The network perspective to psychology conceptualizes observed variables (e.g., attitudes, symptoms, and moods) as causal agents in a complex interplay of psychological, biological, sociological and other components. Aiming to map out this interplay, re-cent psychometric developments have led to methodology for estimating network models from psychological datasets (network psychometrics). Cross-sectional data (one measurement per person) may be used to estimate non-dynamical networks, giving insight in how people differ, and time-series data (repeated measures of a person) may be used to estimate personalized dynamical networks, giving insight in within-person dynamics. These methods have since grown popular in diverse fields of research. This project proposes to use network psychometrics in an innovative adaptive measurement system, leading to solutions for diverse problems such as diagnostic tools (e.g., diagnosing which symptoms of the entire DSM a person endorses), patient monitoring (e.g., assessing changes in severity of select symptoms over time), online applications (e.g., voting recommendation apps), and the analysis and gathering of large epidemiological datasets. Data gathering in network psychometrics is currently static: first a dataset is collected, and then a statistical model is estimated. Participants are required to respond to many questions, possibly several times per day over a period of weeks. In network-based adaptive assessment, only informative items will be administered that best predict other responses (adaptive administration), and a network model will be updated as new data becomes available (adaptive model estimation). The proposal comprises three subprojects: (I) determining item selection rules given known network structures in non-dynamical (I-A) and dynamical (I-B) models, (II) handling of missing data in network estimation, and (III) updating time-varying dynamical networks as new data becomes available. Results will be implemented in freely available open-source software packages as well as applied in online applications and clinical practice.

What happens when multiple research teams analyze the same data? Surprisingly, their results often vary, sometimes dramatically. This occurs because researchers use different methodologies that can lead to conflicting conclusions. While the scientific community has begun to embrace the many-analysts approach—where multiple teams analyze the same question using the same data—there is currently no reliable way to combine the findings. In this project, a new statistical method will be developed to synthesize the results from many-analysts projects. This approach has the potential to make science more reliable and to improve the way researchers tackle scientific questions.

Response inhibition refers to the ability to stop an ongoing response, such as rapidly halting when the traffic light turns red. Response inhibition is the hallmark of executive functions and has received-and continues to receive-considerable attention in the field of experimental, clinical, and neuropsychology. The proposed project focuses on two recently developed cognitive process models of inhibition: the stop-signal race diffusion model (SS-RDM) and the stop-signal linear ballistic accumulator (SS-LBA). Both models conceptualize inhibition as a race between a set of evidence accumulators: one set that is associated with the ongoing response, and another that is associated with the stop response. The difference between the models lies in the mathematical formulation of evidence accumulation. Contrary to traditional models of response inhibition, process models provide parameter estimates that can be directly interpreted in terms of well-defined cognitive processes, such as the rate of evidence accumulation and response caution. Despite this conceptual advantage, the applicability of the SS-RDM and SS-LBA is limited by the large number of observations that are necessary for accurate parameter estimation and by the lack of adequate hypothesis testing techniques. The current project proposes to overcome these limitations with Bayesian inference. My first goal is to provide a Bayesian hierarchical implementation of the SS-RDM and SS-LBA that can substantially decrease the necessary number of observations. My second goal is to develop a Bayesian model selection method that allows researchers to formally evaluate nested and non-nested hypotheses in the SS-RDM and SS-LBA using reversible jump Markov chain Monte Carlo sampling. My overall objective is to create an integrated framework and corresponding software that will enable investigators to address fundamental and applied research questions about the nature and development of response inhibition using relatively small data sets and state-of-the-art Bayesian hypothesis testing techniques.

Network approaches to psychological phenomena, which originate in my work, have generated novel psychometric models in several fields, from psychopathology to attitude research and from intelligence to personality. In each of these fields, the conceptualization of psychometric constructs (e.g., depression) as sets of interacting components in a system (e.g., insomnia->fatigue->loss of interest) proves more fertile than classical models, in which these components are treated as passive psychometric indicators of a latent construct. In the past decade, my team has focused on pioneering statistical techniques that can estimate networks from data (i.e., “network psychometrics”). In this project, I want to supplement these data-oriented approaches with theoretically motivated modelling by developing a Network Construction Methodology (NCM). NCM is a modular system, in which users can generate network theories by combining independently functioning modules in a network structure. Because the networks so constructed inherit the dynamic behaviour and time scale from the underlying modules, they naturally combine into complex dynamical systems. NCM is developed in the context of psychopathology, a field which the network approach has already deeply penetrated. Three Ph. D. projects develop modules that correspond to: (1) belief traps, which are formed by feedback between beliefs and evidence, (2) fear loops, which are formed by feedback between fear and avoidance, and (3) regulatory loops, which represent the maintenance of biological and psychological homeostasis through behaviour. Two postdocs work on the systematization of existing network knowledge and integration of the modules into network structures, while a programmer is tasked with the development of a user interface that allows us to create a user platform through which different labs can work together. NCM will create a powerful methodological tool for theory generation, move network approaches to a new level of sophistication, and deeply change the way psychologists think about theory formation.

Ambition. Mental health is a substantial global challenge. The scale of mental illnesses is overwhelming: worldwide, one in four adults and one in ten children suffer from a mental disorder on any given moment. Mental disorders not only cause considerable suffering, they also cause huge economic waste. Alarming is the modest success rates of current interventions for mental disorders: only about 40% of patients achieve sustained recovery. To achieve better treatments, we need to know more about the key processes that are crucial to long-term clinical improvement. Excellent researchers from diverse disciplines join forces and propose to study an exciting new and challenging view on mental disorders. Our vision is that mental illnesses are not caused by a common pathogenic pathway. Instead, mental illnesses are dynamic and complex networks of symptoms that interact with one another over time, driven by an elaborate group of behavioural, cognitive, neurocognitive and interpersonal processes that are transdiagnostic in nature. So far, these dynamic symptom interactions and transdiagnostic processes have not been integrated within one comprehensive framework of mental illness. This research programme aims to do exactly that: it offers a novel paradigm to understand mental illnesses alongside genuinely new possibilities for more effective treatments and is, therefore, a roadmap to healthier societies. Consortium. The current consortium is unique in the world: it connects 19 influential scientists (47% female) who have demonstrated an ability to lead and inspire research teams and to innovate their scientific fields. The 14 senior members of the current consortium are internationally leading and have excellent track records. They rank among the global top in their fields. They are joined by twelve gifted young scientists (five in the consortium) in various stages of their careers, and all with the potential to act at the forefront of science and to become international leaders. The continuity of this innovative research program is further promised by the creation of a Young Talent Program for optimal development of the next generation leaders. Research Program. The ultimate goal of our research is to advance psychological treatments, so that more effective interventions will serve as an essential part of our set of approaches that are needed to make an impact upon the burden of mental disorders worldwide. Our research program is organised into a matrix of six research teams and three coherent layers of research: mapping, zooming and targeting. The research teams are organised along specific transdiagnostic processes and not along single disorders. We propose a new and challenging view on the origin, maintenance and change of all mental disorders. Our vision is that mental illnesses are not caused by a common pathogenic pathway. Instead, mental illnesses are dynamic and complex networks of symptoms that interact with one another over time. We will focus on (1) the estimation of complex and dynamic networks of individual patients, which can change over time, (2) disorder-transcending network-based interventions tailored for the individual, and (3) in-depth fundamental studies into transdiagnostic behavioural, cognitive, neurocognitive and interpersonal processes that can drive connectivity between symptoms. Organization and management. We will set up a lean and effective organisational structure with good administration and management, to facilitate pleasant, effective and inspiring cooperation and to pave the way for real scientific breakthroughs. It is not only our ambition but also our duty to bring the acquired knowledge on better treatments for mental illnesses truly into the practice of mental health care. We will frequently interact with stakeholders; clinicians, patients, carers, funders, commissioners, managers, policy planners, change experts, and the general public all have a part to play in innovating psychological therapies, and a focus on any one of the ideas presented by our research team has the potential to bring about substantial and much-needed improvements. The ‘translation and communication’ team is led by a professional communication officer to achieve optimal use of acquired knowledge. We are also strongly focused on the development and training of a next generation of scientists to ensure the continuity of our research at top level and in the international forefront. To achieve this ambition, we start a talent development program that focuses on the training, mentoring and monitoring of young researchers, from promising young talents to young international leaders. We also enable Talent Tenure Tracks for midcareer scientists. In addition, and this is certainly not unimportant, we will offer a safe environment to stimulate our young talents’ growth and development into the next generation of leading researchers in our field. Main applicant Jansen is the intended scientific director, she has extensive experience in management.