clinical practice. This integration of research and care has the potential to significantly enhance the management of BD by providing clinicians with data-driven tools to tailor treatment to individual patients. Moreover, the LHN model promotes ongoing learning and adaptation, which is essential in a field as complex and evolving as BD. Methodological strengths, data sharing and open science commitments The study’s design reflects input from a wide range of stakeholders, including academic investigators, clinicians, and people with lived experience, and places a strong emphasis on methodological rigor and standardization. The use of centralized raters for diagnostic and clinical assessments will minimize site-to-site variability - a common issue in multi-center studies. The DCC plays a pivotal role in ensuring data consistency and quality across sites. By aggregating, harmonizing, and storing data in a centralized platform, the DCC enables efficient data sharing and analysis, which is critical for the study’s success. The first public data release from IN-BD is anticipated by mid-2026, following the initial data freeze and quality assurance processes. This release will be announced on the BD2 website (www.bipolardiscoveries.org) and will include curated, de-identified datasets from multiple modalities, including clinical, neuroimaging, wearable and digital data. In alignment with BD2’s open science policy, all qualified researchers will be encouraged to access and utilize these data to pursue novel scientific questions, generate hypotheses, and validate findings across domains. BD2 is committed to fostering a collaborative research ecosystem in which data, tools, and insights are openly shared to accelerate progress and reduce redundancy. Following the inaugural release, annual data updates are expected, with expanded participant numbers, longitudinal follow-up and additional data types made available through a secure access process. By sharing high-quality, harmonized data broadly with the scientific community, BD2 aims to multiply the impact of this investment and catalyze discoveries that improve outcomes for people living with BD. Implications for precision medicine in bipolar disorder The IN-BD has the potential to significantly advance the field of precision medicine in BD. By identifying the clinical, neural and biological factors that drive individual differences in BD outcomes, the study could pave the way for more personalized treatment approaches. Precision medicine, which involves tailoring treatment to the specific characteristics of each patient, is particularly relevant in BD due to the disorder’s heterogeneity. The longitudinal nature of the study allows for the examination of how these factors change over time and influence the course of the disease, providing valuable insights into the mechanisms underlying BD and informing the development of targeted interventions. Broader impact and future directions The IN-BD is a vital part of a broader vision within BD2, serving as the clinical arm of the initiative. Other components, including the Genomics Platform, the Omics Platform, and Discovery Research Grants, focus on the examination of causal biological mechanisms underlying the illness and discovery markers of treatment response. The clinical IN-BD and the translational genetics, omics, and discovery platforms will bidirectionally influence and inform each other with the goal of holistically advancing the foundational understanding and care for people with bipolar disorder. The IN-BD allows for near-real-time implementation of findings from translational studies, making it a novel and essential aspect of the program that impacts people with BD in a short timeframe. The IN-BD Longitudinal Cohort Protocol represents a significant step forward in BD research, with the potential to address many of the limitations that have hindered progress in the field. The study’s comprehensive approach, which combines deep phenotyping, longitudinal data collection, and a Learning Health Network, offers a model for future research in other complex and heterogeneous disorders. As the study progresses and data become available, it will be important to continue to evaluate the impact of the LHN on clinical practice and patient outcomes. Additionally, the success of the IN-BD could inspire similar initiatives in other areas of mental health research, further advancing the field of precision psychiatry. The BD2 Integrated Network Longitudinal Cohort Protocol is an innovative effort that has the potential to transform our understanding of BD and improve outcomes for individuals living with this challenging condition. Through its novel design, emphasis on standardization, and integration of research and clinical practice, the study is poised to make significant contributions to the field and serve as a model for future research endeavors.