Extending psychometric network analysis: Empirical evidence against g in favor of mutualism?

Perturbation graphs, invariant causal prediction and causal relations in psychology

Networks (graphs) in psychology are often restricted to settings without interventions. Here we consider a framework borrowed from biology that involves multiple interventions from different contexts (observations and experiments) in a single analysis. The method is called perturbation graphs. In gene regulatory networks, the induced change in one gene is measured on all other genes in the analysis, thereby assessing possible causal relations. This is repeated for each gene in the analysis. A perturbation graph leads to the correct set of causes (not nec-essarily direct causes). Subsequent pruning of paths in the graph (called transitive reduction) should reveal direct causes. We show that transitive reduction will not in general lead to the correct underlying graph. We also show that invariant causal prediction is a generalisation of the perturbation graph method and does reveal direct causes, thereby replacing transitive re-duction. We conclude that perturbation graphs provide a promising new tool for experimental designs in psychology, and combined with invariant causal prediction make it possible to re-veal direct causes instead of causal paths. As an illustration we apply these ideas to a data set about attitudes on meat consumption and to a time series of a patient diagnosed with major depression disorder.

Depressive symptoms, cognitive functions and daily activities: An extended network analysis in monozygotic and dizygotic twins

BACKGROUND

The network theory posits that depression emerges as the result of individual symptoms triggering each other. Risk factors for depression can impact these between-symptoms interactions through extended networks. The study aimed to model the extended network of depressive symptoms and known depression risk factors - objective cognitive function, intellectual, physical, and social daily activities, and then, compare the observed networks between monozygotic (MZ) and dizygotic (DZ) co-twins.

METHODS

Twin pairs, 722 MZ and 2200 DZ, aged 40-79, were selected from the Dansh Twin Registry for having complete measures of depressive symptoms (e.g., sadness), cognitive functions (e.g., verbal memory), physical (e.g., brisk walk), intellectual (e.g., reading newspapers) and social activities (e.g., phone calls). Gaussian graphical models were used to estimate and compare the networks first between co-twins and then, between MZ to DZ twin pairs separately.

RESULTS

Specific intellectual, physical and social activities were central in the extended networks of depressive symptoms and, with the exception of processing speed, more central than cognition. The extended networks' structure was more homogeneous between MZ co-twins relative to DZ co-twins. Cognitive nodes were more central in MZ than DZ co-twins.

LIMITATIONS

Cross-sectional design, participants were middle-aged or older, mostly affective (non-somatic) depressive symptoms.

CONCLUSIONS

In depression networks, core connecting elements were intellectual, physical and social activities. The interaction between cognition and daily activities seems critical for triggering depressive symptoms. Thus, clinical interventions aimed at preventing depression and associated cognitive deficits should focus on maintenance and/or engagement in stimulating daily activities.

More similar than different: Characterizing special interests in autistic boys and girls based on caregiver report

Almost all autistic youths have special interests (SIs), which are focused, intense areas of passion and interest in a particular topic. Emerging research suggests that there are gender differences in SIs among autistic youth; however, commonly used measures that assess for the presence of SIs may not fully capture the granular nature of those differences between autistic boys and girls. Characterizing these differences is important for autism identification in girls, as SIs in autistic girls may often be overlooked by caregivers, teachers, and clinicians due to their more "typical" and more socially oriented content areas compared to autistic boys. This study therefore aimed to more fully characterize gender differences in SIs using a newly developed caregiver-report measure of SIs (the Special Interests Survey; SIS). Caregivers of 1921 autistic youth completed the SIS. Analyses revealed many similarities between boys and girls; there were no gender differences in mean age SI onset, caregivers' perceptions of uniqueness or interferences of endorsed SIs, or duration of previous SIs. There were gender differences in endorsement of less than half (39%) of the SI categories measured, and there were minor differences in the number of endorsed current and past SIs. Categories with significant gender differences fell along typical gender lines (e.g., more boys interested in math and construction, more girls interested in animals and arts/crafts). This study extends the growing literature on SIs and gender differences in autism and has important implications for supporting autistic youth and their families.

Insights on cognitive reorganization after hemispherectomy in Rasmussen's encephalitis. A narrative review

Rasmussen's encephalitis is a rare neurological pathology affecting one cerebral hemisphere, therefore, posing unique challenges. Patients may undergo hemispherectomy, a surgical procedure after which cognitive development occurs in the isolated contralateral hemisphere. This rare situation provides an excellent opportunity to evaluate brain plasticity and cognitive recovery at a hemispheric level. This literature review synthesizes the existing body of research on cognitive recovery following hemispherectomy in Rasmussen patients, considering cognitive domains and modulatory factors that influence cognitive outcomes. While language function has traditionally been the focus of postoperative assessments, there is a growing acknowledgment of the need to broaden the scope of language investigation in interaction with other cognitive domains and to consider cognitive scaffolding in development and recovery. By synthesizing findings reported in the literature, we delineate how language functions may find support from the right hemisphere after left hemispherectomy, but also how, beyond language, global cognitive functioning is affected. We highlight the critical influence of several factors on postoperative cognitive outcomes, including the timing of hemispherectomy and the baseline preoperative cognitive status, pointing to early surgical intervention as predictive of better cognitive outcomes. However, further specific studies are needed to confirm this correlation. This review aims to emphasize a better understanding of mechanisms underlying hemispheric specialization and plasticity in humans, which are particularly important for both clinical and research advancements. This narrative review underscores the need for an integrative approach based on cognitive scaffolding to provide a comprehensive understanding of mechanisms underlying the reorganization in Rasmussen patients after hemispherectomy.

Unveiling the cognitive network organization through cognitive performance

The evaluation of cognitive functions interactions has become increasingly implemented in the cognition exploration. In the present study, we propose to examine the organization of the cognitive network in healthy participants through the analysis of behavioral performances in several cognitive domains. Specifically, we aim to explore cognitive interactions profiles, in terms of cognitive network, and as a function of participants' handedness. To this end, we proposed several behavioral tasks evaluating language, memory, executive functions, and social cognition performances in 175 young healthy right-handed and left-handed participants and we analyzed cognitive scores, from a network perspective, using graph theory. Our results highlight the existence of intricate interactions between cognitive functions both within and beyond the same cognitive domain. Language functions are interrelated with executive functions and memory in healthy cognitive functioning and assume a central role in the cognitive network. Interestingly, for similar high performance, our findings unveiled differential organizations within the cognitive network between right-handed and left-handed participants, with variations observed both at a global and nodal level. This original integrative network approach to the study of cognition provides new insights into cognitive interactions and modulations. It allows a more global understanding and consideration of cognitive functioning, from which complex behaviors emerge.

Variation in global network properties across risk factors for adolescent internalizing symptoms: evidence of cumulative effects on structure and connectivity

BACKGROUND

Identifying adolescents at risk of internalizing problems is a key priority. However, studies have tended to consider such problems in simple ways using diagnoses, or item summaries. Network theory and methods instead allow for more complex interaction between symptoms. Two key hypotheses predict differences in global network properties for those at risk: altered structure and increased connectivity.

METHODS

The current study evaluated these hypotheses for nine risk factors (e.g. income deprivation and low parent/carer support) individually and cumulatively in a large sample of 12-15 year-olds (N = 34 564). Recursive partitioning and bootstrapped networks were used to evaluate structural and connectivity differences.

RESULTS

The pattern of network interactions was shown to be significantly different via recursive partitioning for all comparisons across risk-present/absent groups and levels of cumulative risk, except for income deprivation. However, the magnitude of differences appeared small. Most individual risk factors also showed relatively small effects for connectivity. Exceptions were noted for gender and sexual minority risk groups, as well as low parent/carer support, where larger effects were evident. A strong linear trend was observed between increasing cumulative risk exposure and connectivity.

CONCLUSIONS

A robust approach to considering the effect of risk exposure on global network properties was demonstrated. Results are consistent with the ideas that pathological states are associated with higher connectivity, and that the number of risks, regardless of their nature, is important. Gender/sexual minority status and low parent/carer support had the biggest individual impacts on connectivity, suggesting these are particularly important for identification and prevention.

Why Do Bi-Factor Models Outperform Higher-Order g Factor Models? A Network Perspective

Bi-factor models of intelligence tend to outperform higher-order g factor models statistically. The literature provides the following rivalling explanations: (i) the bi-factor model represents or closely approximates the true underlying data-generating mechanism; (ii) fit indices are biased against the higher-order g factor model in favor of the bi-factor model; (iii) a network structure underlies the data. We used a Monte Carlo simulation to investigate the validity and plausibility of each of these explanations, while controlling for their rivals. To this end, we generated 1000 sample data sets according to three competing models-a bi-factor model, a (nested) higher-order factor model, and a (non-nested) network model-with 3000 data sets in total. Parameter values were based on the confirmatory analyses of the Wechsler Scale of Intelligence IV. On each simulated data set, we (1) refitted the three models, (2) obtained the fit statistics, and (3) performed a model selection procedure. We found no evidence that the fit measures themselves are biased, but conclude that biased inferences can arise when approximate or incremental fit indices are used as if they were relative fit measures. The validity of the network explanation was established while the outcomes of our network simulations were consistent with previously reported empirical findings, indicating that the network explanation is also a plausible one. The empirical findings are inconsistent with the (also validated) hypothesis that a bi-factor model is the true model. In future model selection procedures, we recommend that researchers consider network models of intelligence, especially when a higher-order g factor model is rejected in favor of a bi-factor model.

The interrelatedness of cognitive abilities in very preterm and full-term born children at 5.5 years of age: a psychometric network analysis approach

BACKGROUND

Very preterm (VP) birth is associated with a considerable risk for cognitive impairment, putting children at a disadvantage in academic and everyday life. Despite lower cognitive ability on the group level, there are large individual differences among VP born children. Contemporary theories define intelligence as a network of reciprocally connected cognitive abilities. Therefore, intelligence was studied as a network of interrelated abilities to provide insight into interindividual differences. We described and compared the network of cognitive abilities, including strength of interrelations between and the relative importance of abilities, of VP and full-term (FT) born children and VP children with below-average and average-high intelligence at 5.5 years.

METHODS

A total of 2,253 VP children from the EPIPAGE-2 cohort and 578 FT controls who participated in the 5.5-year-follow-up were eligible for inclusion. The WPPSI-IV was used to measure verbal comprehension, visuospatial abilities, fluid reasoning, working memory, and processing speed. Psychometric network analysis was applied to analyse the data.

RESULTS

Cognitive abilities were densely and positively interconnected in all networks, but the strength of connections differed between networks. The cognitive network of VP children was more strongly interconnected than that of FT children. Furthermore, VP children with below average IQ had a more strongly connected network than VP children with average-high IQ. Contrary to our expectations, working memory had the least central role in all networks.

CONCLUSIONS

In line with the ability differentiation hypothesis, children with higher levels of cognitive ability had a less interconnected and more specialised cognitive structure. Composite intelligence scores may therefore mask domain-specific deficits, particularly in children at risk for cognitive impairments (e.g., VP born children), even when general intelligence is unimpaired. In children with strongly and densely connected networks, domain-specific deficits may have a larger overall impact, resulting in lower intelligence levels.

The Psychometric Structure of Executive Functions: A Satisfactory Measurement Model? An Examination Using Meta-Analysis and Network Modeling

A long-standing debate among cognitive scientists has focused on describing the underlying nature of executive functions, which has important implications for both theoretical and applied research. Miyake et al.'s three-factor model has often been considered the gold-standard representation of executive functions and has driven much research in the field. More recently, however, there have been increasing concerns that the three-factor model does not adequately describe a highly complex construct such as executive functions. The current project presents two studies that examine the veracity of Miyake et al.'s model and propose a new approach (i.e., network modeling) for detecting the underlying nature of executive functions. The current results raise questions about the psychometric strength and adequacy of the three-factor model. Further, the studies presented here provide evidence that network modeling provides a better understanding of executive functions as it better captures (relative to latent variable modeling) the complexity of cognitive processes. Theoretical and applied implications are discussed.

The general psychopathology factor (p) from adolescence to adulthood: Exploring the developmental trajectories of p using a multi-method approach

Considerable attention has been directed towards studying co-occurring psychopathology through the lens of a general factor (p-factor). However, the developmental trajectory and stability of the p-factor have yet to be fully understood. The present study examined the explanatory power of dynamic mutualism theory - an alternative framework that suggests the p-factor is a product of lower-level symptom interactions that strengthen throughout development. Data were drawn from a population-based sample of girls (N = 2450) who reported on the severity of internalizing and externalizing problems each year from age 14 to age 21. Predictions of dynamic mutualism were tested using three distinct complementary statistical approaches including: longitudinal bifactor models, random-intercept cross-lagged panel models (RI-CLPMs), and network models. Across methods, study results document preliminary support for mutualistic processes in the development of co-occurring psychopathology (that is captured in p). Findings emphasize the importance of exploring alternative frameworks and methods for better understanding the p-factor and its development.

Investigating the Structure of the French WISC-V (WISC-VFR) for Five Age Groups Using Psychometric Network Modeling

Since the seminal work of Spearman, exploratory and confirmatory factor analysis represents the standard method of examining the dimensionality of psychological instruments. Recently, within the network psychometrics approach, a new procedure was proposed to estimate the dimensionality of psychological instruments: exploratory graph analysis (EGA). This study investigated the structure of the French Wechsler Intelligence Scale for Children-Fifth Edition (WISC-VFR) with five standardization sample age groups (6-7, 8-9, 10-11, 12-13, and 14-16 years) using EGA. The primary research questions include (a) how many WISC-VFR dimensions are identified in each age subgroup? (b) how are subtest scores associated within the dimensions? Because the number and the content of the dimensions identified by EGA could vary with samples, the secondary research questions include (c) is there evidence of reproducibility and generalizability of the dimensions identified by EGA? We used another procedure called bootstrap exploratory graph analysis (bootEGA). EGA and bootEGA suggested only three dimensions, which are consistent with processing speed, verbal comprehension, and the "old" perceptual reasoning factor. Results did not support the distinction between visual-spatial and fluid reasoning dimensions. EGA and bootEGA represent new tools to assess the construct validity of psychological instruments, such as the WISC-VFR.

Better together: novel methods for measuring and modeling development of executive function diversity while accounting for unity

INTRODUCTION

Executive functions (EFs) are linked to positive outcomes across the lifespan. Yet, methodological challenges have prevented precise understanding of the developmental trajectory of their organization.

METHODS

We introduce novel methods to address challenges for both measuring and modeling EFs using an accelerated longitudinal design with a large, diverse sample of students in middle childhood (N = 1,286; ages 8 to 14). We used eight adaptive assessments hypothesized to measure three EFs, working memory, context monitoring, and interference resolution. We deployed adaptive assessments to equate EF challenge across ages and a data-driven, network analytic approach to reveal the evolving diversity of EFs while simultaneously accounting for their unity.

RESULTS AND DISCUSSION

Using this methodological paradigm shift brought new precision and clarity to the development of these EFs, showing these eight tasks are organized into three stable components by age 10, but refinement of composition of these components continues through at least age 14.

Bootstrap Exploratory Graph Analysis of the WISC-V with a Clinical Sample

One important aspect of construct validity is structural validity. Structural validity refers to the degree to which scores of a psychological test are a reflection of the dimensionality of the construct being measured. A factor analysis, which assumes that unobserved latent variables are responsible for the covariation among observed test scores, has traditionally been employed to provide structural validity evidence. Factor analytic studies have variously suggested either four or five dimensions for the WISC-V and it is unlikely that any new factor analytic study will resolve this dimensional dilemma. Unlike a factor analysis, an exploratory graph analysis (EGA) does not assume a common latent cause of covariances between test scores. Rather, an EGA identifies dimensions by locating strongly connected sets of scores that form coherent sub-networks within the overall network. Accordingly, the present study employed a bootstrap EGA technique to investigate the structure of the 10 WISC-V primary subtests using a large clinical sample (N = 7149) with a mean age of 10.7 years and a standard deviation of 2.8 years. The resulting structure was composed of four sub-networks that paralleled the first-order factor structure reported in many studies where the fluid reasoning and visual-spatial dimensions merged into a single dimension. These results suggest that discrepant construct and scoring structures exist for the WISC-V that potentially raise serious concerns about the test interpretations of psychologists who employ the test structure preferred by the publisher.

Faster ≠ Smarter: Children with Higher Levels of Ability Take Longer to Give Incorrect Answers, Especially When the Task Matches Their Ability

The stereotype that children who are more able solve tasks quicker than their less capable peers exists both in and outside education. The F > C phenomenon and the distance–difficulty hypothesis offer alternative explanations of the time needed to complete a task; the former by the response correctness and the latter by the relative difference between the difficulty of the task and the ability of the examinee. To test these alternative explanations, we extracted IRT-based ability estimates and task difficulties from a sample of 514 children, 53% girls, M(age) = 10.3 years; who answered 29 Piagetian balance beam tasks. We used the answer correctness and task difficulty as predictors in multilevel regression models when controlling for children’s ability levels. Our results challenge the ‘faster equals smarter’ stereotype. We show that ability levels predict the time needed to solve a task when the task is solved incorrectly, though only with moderately and highly difficult items. Moreover, children with higher ability levels take longer to answer items incorrectly, and tasks equal to children’s ability levels take more time than very easy or difficult tasks. We conclude that the relationship between ability, task difficulty, and answer correctness is complex, and warn education professionals against basing their professional judgment on students’ quickness.

Carroll's Three-Stratum (3S) Cognitive Ability Theory at 30 Years: Impact, 3S-CHC Theory Clarification, Structural Replication, and Cognitive-Achievement Psychometric Network Analysis Extension

Carroll's treatise on the structure of human cognitive abilities is a milestone in psychometric intelligence research. Thirty years later, Carroll's work continues to influence research on intelligence theories and the development and interpretation of intelligence tests. A historical review of the relations between the 3S and CHC theories necessitates the recommendation that the theories of Cattell, Horn, and Carroll be reframed as a family of obliquely correlated CHC theories-not a single CHC theory. Next, a previously unpublished Carroll exploratory factor analysis of 46 cognitive and achievement tests is presented. A complimentary bifactor analysis is presented that reinforces Carroll's conclusion that his 3S model more accurately represents the structure of human intelligence than two prominent alternative models. Finally, a Carroll-recommended higher-stratum psychometric network analysis (PNA) of CHC cognitive, reading, and math variables is presented. The PNA results demonstrate how PNA can complement factor analysis and serve as a framework for identifying and empirically evaluating cognitive-achievement causal relations and mechanisms (e.g., developmental cascade and investment theories), with an eye toward improved cognitive-achievement intervention research. It is believed that Carroll, given his long-standing interest in school learning, would welcome the integration of theory-driven factor and PNA research.

Using theory to guide exploratory network analyses

The use of exploratory network analysis has increased in psychopathology research over the past decade. A benefit of exploratory network analysis is the wealth of information it can provide; however, a single analysis may generate more inferences than what can be discussed in one manuscript (e.g., centrality indices of each node). This necessitates that authors choose which results to discuss in further detail and which to omit. Without a guide for this process, the likelihood of a biased interpretation is high. We propose that the integration of theory throughout the research process makes the interpretation of exploratory networks more manageable for the researcher and more likely to result in an interpretation that advances science. The goals of this paper are to differentiate between exploratory and confirmatory network analyses, discuss the utility of exploratory work, and provide a practical framework that uses theory as a guide to interpret exploratory network analyses.

Changes in ImPACT Cognitive Subtest Networks Following Sport-Related Concussion

OBJECTIVE

High school athletes are administered ImPACT at the start of the academic year or sport season and again after suspected concussion. Concussion management involves the comparison of baseline and post-injury cognitive scores with declines in scores providing evidence for concussive injury. A network framework may provide additional information about post-concussive cognitive changes and expand characterization of sport-related concussion (SRC) recovery.

DESIGN

Retrospective cohort study.

SETTING

High school.

PARTICIPANTS

High school athletes (n = 1553) were administered ImPACT at baseline (T1), post-SRC (T2 = 72 h of injury), and prior to return to play (T3 = within two weeks post-injury).

INDEPENDENT VARIABLES

ImPACT cognitive subtest scores.

MAIN OUTCOME MEASURES

Cognitive networks were calculated and compared over three time points. Centrality indices were calculated to determine the relative importance of cognitive variables within networks.

RESULTS

Network connectivity increased from T1 to T2 and remained hyperconnected at T3. There was evidence of network reorganization between T1 and T3. Processing speed was central within each network, and visual memory and impulsivity became more central over time.

CONCLUSIONS

The results suggest potential evidence of cognitive network change over time. Centrality findings suggest research specific to visual memory and impulse control difficulties during the post-concussion recovery period is warranted. Network analysis may provide additional information about cognitive recovery following SRC and could potentially serve as an effective means of monitoring persisting cognitive symptoms after concussion.

A Psychometric Network Analysis of CHC Intelligence Measures: Implications for Research, Theory, and Interpretation of Broad CHC Scores "Beyond g"

For over a century, the structure of intelligence has been dominated by factor analytic methods that presume tests are indicators of latent entities (e.g., general intelligence or g). Recently, psychometric network methods and theories (e.g., process overlap theory; dynamic mutualism) have provided alternatives to g-centric factor models. However, few studies have investigated contemporary cognitive measures using network methods. We apply a Gaussian graphical network model to the age 9-19 standardization sample of the Woodcock-Johnson Tests of Cognitive Ability-Fourth Edition. Results support the primary broad abilities from the Cattell-Horn-Carroll (CHC) theory and suggest that the working memory-attentional control complex may be central to understanding a CHC network model of intelligence. Supplementary multidimensional scaling analyses indicate the existence of possible higher-order dimensions (PPIK; triadic theory; System I-II cognitive processing) as well as separate learning and retrieval aspects of long-term memory. Overall, the network approach offers a viable alternative to factor models with a g-centric bias (i.e., bifactor models) that have led to erroneous conclusions regarding the utility of broad CHC scores in test interpretation beyond the full-scale IQ, g.

Cognitive impairment networks in Alzheimer's disease: Analysis of three double-blind randomized, placebo-controlled, clinical trials of donepezil

Psychometric network analysis is an alternative theoretically-driven analytic approach that has the potential to conceptualize cognitive impairment in Alzheimer's disease differently than was previously assumed and consequently detect unknown treatment effects. Based on individual participant data, extracted from three double-blind, randomized placebo-controlled clinical trials, psychometric networks were computed on observed Alzheimer's Disease Assessment Scale Cognitive Subscale scores at baseline (N=1,554) and on predicted change scores at 24 weeks of follow-up for participants who received donepezil (N=797) or placebo (N=484). A novel conceptualization of cognitive impairment in Alzheimer's disease was displayed through the baseline network, that had 90% (n=27) positive statistically significant (p<0.05) associations, and a most central aspect of ideational praxis. Following 24 weeks, treatment effects emerged via the differences between the change score networks. The donepezil network had more statistically significant (p<0.05) positive associations and a higher global strength (n=15; S=1.22; p=0.03), than the placebo network (n=8; S=0.57). This suggests that for those who were treated with donepezil compared with placebo, cognition is a more unified construct. The main aspects of change in cognitive impairment were comprehension of spoken language for the donepezil network and spoken language ability for the placebo network. Comprehension of spoken language apears to be most sensitive to psychopharmaceutical interventions and should therefore be closely monitored. Overall, our psychometric network analysis presents a new conceptualization of cognitive impairment in Alzheimer's disease, points to previously unknown treatment effects and highlights well-defined aspects of cognitive impairment  that may translate into future treatment targets.

Meta-analytic Gaussian Network Aggregation

A growing number of publications focus on estimating Gaussian graphical models (GGM, networks of partial correlation coefficients). At the same time, generalizibility and replicability of these highly parameterized models are debated, and sample sizes typically found in datasets may not be sufficient for estimating the underlying network structure. In addition, while recent work emerged that aims to compare networks based on different samples, these studies do not take potential cross-study heterogeneity into account. To this end, this paper introduces methods for estimating GGMs by aggregating over multiple datasets. We first introduce a general maximum likelihood estimation modeling framework in which all discussed models are embedded. This modeling framework is subsequently used to introduce meta-analytic Gaussian network aggregation (MAGNA). We discuss two variants: fixed-effects MAGNA, in which heterogeneity across studies is not taken into account, and random-effects MAGNA, which models sample correlations and takes heterogeneity into account. We assess the performance of MAGNA in large-scale simulation studies. Finally, we exemplify the method using four datasets of post-traumatic stress disorder (PTSD) symptoms, and summarize findings from a larger meta-analysis of PTSD symptom.

A network analysis of interoception, self-awareness, empathy, alexithymia, and autistic traits

Altered interoception has been consistently found in people with autism spectrum disorder (ASD), and this impairment may contribute to social cognitive dysfunctions. However, little is known regarding the intercorrelations between interoceptive sensibility, autistic, alexithymic, empathic, and self-related traits. We recruited 1360 non-clinical college students and adults to investigate the complex inter-relationship between these variables using network analysis. The resultant network revealed patterns connecting autistic traits to interoceptive sensibility, empathy, alexithymia, and self-awareness, with reasonable stability and test-retest consistency. The node of alexithymia exhibited the highest centrality and expected influence. As revealed by the network comparison test, networks constructed in high- and low-autistic subgroups were comparable in global strength and structure. Our findings suggested that alexithymia serves as an important node, bridging interoceptive deficits, self-awareness, and empathic impairments of autism spectrum disorder. The co-morbidity of alexithymia should be considered carefully in future studies of interoceptive impairments and social deficits in ASD.

Utilizing Cognitive Training to Improve Working Memory, Attention, and Impulsivity in School-Aged Children with ADHD and SLD

Students’ use of working memory (WM) is a key to academic success, as many subject areas and various tasks school-aged children encounter require the ability to attend to, work with, and recall information. Children with poor WM ability typically struggle with academic work compared to similar-aged peers without WM deficits. Further, WM has been shown to be significantly correlated with inattention and disorganization in those with ADHD, and WM deficits have also been identified as a potential underpinning of specific learning disorder (SLD). As an intervention technique, the use of computerized cognitive training has demonstrated improved attention and working memory skills in children with WM deficits, and children that have completed cognitive training protocols have demonstrated performance improvements in reading and math. The current study aimed to examine the effectiveness of cognitive training (conducted in a clinical setting) for students diagnosed with ADHD and SLD. Using paired-samples t-tests and a psychometric network modeling technique, results from data obtained from a sample of 43 school-aged children showed (1) that attention and working memory improved following cognitive training and (2) that cognitive training might be related to cognitive structural changes found pre- to post-training among the variables being measured. Implications for clinical practice and school-based interventions are discussed.

Network neuropsychology: The map and the territory

In "network neuropsychology", network modelling and graph theory is applied to the neuropsychological test scores of patients with neurological disorders to investigate cognitive functioning. This review identifies the emerging literature on several disorders before focusing on the assumptions about cognition underlying the studies; specifically, that cognition can be thought of as a network of interrelated variables and that changes in these interrelationships, or cognitive rearrangement, can occur in neurological disorders. Next the review appraises how well network models can provide a "map" of this cognitive "territory". In particular, the review considers the lack of correspondence between the variables and properties of network models and cognitive functioning. The challenges of explicitly accounting for latent cognitive constructs and making inferences about cognition based on associative, as opposed to dissociative, methods are also discussed. It is concluded that the validity of network neuropsychological models is yet to be established and that cognitive theory and experiments, as well as network models, are needed to develop and interpret better maps.

Age and Sex Invariance of the Woodcock-Johnson IV Tests of Cognitive Abilities: Evidence from Psychometric Network Modeling

The Woodcock-Johnson IV Tests of Cognitive Abilities (WJ IV COG) is a comprehensive assessment battery designed to assess broad and narrow cognitive abilities, as defined by the Cattell-Horn-Carroll (CHC) theory of intelligence. Previous studies examined the invariance of the WJ assessments across sex and age groups using factor analytic methods. Psychometric network modeling is an alternative methodology that can address both direct and indirect relationships among the observed variables. In this study, we employed psychometric network modeling to examine the invariance of the WJ IV COG across sex and age groups. Using a normative sample (n = 4212 participants) representative of the United States population, we tested the extent to which the factorial structure of the WJ IV COG aligned with CHC theory for the school-aged sample. Next, we used psychometric network modeling as a data-driven method to investigate whether the network structure of the WJ IV COG remains similar across different sex and age (age 6 to 19, inclusively) groups. Our results showed that the WJ IV COG maintained the same network structure across all age and sex groups, although the network structure at younger ages indicated weaker relationships among some subtests. Overall, the results provide construct validity evidence for the WJ IV COG, based on both theoretical and data-driven methods.

Investigating the Dimensionality of Early Numeracy Using the Bifactor Exploratory Structural Equation Modeling Framework

The few studies that have analyzed the factorial structure of early number skills have mainly used confirmatory factor analysis (CFA) and have yielded inconsistent results, since early numeracy is considered to be unidimensional, multidimensional or even underpinned by a general factor. Recently, the bifactor exploratory structural equation modeling (bifactor-ESEM)-which has been proposed as a way to overcome the shortcomings of both the CFA and the exploratory structural equation modeling (ESEM)-proved to be valuable to account for the multidimensionality and the hierarchical nature of several psychological constructs. The present study is the first to investigate the dimensionality of early number skills measurement through the application of the bifactor-ESEM framework. Using data from 644 prekindergarten and kindergarten children (4 to 6 years old), several competing models were contrasted: the one-factor CFA model; the independent cluster model (ICM-CFA); the exploratory structural equation modeling (ESEM); and their bifactor counterpart (bifactor-CFA and bifactor-ESEM, respectively). Results indicated acceptable fit indexes for the one-factor CFA and the ICM-CFA models and excellent fit for the others. Among these, the bifactor-ESEM with one general factor and three specific factors (Counting, Relations, Arithmetic) not only showed the best model fit, but also the best coherent factor loadings structure and full measurement invariance across gender. The bifactor-ESEM appears relevant to help disentangle and account for general and specific factors of early numerical ability. While early numerical ability appears to be mainly underpinned by a general factor whose exact nature still has to be determined, this study highlights that specific latent dimensions with substantive value also exist. Identifying these specific facets is important in order to increase quality of early numerical ability measurement, predictive validity, and for practical implications.

Internalizing symptoms, well-being, and correlates in adolescence: A multiverse exploration via cross-lagged panel network models

Internalizing symptoms are the most prevalent mental health problem in adolescents, with sharp increases seen, particularly for girls, and evidence that young people today report more problems than previous generations. It is therefore critical to measure and monitor these states on a large scale and consider correlates. We used novel panel network methodology to explore relationships between internalizing symptoms, well-being, and inter/intrapersonal indicators. A multiverse design was used with 32 conditions to consider the stability of results across arbitrary researcher decisions in a large community sample over three years (N = 15,843, aged 11-12 at Time 1). Networks were consistently similar for girls and boys. Stable trait-like effects within anxiety, attentional, and social indicators were found. Within-person networks were densely connected and suggested mental health and inter/intrapersonal correlates related to one another in similar complex ways. The multiverse design suggested the particular operationalization of items can substantially influence conclusions. Nevertheless, indicators such as thinking clearly, unhappiness, dealing with stress, and worry showed more consistent centrality, suggesting these indicators may play particularly important roles in the development of mental health in adolescence.

Bridging Brain and Cognition: A Multilayer Network Analysis of Brain Structural Covariance and General Intelligence in a Developmental Sample of Struggling Learners

Network analytic methods that are ubiquitous in other areas, such as systems neuroscience, have recently been used to test network theories in psychology, including intelligence research. The network or mutualism theory of intelligence proposes that the statistical associations among cognitive abilities (e.g., specific abilities such as vocabulary or memory) stem from causal relations among them throughout development. In this study, we used network models (specifically LASSO) of cognitive abilities and brain structural covariance (grey and white matter) to simultaneously model brain-behavior relationships essential for general intelligence in a large (behavioral, N = 805; cortical volume, N = 246; fractional anisotropy, N = 165) developmental (ages 5-18) cohort of struggling learners (CALM). We found that mostly positive, small partial correlations pervade our cognitive, neural, and multilayer networks. Moreover, using community detection (Walktrap algorithm) and calculating node centrality (absolute strength and bridge strength), we found convergent evidence that subsets of both cognitive and neural nodes play an intermediary role 'between' brain and behavior. We discuss implications and possible avenues for future studies.

Positive and negative urgency as a single coherent construct: Evidence from a large-scale network analysis in clinical and non-clinical samples

Aims

Negative and positive urgency are emotion‐related impulsivity traits that are thought to be transdiagnostic factors in psychopathology. However, it has recently been claimed that these two traits are closely related to each other and that considering them separately might have limited conceptual and methodological value. The present study aimed to examine whether positive and negative urgency constructs constitute separate impulsivity traits.

Methods

In contrast to previous studies that have used latent variable approaches, this study employed an item‐based network analysis conducted in two different samples: a large sample of non‐clinical participants (N = 18,568) and a sample of clinical participants with psychiatric disorders (N = 385).

Results

The network analysis demonstrated that items denoting both positive and negative urgency cohere as a single cluster of items termed “general urgency” in both clinical and non‐clinical samples, thereby suggesting that differentiating positive and negative urgency as separate constructs is not necessary.

Conclusion

These findings have important implications for the conceptualization and assessment of urgency and, more broadly, for future research on impulsivity, personality, and psychopathology.

Exploratory Factor Analyses of the French WISC-V (WISC-VFR) for Five Age Groups: Analyses Based on the Standardization Sample

This study investigated the factor structure of the French Wechsler Intelligence Scale for Children–Fifth Edition with five standardization sample age groups (6-7, 8-9, 10-11, 12-13, 14-16 years) using hierarchical exploratory factor analysis followed by Schmid–Leiman procedure. The primary research questions included (a) how many French Wechsler Intelligence Scale for Children–Fifth Edition factors should be extracted and retained in each age subgroup, (b) how are subtests associated with the latent factors, (c) was there evidence for the publisher’s claim of five first-order factors and separate Visual Spatial and Fluid Reasoning factors, (d) what proportion of variance was due to general intelligence versus the first-order group ability factors following a Schmid–Leiman procedure, and (e) do results support the age differentiation hypothesis? Results suggested that four factors might be sufficient for all five age groups and results did not support the distinction between Visual Spatial and Fluid Reasoning factors. While the general factor accounted for the largest portions of variance, the four first-order factors accounted for small unique portions of variance. Results did not support the age differentiation hypothesis because the number of factors remained the same across age groups, and there was no change in the percentage of variance accounted for by the general factor across age groups.

Multivariate Patterns of Brain-Behavior-Environment Associations in the Adolescent Brain and Cognitive Development Study

BACKGROUND

Adolescence is a critical developmental stage. A key challenge is to characterize how variation in adolescent brain organization relates to psychosocial and environmental influences.

METHODS

We used canonical correlation analysis to discover distinct patterns of covariation between measures of brain organization (brain morphometry, intracortical myelination, white matter integrity, and resting-state functional connectivity) and individual, psychosocial, and environmental factors in a nationally representative U.S. sample of 9623 individuals (aged 9-10 years, 49% female) participating in the Adolescent Brain and Cognitive Development (ABCD) study.

RESULTS

These analyses identified 14 reliable modes of brain-behavior-environment covariation (canonical r_discovery = .21 to .49, canonical r_test = .10 to .39, p_{false discovery rate corrected} < .0001). Across modes, neighborhood environment, parental characteristics, quality of family life, perinatal history, cardiometabolic health, cognition, and psychopathology had the most consistent and replicable associations with multiple measures of brain organization; positive and negative exposures converged to form patterns of psychosocial advantage or adversity. These showed modality-general, respectively positive or negative, associations with brain structure and function with little evidence of regional specificity. Nested within these cross-modal patterns were more specific associations between prefrontal measures of morphometry, intracortical myelination, and functional connectivity with affective psychopathology, cognition, and family environment.

CONCLUSIONS

We identified clusters of exposures that showed consistent modality-general associations with global measures of brain organization. These findings underscore the importance of understanding the complex and intertwined influences on brain organization and mental function during development and have the potential to inform public health policies aiming toward interventions to improve mental well-being.

Investigating the Structure of Intelligence Using Latent Variable and Psychometric Network Modeling: A Commentary and Reanalysis

In a recent publication in the Journal of Intelligence, Dennis McFarland mischaracterized previous research using latent variable and psychometric network modeling to investigate the structure of intelligence. Misconceptions presented by McFarland are identified and discussed. We reiterate and clarify the goal of our previous research on network models, which is to improve compatibility between psychological theories and statistical models of intelligence. WAIS-IV data provided by McFarland were reanalyzed using latent variable and psychometric network modeling. The results are consistent with our previous study and show that a latent variable model and a network model both provide an adequate fit to the WAIS-IV. We therefore argue that model preference should be determined by theory compatibility. Theories of intelligence that posit a general mental ability (general intelligence) are compatible with latent variable models. More recent approaches, such as mutualism and process overlap theory, reject the notion of general mental ability and are therefore more compatible with network models, which depict the structure of intelligence as an interconnected network of cognitive processes sampled by a battery of tests. We emphasize the importance of compatibility between theories and models in scientific research on intelligence.

How to Compare Psychometric Factor and Network Models

In memory of Dr. Dennis John McFarland, who passed away recently, our objective is to continue his efforts to compare psychometric networks and latent variable models statistically. We do so by providing a commentary on his latest work, which he encouraged us to write, shortly before his death. We first discuss the statistical procedure McFarland used, which involved structural equation modeling (SEM) in standard SEM software. Next, we evaluate the penta-factor model of intelligence. We conclude that (1) standard SEM software is not suitable for the comparison of psychometric networks with latent variable models, and (2) the penta-factor model of intelligence is only of limited value, as it is nonidentified. We conclude with a reanalysis of the Wechlser Adult Intelligence Scale data McFarland discussed and illustrate how network and latent variable models can be compared using the recently developed R package Psychonetrics. Of substantive theoretical interest, the results support a network interpretation of general intelligence. A novel empirical finding is that networks of intelligence replicate over standardization samples.

Network analysis of the structure and change in the mini-mental state examination: a nationally representative sample

PURPOSE

The structure of the mini-mental state examination (MMSE) is inconsistent across factor analytic studies, and yet to be examined based on network analysis. The current study aims to identify the (I) cross-sectional network structure and (II) longitudinal network changes of the MMSE.

METHODS

The MMSE was administered to a nationally representative sample of older adults (age 50 and over) in Ireland twice over 4 years (2012-2013: N = 7207; 2016: N = 5715). Psychometric network analysis was computed at each time point to identify structure, strength and magnitude of the network associations. Item clustering was examined, and modularity scores were computed to measure the overall strength of clustering. Centrality indices were used to identify the main aspects of the MMSE. Longitudinal differences between the networks were examined.

RESULTS

Cross-sectionally, the MMSE network structure clustered into a single community (modularity score = 0) with orientation items identified as most central. Longitudinally, the MMSE was time invariant regarding structure, centrality and magnitude of the positive associations between the items. The average magnitude of the negative associations increased over time[(t(65.15) = 3.78, p < 0.001; time 1: M = - 0.59, SD = 0.58 time 2: M = - 1.65, SD = 1.97] as did their percentage.

CONCLUSION

Network analysis of the MMSE showed that the measure consisted of a single entity of cognitive functioning irrespective of time. Orientation items were repeatedly identified as most central. Longitudinal changes of the network were evident in increased negative associations between selected cognitive components after 4 years of follow-up. These changes may be explained by neuro-cognitive compensation processes.

Toward an Integrative Psychometric Model of Emotions

Emotions are part and parcel of the human condition, but their nature is debated. Three broad classes of theories about the nature of emotions can be distinguished: affect-program theories, constructionist theories, and appraisal theories. Integrating these broad classes of theories into a unifying theory is challenging. An integrative psychometric model of emotions can inform such a theory because psychometric models are intertwined with theoretical perspectives about constructs. To identify an integrative psychometric model, we delineate properties of emotions stated by emotion theories and investigate whether psychometric models account for these properties. Specifically, an integrative psychometric model of emotions should allow (a) identifying distinct emotions (central in affect-program theories), (b) between- and within-person variations of emotions (central in constructionist theories), and (c) causal relationships between emotion components (central in appraisal theories). Evidence suggests that the popular reflective and formative latent variable models-in which emotions are conceptualized as unobservable causes or consequences of emotion components-cannot account for all properties. Conversely, a psychometric network model-in which emotions are conceptualized as systems of causally interacting emotion components-accounts for all properties. The psychometric network model thus constitutes an integrative psychometric model of emotions, facilitating progress toward a unifying theory.

Psychometric network models from time-series and panel data

Researchers in the field of network psychometrics often focus on the estimation of Gaussian graphical models (GGMs)-an undirected network model of partial correlations-between observed variables of cross-sectional data or single-subject time-series data. This assumes that all variables are measured without measurement error, which may be implausible. In addition, cross-sectional data cannot distinguish between within-subject and between-subject effects. This paper provides a general framework that extends GGM modeling with latent variables, including relationships over time. These relationships can be estimated from time-series data or panel data featuring at least three waves of measurement. The model takes the form of a graphical vector-autoregression model between latent variables and is termed the ts-lvgvar when estimated from time-series data and the panel-lvgvar when estimated from panel data. These methods have been implemented in the software package psychonetrics, which is exemplified in two empirical examples, one using time-series data and one using panel data, and evaluated in two large-scale simulation studies. The paper concludes with a discussion on ergodicity and generalizability. Although within-subject effects may in principle be separated from between-subject effects, the interpretation of these results rests on the intensity and the time interval of measurement and on the plausibility of the assumption of stationarity.

The Effects of Using Partial or Uncorrected Correlation Matrices When Comparing Network and Latent Variable Models

Network models of the WAIS-IV based on regularized partial correlation matrices have been reported to outperform latent variable models based on uncorrected correlation matrices. The present study sought to compare network and latent variable models using both partial and uncorrected correlation matrices with both types of models. The results show that a network model provided better fit to matrices of partial correlations but latent variable models provided better fit to matrices of full correlations. This result is due to the fact that the use of partial correlations removes most of the covariance common to WAIS-IV tests. Modeling should be based on uncorrected correlations since these represent the majority of shared variance between WAIS-IV test scores.

Psychometric Network Analysis of the Hungarian WAIS

The positive manifold-the finding that cognitive ability measures demonstrate positive correlations with one another-has led to models of intelligence that include a general cognitive ability or general intelligence (g). This view has been reinforced using factor analysis and reflective, higher-order latent variable models. However, a new theory of intelligence, Process Overlap Theory (POT), posits that g is not a psychological attribute but an index of cognitive abilities that results from an interconnected network of cognitive processes. These competing theories of intelligence are compared using two different statistical modeling techniques: (a) latent variable modeling and (b) psychometric network analysis. Network models display partial correlations between pairs of observed variables that demonstrate direct relationships among observations. Secondary data analysis was conducted using the Hungarian Wechsler Adult Intelligence Scale Fourth Edition (H-WAIS-IV). The underlying structure of the H-WAIS-IV was first assessed using confirmatory factor analysis assuming a reflective, higher-order model and then reanalyzed using psychometric network analysis. The compatibility (or lack thereof) of these theoretical accounts of intelligence with the data are discussed.