Methodological and Statistical Practices of Using Symptom Networks to Evaluate Mental Health Interventions: A Review and Reflections.

The network approach to psychopathology, which assesses associations between individual symptoms, has recently been applied to evaluate treatments for mental disorders. While various options for conducting network analyses in intervention research exist, an overview and an evaluation of the various approaches are currently missing. Therefore, we conducted a review on network analyses in intervention research. Studies were included if they constructed a symptom network, analyzed data that were collected before, during or after treatment of a mental disorder, and yielded information about the treatment effect. The 56 included studies were reviewed regarding their methodological and analytic strategies. About half of the studies based on data from randomized trials conducted a network intervention analysis, while the other half compared networks between treatment groups. The majority of studies estimated cross-sectional networks, even when repeated measures were available. All but five studies investigated networks on the group level. This review highlights that current methodological practices limit the information that can be gained through network analyses in intervention research. We discuss the strength and limitations of certain methodological and analytic strategies and propose that further work is needed to use the full potential of the network approach in intervention research.

Centrality statistics of symptom networks of schizophrenia: a systematic review.

The network theory of psychological disorders posits that systems of symptoms cause, or are associated with, the expression of other symptoms. Substantial literature on symptom networks has been published to date, although no systematic review has been conducted exclusively on symptom networks of schizophrenia, schizoaffective disorder, and schizophreniform (people diagnosed with schizophrenia; PDS). This study aims to compare statistics of the symptom network publications on PDS in the last 21 years and identify congruences and discrepancies in the literature. More specifically, we will focus on centrality statistics. Thirty-two studies met the inclusion criteria. The results suggest that cognition, and social, and occupational functioning are central to the network of symptoms. Positive symptoms, particularly delusions were central among participants in many studies that did not include cognitive assessment. Nodes representing cognition were most central in those studies that did. Nodes representing negative symptoms were not as central as items measuring positive symptoms. Some studies that included measures of mood and affect found items or subscales measuring depression were central nodes in the networks. Cognition, and social, and occupational functioning appear to be core symptoms of schizophrenia as they are more central in the networks, compared to variables assessing positive symptoms. This seems consistent despite heterogeneity in the design of the studies.

The added value of daily diary data in 1- and 3-year prediction of psychopathology and psychotic experiences in individuals at risk for psychosis.

This study aimed to assess whether adding information on psychological experiences derived from a daily diary to baseline cross-sectional data could improve short- (1-year) and long-term (3-years) prediction of psychopathology and positive psychotic experiences (PEs). We used 90-day daily diary data from 96 individuals in early subclinical risk stages for psychosis. Stepwise linear regression models were built for psychopathology and PEs at 1- and 3-years follow-up, adding: (1) baseline questionnaires, (2) the mean and variance of daily psychological experiences, and (3) individual symptom network density. We assessed whether similar results could be achieved with a subset of the data (7-14- and 30-days). The mean and variance of the diary improved model prediction of short- and long-term psychopathology and PEs, compared to prediction based on baseline questionnaires solely. Similar results were achieved with 7-14- and 30-day subsets. Symptom network density did not improve model prediction except for short-term prediction of PEs. Simple metrics, i.e., the mean and variance from 7 to 14 days of daily psychological experiences assessments, can improve short- and long-term prediction of both psychopathology and PEs in individuals in early subclinical stages for psychosis. Diary data could be a valuable addition to clinical risk prediction models for psychopathology development.

Cannabis Use Disorder Symptoms in Weekly Cannabis Users: A Network Comparison Between Daily Cigarette Users and Nondaily Cigarette Users.

Background: Concurrent use (co-use) of cannabis and tobacco is common and associated with worse clinical outcomes compared with cannabis use only. The mechanisms and interactions of cannabis use disorder (CUD) symptoms underlying co-use remain poorly understood. Methods: We examined differences in the symptom presence and symptom network configurations between weekly cannabis users who use tobacco daily (co-users, n=789) or non- or nondaily (nondaily co-users, n=428). Results: First, we identified a range of symptoms (craving, failed reduce or quit attempts, neglected responsibilities, and negative social effects) that are most central to the highly interconnected CUD symptom network. Risky cannabis use was mostly associated with negative social and health effects, and independent of other CUD symptoms. Craving symptoms act as a bridge between different CUD and withdrawal symptoms. Among co-users, (1) craving is more strongly associated with negative psychosocial effects, (2) feelings of depression and negative health effects are more central to the network, and (3) the negative health effects are more strongly associated with failed attempts to reduce or quit attempts compared with nondaily co-users. Discussion: Our results go beyond existing findings focused on the mere increase in CUD symptom presence, and speak to the potential synergistic effects of co-use on dependence and withdrawal symptoms. We outline clinical implications with respect to targeting specific CUD symptoms in co-users, and point to future research to disentangle tobacco and cannabis craving symptoms.

Deep phenotypic analysis of psychiatric features in genetically defined cohorts: application to XYY syndrome.

BACKGROUND: Recurrent gene dosage disorders impart substantial risk for psychopathology. Yet, understanding that risk is hampered by complex presentations that challenge classical diagnostic systems. Here, we present a suite of generalizable analytic approaches for parsing this clinical complexity, which we illustrate through application to XYY syndrome. METHOD: We gathered high-dimensional measures of psychopathology in 64 XYY individuals and 60 XY controls, plus additional interviewer-based diagnostic data in the XYY group. We provide the first comprehensive diagnostic description of psychiatric morbidity in XYY syndrome and show how diagnostic morbidity relates to functioning, subthreshold symptoms, and ascertainment bias. We then map behavioral vulnerabilities and resilience across 67 behavioral dimensions before borrowing techniques from network science to resolve the mesoscale architecture of these dimensions and links to observable functional outcomes. RESULTS: Carriage of an extra Y-chromosome increases risk for diverse psychiatric diagnoses, with clinically impactful subthreshold symptomatology. Highest rates are seen for neurodevelopmental and affective disorders. A lower bound of < 25% of carriers are free of any diagnosis. Dimensional analysis of 67 scales details the profile of psychopathology in XYY, which survives control for ascertainment bias, specifies attentional and social domains as the most impacted, and refutes stigmatizing historical associations between XYY and violence. Network modeling compresses all measured symptom scales into 8 modules with dissociable links to cognitive ability, adaptive function, and caregiver strain. Hub modules offer efficient proxies for the full symptom network. CONCLUSIONS: This study parses the complex behavioral phenotype of XYY syndrome by applying new and generalizable analytic approaches for analysis of deep-phenotypic psychiatric data in neurogenetic disorders.

Transdiagnostic symptom networks in relation to mental health service use in community youth.

OBJECTIVE: The objective of this study is to scrutinize whether psychopathology symptom networks differ between those with and without lifetime: treatment seeking, treatment and treatment of longer duration. METHODS: We created non-exclusive groups of subjects with versus without lifetime treatment seeking, treatment and treatment of mid-long-term duration. We estimated Ising models and carried out network comparison tests (NCTs) to compare (a) overall connectivity and (b) network structure. Furthermore, we examined node strength. We used propensity score matching (PSM) to minimize potential confounding by indication for service use. RESULTS: Based on data from 9,172 participants, there were no statistically significant differences in overall connectivity and network structure in those with versus without lifetime: treatment seeking (p = .75 and p = .82, respectively), treatment (p = .63 and p = .49, respectively) and treatment of mid-longterm duration (p = .15 and p = .62, respectively). Notably, comparing networks with versus without service use consistently revealed higher node strength in 'obsessions' and 'aggression' and lower node strength in 'elevated mood' in all networks with service use. CONCLUSIONS: Findings suggest that after adjusting for potential confounding by indication for service use, there was no indication of an association in overall connectivity or network structure for lifetime treatment seeking, treatment and treatment of longer duration. However, selected structurally important symptoms differed consistently in all three comparisons. Our findings highlight the potential of network analysis methods to examine treatment mechanisms and outcomes. Specifically, more granular network characteristics on the node level may complement and enrich traditional outcomes in clinical research.

Transdiagnostic Psychopathology in a Help-Seeking Population of an Early Recognition Center for Mental Disorders: Protocol for an Experience Sampling Study.

BACKGROUND: Prevention in psychiatry provides a promising way to address the burden of mental illness. However, established approaches focus on specific diagnoses and do not address the heterogeneity and manifold potential outcomes of help-seeking populations that present at early recognition services. Conceptualizing the psychopathology manifested in help-seeking populations from a network perspective of interacting symptoms allows transdiagnostic investigations beyond binary disease categories. Furthermore, modern technologies such as smartphones facilitate the application of the Experience Sampling Method (ESM). OBJECTIVE: This study is a combination of ESM with network analyses to provide valid insights beyond the established assessment instruments in a help-seeking population. METHODS: We will examine 75 individuals (aged 18-40 years) of the help-seeking population of the Cologne early recognition center. For a maximally naturalistic sample, only minimal exclusion criteria will be applied. We will collect data for 14 days using a mobile app to assess 10 transdiagnostic symptoms (ie, depressive, anxious, and psychotic symptoms) as well as distress level 5 times a day. With these data, we will generate average group-level symptom networks and personalized symptom networks using a 2-step multilevel vector autoregressive model. Additionally, we will explore associations between symptom networks and sociodemographic, risk, and resilience factors, as well as psychosocial functioning. RESULTS: The protocol was designed in February 2020 and approved by the Ethics Committee of the University Hospital Cologne in October 2020. The protocol was reviewed and funded by the Köln Fortune program in September 2020. Data collection began in November 2020 and was completed in November 2021. Of the 258 participants who were screened, 93 (36%) fulfilled the inclusion criteria and were willing to participate in the study. Of these 93 participants, 86 (92%) completed the study. The first results are expected to be published in 2022. CONCLUSIONS: This study will provide insights about the feasibility and utility of the ESM in a help-seeking population of an early recognition center. Providing the first explorative phenotyping of transdiagnostic psychopathology in this population, our study will contribute to the innovation of early recognition in psychiatry. The results will help pave the way for prevention and targeted early intervention in a broader patient group, and thus, enable greater intended effects in alleviating the burden of psychiatric disorders. INTERNATIONAL REGISTERED REPORT IDENTIFIER (IRRID): DERR1-10.2196/35206.

Cross-sectional and longitudinal AUD symptom networks: They tell different stories.

Modern theoretical models of Alcohol Use Disorder (AUD) highlight the different functional roles played by various mechanisms associated with different symptoms. Symptom network models (SNMs) offer one approach to modeling AUD symptomatology in a way that could reflect these processes and provide important information on the progression and persistence of disorder. However, much of the research conducted using SNMs relies on cross-sectional data, which has raised questions regarding the extent they reflect dynamic processes. The current study aimed to (a) examine symptom networks of AUD and (b) compare the extent to which cross-sectional network models had similar structures and interpretations as longitudinal network models. 17,360 participants from Wave 1 (2001-2002) and Wave 2 (2003-2004) of the National Epidemiological Survey on Alcohol and Related Conditions (NESARC) were used to model cross-sectional and longitudinal AUD symptom networks. The cross-sectional analyses demonstrate high replicability across waves and central symptoms consistent with other cross-sectional studies on addiction networks. The longitudinal network shared much less similarity than the cross-sectional networks and had a substantially different structure. Given the increasing attention given to the network perspective in psychopathology research, the results of this study raise concerns about interpreting cross-sectional symptom networks as representative of temporal changes occurring within a psychological disorder. We conclude that the psychological symptom network literature should be bolstered with additional research on longitudinal network models.

Disentangling relationships in symptom networks using matrix permutation methods.

Common outputs of software programs for network estimation include association matrices containing the edge weights between pairs of symptoms and a plot of the symptom network. Although such outputs are useful, it is sometimes difficult to ascertain structural relationships among symptoms from these types of output alone. We propose that matrix permutation provides a simple, yet effective, approach for clarifying the order relationships among the symptoms based on the edge weights of the network. For directed symptom networks, we use a permutation criterion that has classic applications in electrical circuit theory and economics. This criterion can be used to place symptoms that strongly predict other symptoms at the beginning of the ordering, and symptoms that are strongly predicted by other symptoms at the end. For undirected symptom networks, we recommend a permutation criterion that is based on location theory in the field of operations research. When using this criterion, symptoms with many strong ties tend to be placed centrally in the ordering, whereas weakly-tied symptoms are placed at the ends. The permutation optimization problems are solved using dynamic programming. We also make use of branch-search algorithms for extracting maximum cardinality subsets of symptoms that have perfect structure with respect to a selected criterion. Software for implementing the dynamic programming algorithms is available in MATLAB and R. Two networks from the literature are used to demonstrate the matrix permutation algorithms.

Perceived causal symptom network of adolescent mental health issues.

Adolescent mental health is difficult to capture in categories such as depression or specific anxiety disorders. An alternative is to approach psychiatric symptoms as causal networks, potentially revealing feedback loops that maintain a pathological state. One approach to creating such networks, implemented in the PECAN methodology, is to ask adolescents about their perceptions of the causes to their symptoms. For this purpose, a transdiagnostic item list was created, and adolescents who screened positive for depression (N = 55) completed twice in two weeks a survey quantifying perceptions of causality between their mental health problems. A network that was averaged across all participants was reliable and revealed three strong feedback loops: a first loop running through stress, insomnia, fatigue, procrastination, and back to stress; a second loop between stress and overthinking; and a third loop between stress and procrastination. Although all adolescents in the study screened positive for depression, symptoms of depression were not particularly central to the network. Instead, the most central symptoms were procrastination and overthinking. The average test-retest reliability for individual networks was low, limiting clinical application. In conclusion, PECAN was found to be reliable and useful when creating a group-level network of adolescent mental health problems. While informative at a group level, the method should be improved before it can be used to inform treatment at the individual level.

Symptom networks in major depression do not diverge across sex, familial risk, and environmental risk.

BACKGROUND: Major depression (MD) is a heterogeneous disorder in terms of its symptoms. Symptoms vary by presence of risk factors such as female sex, familial risk, and environmental adversity. However, it is unclear if these factors also influence interactions between symptoms. This study investigates if symptom networks diverge across sex, familial risk, and adversity. METHODS: We included 9713 subjects from the general population who reported a lifetime episode of MD based on DSM-IV criteria. The survey assessed a wide set of symptoms, both from within the DSM criteria as well as other symptoms commonly experienced in MD. We compared symptom endorsement rates across sex, age at onset, family history and environmental adversity. We used the Network Comparison Test to test for symptom network differences across risk factors. RESULTS: We found differences in symptom endorsement between groups. For instance, participants with an early onset of MD reported suicidal ideation nearly twice as often compared to participants with a later onset. We did not find any robust differences in symptom networks, which suggests that symptom networks do not diverge across sex, familial risk, and adversity. LIMITATIONS: We estimated symptom networks of individuals during their worst lifetime episode of MD. Network differences might exist in a prodromal stage, while disappearing in full-blown MD (equifinality). Furthermore, as we used retrospective reports, results could be prone to recall bias. CONCLUSIONS: Despite MD's heterogeneous symptomatology, interactions between symptoms are stable across risk factors and sex.

Ambulatory assessment in psychopathology research: Current achievements and future ambitions.

Ambulatory assessment (AA) - a collection of methods that aim to track individuals in the realm of everyday life via repeated self-reports or passive mobile sensing - is well established in contemporary psychopathology research. Unravelling the dynamic signature of patients' symptoms and emotions over time and in their own personal ecology, AA methodology has improved our understanding of the real-time pathogenic processes that underlie mental ill-being. In this article, we evaluate the current strengths and shortcomings of AA in psychopathology research and spell out important ambitions for next-generation AA studies to consider. Regarding AA's current achievements, a selective review of recent AA studies underscores the ecological qualities of this method, its ability to bypass retrospective biases in self-report and the introduction of a within-person perspective. Regarding AA's future ambitions, we advocate for a stronger idiosyncratic focus, the incorporation of contextual information and more psychometric scrutiny.

On Unreplicable Inferences in Psychopathology Symptom Networks and the Importance of Unreliable Parameter Estimates.

We recently wrote an article comparing the conclusions that followed from two different approaches to quantifying the reliability and replicability of psychopathology symptom networks. Two commentaries on the article have raised five core criticisms, which are addressed in this response with supporting evidence. 1) We did not over-generalize about the replicability of symptom networks, but rather focused on interpreting the contradictory conclusions of the two sets of methods we examined. 2) We closely followed established recommendations when estimating and interpreting the networks. 3) We also closely followed the relevant tutorials, and used examples interpreted by experts in the field, to interpret the bootnet and NetworkComparisonTest results. 4) It is possible for statistical control to increase reliability, but that does not appear to be the case here. 5) Distinguishing between statistically significant versus substantive differences makes it clear that the differences between the networks affect the inferences we would make about symptom-level relationships (i.e., the basis of the purported utility of symptom networks). Ultimately, there is an important point of agreement between our article and the commentaries: All of these applied examples of cross-sectional symptom networks are demonstrating unreliable parameter estimates. While the commentaries propose that the resulting differences between networks are not genuine or meaningful because they are not statistically significant, we propose that the unreplicable inferences about the symptom-level relationships of interest fundamentally undermine the utility of the symptom networks.

The network approach to psychopathology: a review of the literature 2008-2018 and an agenda for future research.

The network approach to psychopathology posits that mental disorders can be conceptualized and studied as causal systems of mutually reinforcing symptoms. This approach, first posited in 2008, has grown substantially over the past decade and is now a full-fledged area of psychiatric research. In this article, we provide an overview and critical analysis of 363 articles produced in the first decade of this research program, with a focus on key theoretical, methodological, and empirical contributions. In addition, we turn our attention to the next decade of the network approach and propose critical avenues for future research in each of these domains. We argue that this program of research will be best served by working toward two overarching aims: (a) the identification of robust empirical phenomena and (b) the development of formal theories that can explain those phenomena. We recommend specific steps forward within this broad framework and argue that these steps are necessary if the network approach is to develop into a progressive program of research capable of producing a cumulative body of knowledge about how specific mental disorders operate as causal systems.

A new conception and subsequent taxonomy of clinical psychological problems.

BACKGROUND: A taxonomy of the objects of study, theory, assessment, and intervention is critical to the development of all clinical sciences. Clinical psychology has been conceptually and administratively dominated by the taxonomy of an adjacent discipline - psychiatry's Diagnostic and statistical manual of mental disorders (DSM). Many have called for a 'paradigm shift' away from a medical nosology of diseases toward clinical psychology's own taxonomy of clinical psychological problems (CPPs), without being able to specify what is to be listed and classified. MAIN TEXT: An examination of DSM's problems for clinical psychology, especially its lack of clinical utility, and a search for the essence of CPPs in what clinical psychologists actually do, leads to the proposal that: The critical psychological-level phenomenon underlying CPPs is the occurrence of 'problem-maintaining circles' (PMCs) of causally related cognitions, emotions, behaviours, and/or stimuli. This concept provides an empirically-derived, theory-based, treatment-relevant, categorical, essentialist, parsimonious, and nonstigmatizing definition of CPPs. It distinguishes psychological problems in which PMCs have not (yet?) formed, and which may respond to 'counseling', clinical psychological problems in which active PMCs require clinical intervention, and psychopathological problems which are unlikely to be 'cured' by PMC-breaking alone. CONCLUSION: A subsequent classification and coding system of PMCs is proposed, and expected benefits to research, communication, and the quality of case formulation in clinical psychology are described, reliant upon a development effort of some meaningful fraction of that which has been devoted to the DSM.

Outcome measurement in mental health services: insights from symptom networks.

BACKGROUND: In mental health, outcomes are currently measured by changes of individual scores. However, such an analysis on individual scores does not take into account the interaction between symptoms, which could yield crucial information while investigating outcomes. Network analysis techniques can be used to routinely study these systems of interacting symptoms. The present study aimed at comparing outcomes using individual scores vs. symptom networks, after a 1 year intervention at a local community mental health centre. METHODS: We used the Health of the Nation Outcomes Scales, which defines a set of 12 scales investigating mental health and social functioning. We first assessed how individual scores varied from baseline to end point and which items were associated to treatment response. Second, using network analysis techniques, we measured the overall connectivity of the networks and determined the most important symptoms. RESULTS: The individual scores analysis revealed a significant improvement amongst most scales. No specific factors were related to treatment response at end point. At end point, network analysis revealed a very densely connected network while agitation and substance use were the most connected symptoms. CONCLUSIONS: Individual scores and symptom network analysis resulted in very different outcomes, with network analysis toning down positive results gained from individual scores analysis. The strong connectivity of patients' network at end point may reflect their increased complexity. Allocating more resources to interventions tailored to symptoms that are the most connected would decrease network connectivity and improve patients' prognosis. When investigating outcomes, network analysis could give insights complementary to standard analysis on individual scores.

A Network Approach to Understanding the Structure of Core Symptoms of Psychopathic Personality Disturbance in Adolescent Offenders.

A central aim of research on psychopathic personality disturbance (PPD) involves identifying core features of the construct. This has been addressed primarily through prototypicality studies and research using item-response theory. More recently, the logic of social network analysis was extended to psychopathology research to examine which symptoms were most central to PPD networks. Such studies identified affective symptoms of the disorder as especially central among adult offenders. To build upon this prior research, the current study used data on male offenders from the Incarcerated Serious and Violent Young Offender Study to examine the network structure of the Comprehensive Assessment of Psychopathic Personality - Institutional Rating Scale (CAPP-IRS; n = 224) and Psychopathy Checklist: Youth Version (PCL:YV; n = 445). Using multiple measures of PPD helped avoid equating measures with constructs. In both the CAPP-IRS and PCL:YV networks, in line with prior studies, attachment/affective features of the disorder were most central. Several recommendations are made for future research, including the need to study the longitudinal development of PPD using a network approach.

A network theory of mental disorders.

In recent years, the network approach to psychopathology has been advanced as an alternative way of conceptualizing mental disorders. In this approach, mental disorders arise from direct interactions between symptoms. Although the network approach has led to many novel methodologies and substantive applications, it has not yet been fully articulated as a scientific theory of mental disorders. The present paper aims to develop such a theory, by postulating a limited set of theoretical principles regarding the structure and dynamics of symptom networks. At the heart of the theory lies the notion that symptoms of psychopathology are causally connected through myriads of biological, psychological and societal mechanisms. If these causal relations are sufficiently strong, symptoms can generate a level of feedback that renders them self-sustaining. In this case, the network can get stuck in a disorder state. The network theory holds that this is a general feature of mental disorders, which can therefore be understood as alternative stable states of strongly connected symptom networks. This idea naturally leads to a comprehensive model of psychopathology, encompassing a common explanatory model for mental disorders, as well as novel definitions of associated concepts such as mental health, resilience, vulnerability and liability. In addition, the network theory has direct implications for how to understand diagnosis and treatment, and suggests a clear agenda for future research in psychiatry and associated disciplines.